A detection algorithm for cherry fruits based on the improved YOLO-v4 model

TopoGeoFusion: Integrating object topology based feature computation methods into geometrical feature analysis to enhance classification performance

This study used smartphone captured RGB images of gooseberries to automatically sort into standard, premium, or rejected categories based on topology. Main challenges addressed include, separation of touching or overlapping fruits into individual entities and new method called 'TopoGeoFusion' that combines basic geometrical features with topology aware features computed from the fruits to assess the grade or maturity. Quality assessment helps in grading the fruit to determine market suitability and intelligent camera applications. Computer Vision-based techniques have been applied to automatically grade the quality of gooseberries as standard, premium, or rejected according to fruit maturity. Smartphone-captured images of 1697 Indian Star Gooseberries are contributed to the study. This work acquired images consisting multiple fruits with overlapping and non-overlapping boundaries for concurrent quality assessment. Multiple classifiers such as Random Forest, SVM, Naive Bayes, Decision Tree, and KNN were applied to grade the gooseberry fruit. Random Forest classification with a fusion feature model resulted in an accuracy of 100 % towards reject, standard, and premium classes for test sets with four training strategies. The proposed segmentation model proves reliable in fruit detection & extraction with an average mAP of 0.56, resulting in an acceptable model for grade assessment.•The study highlights the effectiveness of TopoGeoFusion in automating the grading process of gooseberry fruits using topologically computed features.•The developed models exhibit high accuracy and reliability, even in challenging scenarios such as overlapping and touching fruits.•The method provides the technique to detect and extract the occluded objects and compute the features based on the partial object's topology.

Entomopathogenic nematode detection and counting model developed based on A-star algorithm

Entomopathogenic nematodes are soil-dwelling living organisms widely employed in the biological control of agricultural insect pests, serving as a significant alternative to pesticides. In laboratory procedures, the counting process remains the most common, labor-intensive, time-consuming, and approximate aspect of studies related to entomopathogenic nematodes. In this context, a novel method has been proposed for the detection and quantification of Steinernema feltiae isolate using computer vision on microscope images. The proposed method involves two primary algorithmic steps: framing and isolation. Compared to YOLO-V5m, YOLO-V7m, and YOLO-V8m, the A-star-based developed network demonstrates significantly improved detection accuracy compared to other networks. The novel method is particularly effective in facilitating the detection of overlapping nematodes. The developed algorithm excludes processes that increase space and time complexity, such as the weight document, which contains the learned parameters of the deep learning model, model integration, and prediction time, resulting in more efficient operation. The results indicate the feasibility of the proposed method for detecting and counting entomopathogenic nematodes.

Enhancing kiwifruit flower pollination detection through frequency domain feature fusion: a novel approach to agricultural monitoring

The pollination process of kiwifruit flowers plays a crucial role in kiwifruit yield. Achieving accurate and rapid identification of the four stages of kiwifruit flowers is essential for enhancing pollination efficiency. In this study, to improve the efficiency of kiwifruit pollination, we propose a novel full-stage kiwifruit flower pollination detection algorithm named KIWI-YOLO, based on the fusion of frequency-domain features. Our algorithm leverages frequency-domain and spatial-domain information to improve recognition of contour-detailed features and integrates decision-making with contextual information. Additionally, we incorporate the Bi-Level Routing Attention (BRA) mechanism with C3 to enhance the algorithm's focus on critical areas, resulting in accurate, lightweight, and fast detection. The algorithm achieves a m A P 0.5 of 91.6% with only 1.8M parameters, the AP of the Female class and the Male class reaches 95% and 93.5%, which is an improvement of 3.8%, 1.2%, and 6.2% compared with the original algorithm. Furthermore, the Recall and F1-score of the algorithm are enhanced by 5.5% and 3.1%, respectively. Moreover, our model demonstrates significant advantages in detection speed, taking only 0.016s to process an image. The experimental results show that the algorithmic model proposed in this study can better assist the pollination of kiwifruit in the process of precision agriculture production and help the development of the kiwifruit industry.

Advancing jasmine tea production: YOLOv7-based real-time jasmine flower detection

BACKGROUND

To produce jasmine tea of excellent quality, it is crucial to select jasmine flowers at their optimal growth stage during harvesting. However, achieving this goal remains a challenge due to environmental and manual factors. This study addresses this issue by classifying different jasmine flowers based on visual attributes using the YOLOv7 algorithm, one of the most advanced algorithms in convolutional neural networks.

RESULTS

The mean average precision (mAP value) for detecting jasmine flowers using this model is 0.948, and the accuracy for five different degrees of openness of jasmine flowers, namely small buds, buds, half-open, full-open and wiltered, is 87.7%, 90.3%, 89%, 93.9% and 86.4%, respectively. Meanwhile, other ways of processing the images in the dataset, such as blurring and changing the brightness, also increased the credibility of the algorithm.

CONCLUSION

This study shows that it is feasible to use deep learning algorithms for distinguishing jasmine flowers at different growth stages. This study can provide a reference for jasmine production estimation and for the development of intelligent and precise flower-picking applications to reduce flower waste and production costs. © 2024 Society of Chemical Industry.

Optimization of intelligent guided vehicle vision navigation based on improved YOLOv2

Addressing the challenge of limited accuracy and real-time performance in intelligent guided vehicle (IGV) image recognition and detection, typically reliant on traditional feature extraction approaches. This study delves into a visual navigation detection method using an improved You Only Look Once (YOLO) model-simplified YOLOv2 (SYOLOv2) to satisfy the complex operating conditions of the port and the limitations of IGV hardware computing. The convolutional neural network structure of YOLOv2 is refined to ensure adaptability to varying weather conditions using a single image. Preprocessing of images involves Contrast Limited Adaptive Histogram Equalization (CLAHE), while an adaptive image resolution detection model, contingent upon vehicle speed, is proposed to enhance the detection performance. The comparative experiments conducted on image datasets reflective of actual road conditions and weather conditions demonstrate notable enhancements in accuracy and frames transmitted per second compared to conventional methods. These improvements signify the efficacy of the proposed approach in meeting the stringent requirements for real-time detection on IGV platforms.

Defect identification of electricity transmission line insulators based on the improved lightweight network model with computer vision assistance

This work aims to ensure the safe operation of electricity transmission lines and reduce costs and maintenance difficulties. It studies the application of computer vision (CV) in the defect identification of electricity transmission lines. In addition, this work proposes a method to improve the lightweight network model to provide an effective identification model to solve the problem of electricity transmission line defects. Firstly, GraphCut segmentation and Laplace algorithms are employed to expand and sharpen the electricity transmission line image. Secondly, in light of the Depth Separable Convolution algorithm, a defect detection model for the electricity transmission line insulator is proposed based on the You Only Look Once 4 (YOLOv4) network. Moreover, MobileNetV1 is utilized to improve this lightweight network model. Finally, this work uses ImageNet, a large public dataset, to validate the proposed model experimentally. The research results reveal that: (1) In the model testing results, all research indicators of the model are greater than 90 %, indicating an excellent detection accuracy of this model. (2) The improved YOLOv4 model can increase the detection speed to 53 frames/s at the cost of 2.4 % accuracy. (3) After image sharpening, the improved YOLOv4 model has promoted the insulator defects' detection ability to a certain extent. The above outcomes suggest that the improved YOLOv4 model can predict more efficiently and accurately and reduce unnecessary false positives. This illustrates that the proposed model is feasible and is expected to be applied to the defect identification of electricity transmission lines in practice. These findings fully demonstrate this work's vital value in enhancing the prediction efficiency and accuracy, thus offering a strong preference for the defect identification of electricity transmission lines in practical applications.

Classification of peanut pod rot based on improved YOLOv5s

Peanut pod rot is one of the major plant diseases affecting peanut production and quality over China, which causes large productivity losses and is challenging to control. To improve the disease resistance of peanuts, breeding is one significant strategy. Crucial preventative and management measures include grading peanut pod rot and screening high-contributed genes that are highly resistant to pod rot should be carried out. A machine vision-based grading approach for individual cases of peanut pod rot was proposed in this study, which avoids time-consuming, labor-intensive, and inaccurate manual categorization and provides dependable technical assistance for breeding studies and peanut pod rot resistance. The Shuffle Attention module has been added to the YOLOv5s (You Only Look Once version 5 small) feature extraction backbone network to overcome occlusion, overlap, and adhesions in complex backgrounds. Additionally, to reduce missing and false identification of peanut pods, the loss function CIoU (Complete Intersection over Union) was replaced with EIoU (Enhanced Intersection over Union). The recognition results can be further improved by introducing grade classification module, which can read the information from the identified RGB images and output data like numbers of non-rotted and rotten peanut pods, the rotten pod rate, and the pod rot grade. The Precision value of the improved YOLOv5s reached 93.8%, which was 7.8%, 8.4%, and 7.3% higher than YOLOv5s, YOLOv8n, and YOLOv8s, respectively; the mAP (mean Average Precision) value was 92.4%, which increased by 6.7%, 7.7%, and 6.5%, respectively. Improved YOLOv5s has an average improvement of 6.26% over YOLOv5s in terms of recognition accuracy: that was 95.7% for non-rotted peanut pods and 90.8% for rotten peanut pods. This article presented a machine vision- based grade classification method for peanut pod rot, which offered technological guidance for selecting high-quality cultivars with high resistance to pod rot in peanut.

Research on duck egg recognition algorithm based on improved YOLOv4

1. The following study addressed the problem of small duck eggs as challenging to detect and identify for pick up in complex free-range duck farm environments. It introduces improvements to the YOLOv4 convolutional neural network target detection algorithm, based on the working conditions of egg-picking robots.2. Specifically, one scale of anchor boxes was removed from the prediction network, and a duck egg labelling dataset was established to make the improved algorithm YOLOv4-ours better match the working state of egg-picking robots and enhance detection performance.3. Through multiple comparative experiments, the YOLOv4-ours object detection algorithm exhibited superior overall performance, achieving a precision of 98.85%, recall of 96.67%, and an average precision of 98.60% and F1 score increased to 97%. Compared to the original YOLOv4 model, these improvements represented increases of 1.89%, 3.41%, 1.32%, and 1.04%, respectively. Furthermore, detection time was reduced from 0.26 seconds per image to 0.20 seconds.4. The enhanced model accurately detected duck eggs in free-range duck housing, effectively meeting the real-time egg identification and picking requirements.

Multitargets Joint Training Lightweight Model for Object Detection of Substation

The object detection of the substation is the key to ensuring the safety and reliable operation of the substation. The traditional image detection algorithms use the corresponding texture features of single-class objects and would not handle other different class objects easily. The object detection algorithm based on deep networks has generalization, and its sizeable complex backbone limits the application in the substation monitoring terminals with weak computing power. This article proposes a multitargets joint training lightweight model. The proposed model uses the feature maps of the complex model and the labels of objects in images as training multitargets. The feature maps have deeper feature information, and the feature maps of complex networks have higher information entropy than lightweight networks have. This article proposes the heat pixels method to improve the adequate object information because of the imbalance of the proportion between the foreground and the background. The heat pixels method is designed as a kind of reverse network calculation and reflects the object's position to the pixels of the feature maps. The temperature of the pixels indicates the probability of the existence of the objects in the locations. Three different lightweight networks use the complex model feature maps and the traditional tags as the training multitargets. The public dataset VOC and the substation equipment dataset are adopted in the experiments. The experimental results demonstrate that the proposed model can effectively improve object detection accuracy and reduce the time-consuming and calculation amount.

YOLO and residual network for colorectal cancer cell detection and counting

The HT-29 cell line, derived from human colon cancer, is valuable for biological and cancer research applications. Early detection is crucial for improving the chances of survival, and researchers are introducing new techniques for accurate cancer diagnosis. This study introduces an efficient deep learning-based method for detecting and counting colorectal cancer cells (HT-29). The colorectal cancer cell line was procured from a company. Further, the cancer cells were cultured, and a transwell experiment was conducted in the lab to collect the dataset of colorectal cancer cell images via fluorescence microscopy. Of the 566 images, 80 % were allocated to the training set, and the remaining 20 % were assigned to the testing set. The HT-29 cell detection and counting in medical images is performed by integrating YOLOv2, ResNet-50, and ResNet-18 architectures. The accuracy achieved by ResNet-18 is 98.70 % and ResNet-50 is 96.66 %. The study achieves its primary objective by focusing on detecting and quantifying congested and overlapping colorectal cancer cells within the images. This innovative work constitutes a significant development in overlapping cancer cell detection and counting, paving the way for novel advancements and opening new avenues for research and clinical applications. Researchers can extend the study by exploring variations in ResNet and YOLO architectures to optimize object detection performance. Further investigation into real-time deployment strategies will enhance the practical applicability of these models.

Accurate real-time obstacle detection of coal mine driverless electric locomotive based on ODEL-YOLOv5s

The accurate identification and real-time detection of obstacles have been considered the premise to ensure the safe operation of coal mine driverless electric locomotives. The harsh coal mine roadway environment leads to low detection accuracy of obstacles based on traditional detection methods such as LiDAR and machine learning, and these traditional obstacle detection methods lead to slower detection speeds due to excessive computational reasoning. To address the above-mentioned problems, we propose a deep learning-based ODEL-YOLOv5s detection model based on the conventional YOLOv5s. In this work, several data augmentation methods are introduced to increase the diversity of obstacle features in the dataset images. An attention mechanism is introduced to the neck of the model to improve the focus of the model on obstacle features. The three-scale prediction of the model is increased to a four-scale prediction to improve the detection ability of the model for small obstacles. We also optimize the localization loss function and non-maximum suppression method of the model to improve the regression accuracy and reduce the redundancy of the prediction boxes. The experimental results show that the mean average precision (mAP) of the proposed ODEL-YOLOv5s model is increased from 95.2 to 98.9% compared to the conventional YOLOv5s, the average precision of small obstacle rock is increased from 89.2 to 97.9%, the detection speed of the model is 60.2 FPS, and it has better detection performance compared with other detection models, which can provide technical support for obstacle identification and real-time detection of coal mine driverless electric locomotives.

Automated hand-marked semantic text recognition from photographs

Automated text recognition techniques have made significant advancements; however, certain tasks still present challenges. This study is motivated by the need to automatically recognize hand-marked text on construction defect tags among millions of photographs. To address this challenge, we investigated three methods for automating hand-marked semantic text recognition (HMSTR)-a modified scene text recognition-based (STR) approach, a two-step HMSTR approach, and a lumped approach. The STR approach involves locating marked text using an object detection model and recognizing it using a competition-winning STR model. Similarly, the two-step HMSTR approach first localizes the marked text and then recognizes the semantic text using an image classification model. By contrast, the lumped approach performs both localization and identification of marked semantic text in a single step using object detection. Among these approaches, the two-step HMSTR approach achieved the highest F1 score (0.92) for recognizing circled text, followed by the STR approach (0.87) and the lumped approach (0.78). To validate the generalizability of the two-step HMSTR approach, subsequent experiments were conducted using check-marked text, resulting in an F1 score of 0.88. Although the proposed methods have been tested specifically with tags, they can be extended to recognize marked text in reports or books.

Instance Segmentation and Berry Counting of Table Grape before Thinning Based on AS-SwinT

Berry thinning is one of the most important tasks in the management of high-quality table grapes. Farmers often thin the berries per cluster to a standard number by counting. With an aging population, it is hard to find adequate skilled farmers to work during thinning season. It is urgent to design an intelligent berry-thinning machine to avoid exhaustive repetitive labor. A machine vision system that can determine the number of berries removed and locate the berries removed is a challenge for the thinning machine. A method for instance segmentation of berries and berry counting in a single bunch is proposed based on AS-SwinT. In AS-SwinT, Swin Transformer is performed as the backbone to extract the rich characteristics of grape berries. An adaptive feature fusion is introduced to the neck network to sufficiently preserve the underlying features and enhance the detection of small berries. The size of berries in the dataset is statistically analyzed to optimize the anchor scale, and Soft-NMS is used to filter the candidate frames to reduce the missed detection of densely shaded berries. Finally, the proposed method could achieve 65.7 APbox, 95.0 A P 0.5 box , 57 A P s box , 62.8 APmask, 94.3 A P 0.5 mask , 48 A P s mask , which is markedly superior to Mask R-CNN, Mask Scoring R-CNN, and Cascade Mask R-CNN. Linear regressions between predicted numbers and actual numbers are also developed to verify the precision of the proposed model. RMSE and R2 values are 7.13 and 0.95, respectively, which are substantially higher than other models, showing the advantage of the AS-SwinT model in berry counting estimation.

Deep learning supported machine vision system to precisely automate the wild blueberry harvester header

An operator of a wild blueberry harvester faces the fatigue of manually adjusting the height of the harvester's head, considering spatial variations in plant height, fruit zone, and field topography affecting fruit yield. For stress-free harvesting of wild blueberries, a deep learning-supported machine vision control system has been developed to detect the fruit height and precisely auto-adjust the header picking teeth rake position. The OpenCV AI Kit (OAK-D) was used with YOLOv4-tiny deep learning model with code developed in Python to solve the challenge of matching fruit heights with the harvester's head position. The system accuracy was statistically evaluated with R2 (coefficient of determination) and σ (standard deviation) measured on the difference in distances between the berries picking teeth and average fruit heights, which were 72, 43% and 2.1, 2.3 cm for the auto and manual head adjustment systems, respectively. This innovative system performed well in weed-free areas but requires further work to operate in weedy sections of the fields. Benefits of using this system include automated control of the harvester's head to match the header picking rake height to the level of the fruit height while reducing the operator's stress by creating safer working environments.

YOLO-Weld: A Modified YOLOv5-Based Weld Feature Detection Network for Extreme Weld Noise

Weld feature point detection is a key technology for welding trajectory planning and tracking. Existing two-stage detection methods and conventional convolutional neural network (CNN)-based approaches encounter performance bottlenecks under extreme welding noise conditions. To better obtain accurate weld feature point locations in high-noise environments, we propose a feature point detection network, YOLO-Weld, based on an improved You Only Look Once version 5 (YOLOv5). By introducing the reparameterized convolutional neural network (RepVGG) module, the network structure is optimized, enhancing detection speed. The utilization of a normalization-based attention module (NAM) in the network enhances the network's perception of feature points. A lightweight decoupled head, RD-Head, is designed to improve classification and regression accuracy. Furthermore, a welding noise generation method is proposed, increasing the model's robustness in extreme noise environments. Finally, the model is tested on a custom dataset of five weld types, demonstrating better performance than two-stage detection methods and conventional CNN approaches. The proposed model can accurately detect feature points in high-noise environments while meeting real-time welding requirements. In terms of the model's performance, the average error of detecting feature points in images is 2.100 pixels, while the average error in the world coordinate system is 0.114 mm, sufficiently meeting the accuracy needs of various practical welding tasks.

A mobile-based system for maize plant leaf disease detection and classification using deep learning

Artificial Intelligence has been used for many applications such as medical, communication, object detection, and object tracking. Maize crop, which is the major crop in the world, is affected by several types of diseases which lower its yield and affect the quality. This paper focuses on this issue and provides an application for the detection and classification of diseases in maize crop using deep learning models. In addition to this, the developed application also returns the segmented images of affected leaves and thus enables us to track the disease spots on each leaf. For this purpose, a dataset of three maize crop diseases named Blight, Sugarcane Mosaic virus, and Leaf Spot is collected from the University Research Farm Koont, PMAS-AAUR at different growth stages on contrasting weather conditions. This data was used for training different prediction models including YOLOv3-tiny, YOLOv4, YOLOv5s, YOLOv7s, and YOLOv8n and the reported prediction accuracy was 69.40%, 97.50%, 88.23%, 93.30%, and 99.04% respectively. Results demonstrate that the prediction accuracy of the YOLOv8n model is higher than the other applied models. This model has shown excellent results while localizing the affected area of the leaf accurately with a higher confidence score. YOLOv8n is the latest model used for the detection of diseases as compared to the other approaches in the available literature. Also, worked on sugarcane mosaic virus using deep learning models has also been reported for the first time. Further, the models with high accuracy have been embedded in a mobile application to provide a real-time disease detection facility for end users within a few seconds.

Waste Detection System Based on Data Augmentation and YOLO_EC

The problem of waste classification has been a major concern for both the government and society, and whether waste can be effectively classified will affect the sustainable development of human society. To perform fast and efficient detection of waste targets in the sorting process, this paper proposes a data augmentation + YOLO_EC waste detection system. First of all, because of the current shortage of multi-objective waste classification datasets, the heavy workload of human data collection, and the limited improvement of data features by traditional data augmentation methods, DCGAN (deep convolution generative adversarial networks) was optimized by improving the loss function, and an image-generation model was established to realize the generation of multi-objective waste images; secondly, with YOLOv4 (You Only Look Once version 4) as the basic model, EfficientNet is used as the backbone feature extraction network to realize the light weight of the algorithm, and at the same time, the CA (coordinate attention) attention mechanism is introduced to reconstruct the MBConv module to filter out high-quality information and enhance the feature extraction ability of the model. Experimental results show that on the HPU_WASTE dataset, the proposed model outperforms other models in both data augmentation and waste detection.

A study on Shine-Muscat grape detection at maturity based on deep learning

The efficient detection of grapes is a crucial technology for fruit-picking robots. To better identify grapes from branch shading that is similar to the fruit color and improve the detection accuracy of green grapes due to cluster adhesion, this study proposes a Shine-Muscat Grape Detection Model (S-MGDM) based on improved YOLOv3 for the ripening stage. DenseNet is fused in the backbone feature extraction network to extract richer underlying grape information; depth-separable convolution, CBAM, and SPPNet are added in the multi-scale detection module to increase the perceptual field of grape targets and reduce the model computation; meanwhile, PANet is combined with FPN to promote inter-network information flow and iteratively extract grape features. In addition, the CIOU regression loss function is used and the prior frame size is modified by the k-means algorithm to improve the accuracy of detection. The improved detection model achieves an AP value of 96.73% and an F1 value of 91% on the test set, which are 3.87% and 3% higher than the original network model, respectively; the average detection speed under GPU reaches 26.95 frames/s, which is 6.49 frames/s higher than the original model. The comparison results with several mainstream detection algorithms such as SSD and YOLO series show that the method has excellent detection accuracy and good real-time performance, which is an important reference value for the problem of accurate identification of Shine-Muscat grapes at maturity.

A Study of an Online Tracking System for Spark Images of Abrasive Belt-Polishing Workpieces

During the manual grinding of blades, the workers can estimate the material removal rate based on their experiences from observing the characteristics of the grinding sparks, leading to low grinding accuracy and low efficiency and affecting the processing quality of the blades. As an alternative to the recognition of spark images by the human eye, we used the deep learning algorithm YOLO5 to perform target detection on spark images and obtain spark image regions. First the spark images generated during one turbine blade-grinding process were collected, and some of the images were selected as training samples, with the remaining images used as test samples, which were labelled with LabelImg. Afterwards, the selected images were trained with YOLO5 to obtain an optimisation model. In the end, the trained optimisation model was used to predict the images of the test set. The proposed method was able to detect spark image regions quickly and accurately, with an average accuracy of 0.995. YOLO4 was also used to train and predict spark images, and the two methods were compared. Our findings show that YOLO5 is faster and more accurate than the YOLO4 target detection algorithm and can replace manual observation, laying a specific foundation for the automatic segmentation of spark images and the study of the relationship between the material removal rate and spark images at a later stage, which has some practical value.

Sensing and Automation Technologies for Ornamental Nursery Crop Production: Current Status and Future Prospects

The ornamental crop industry is an important contributor to the economy in the United States. The industry has been facing challenges due to continuously increasing labor and agricultural input costs. Sensing and automation technologies have been introduced to reduce labor requirements and to ensure efficient management operations. This article reviews current sensing and automation technologies used for ornamental nursery crop production and highlights prospective technologies that can be applied for future applications. Applications of sensors, computer vision, artificial intelligence (AI), machine learning (ML), Internet-of-Things (IoT), and robotic technologies are reviewed. Some advanced technologies, including 3D cameras, enhanced deep learning models, edge computing, radio-frequency identification (RFID), and integrated robotics used for other cropping systems, are also discussed as potential prospects. This review concludes that advanced sensing, AI and robotic technologies are critically needed for the nursery crop industry. Adapting these current and future innovative technologies will benefit growers working towards sustainable ornamental nursery crop production.

Precision detection of crop diseases based on improved YOLOv5 model

Accurate identification of crop diseases can effectively improve crop yield. Most current crop diseases present small targets, dense numbers, occlusions and similar appearance of different diseases, and the current target detection algorithms are not effective in identifying similar crop diseases. Therefore, in this paper, an improved model based on YOLOv5s was proposed to improve the detection of crop diseases. First, the CSP structure of the original model in the feature fusion stage was improved, and a lightweight structure was used in the improved CSP structure to reduce the model parameters, while the feature information of different layers was extracted in the form of multiple branches. A structure named CAM was proposed, which can extract global and local features of each network layer separately, and the CAM structure can better fuse semantic and scale inconsistent features to enhance the extraction of global information of the network. In order to increase the number of positive samples in the model training process, one more grid was added to the original model with three grids to predict the target, and the formula for the prediction frame centroid offset was modified to obtain the better prediction frame centroid offset when the target centroid falled on the special point of the grid. To solve the problem of the prediction frame being scaled incorrectly during model training, an improved DIoU loss function was used to replace the GIoU loss function used in the original YOLOv5s. Finally, the improved model was trained using transfer learning, the results showed that the improved model had the best mean average precision (mAP) performance compared to the Faster R-CNN, SSD, YOLOv3, YOLOv4, YOLOv4-tiny, and YOLOv5s models, and the mAP, F1 score, and recall of the improved model were 95.92%, 0.91, and 87.89%, respectively. Compared with YOLOv5s, they improved by 4.58%, 5%, and 4.78%, respectively. The detection speed of the improved model was 40.01 FPS, which can meet the requirement of real-time detection. The results showed that the improved model outperformed the original model in several aspects, had stronger robustness and higher accuracy, and can provide better detection for crop diseases.

Fast and Non-Destructive Quail Egg Freshness Assessment Using a Thermal Camera and Deep Learning-Based Air Cell Detection Algorithms for the Revalidation of the Expiration Date of Eggs

Freshness is one of the most important parameters for assessing the quality of avian eggs. Available techniques to estimate the degradation of albumen and enlargement of the air cell are either destructive or not suitable for high-throughput applications. The aim of this research was to introduce a new approach to evaluate the air cell of quail eggs for freshness assessment as a fast, noninvasive, and nondestructive method. A new methodology was proposed by using a thermal microcamera and deep learning object detection algorithms. To evaluate the new method, we stored 174 quail eggs and collected thermal images 30, 50, and 60 days after the labeled expiration date. These data, 522 in total, were expanded to 3610 by image augmentation techniques and then split into training and validation samples to produce models of the deep learning algorithms, referred to as "You Only Look Once" version 4 and 5 (YOLOv4 and YOLOv5) and EfficientDet. We tested the models in a new dataset composed of 60 eggs that were kept for 15 days after the labeled expiration label date. The validation of our methodology was performed by measuring the air cell area highlighted in the thermal images at the pixel level; thus, we compared the difference in the weight of eggs between the first day of storage and after 10 days under accelerated aging conditions. The statistical significance showed that the two variables (air cell and weight) were negatively correlated (R² = 0.676). The deep learning models could predict freshness with F1 scores of 0.69, 0.89, and 0.86 for the YOLOv4, YOLOv5, and EfficientDet models, respectively. The new methodology for freshness assessment demonstrated that the best model reclassified 48.33% of our testing dataset. Therefore, those expired eggs could have their expiration date extended for another 2 weeks from the original label date.

Comparison of Deep Learning Methods for Detecting and Counting Sorghum Heads in UAV Imagery

With the rapid development of remote sensing with small, lightweight unmanned aerial vehicles (UAV), efficient and accurate crop spike counting, and yield estimation methods based on deep learning (DL) methods have begun to emerge, greatly reducing labor costs and enabling fast and accurate counting of sorghum spikes. However, there has not been a systematic, comprehensive evaluation of their applicability in cereal crop spike identification in UAV images, especially in sorghum head counting. To this end, this paper conducts a comparative study of the performance of three common DL algorithms, EfficientDet, Single Shot MultiBox Detector (SSD), and You Only Look Once (YOLOv4), for sorghum head detection based on lightweight UAV remote sensing data. The paper explores the effects of overlap ratio, confidence, and intersection over union (IoU) parameters, using the evaluation metrics of precision P, recall R, average precision AP, F1 score, computational efficiency, and the number of detected positive/negative samples (Objects detected consistent/inconsistent with real samples). The experiment results show the following. (1) The detection results of the three methods under dense coverage conditions were better than those under medium and sparse conditions. YOLOv4 had the most accurate detection under different coverage conditions; on the contrary, EfficientDet was the worst. While SSD obtained better detection results under dense conditions, the number of over-detections was larger. (2) It was concluded that although EfficientDet had a good positive sample detection rate, it detected the fewest samples, had the smallest R and F1, and its actual precision was poor, while its training time, although medium, had the lowest detection efficiency, and the detection time per image was 2.82-times that of SSD. SSD had medium values for P, AP, and the number of detected samples, but had the highest training and detection efficiency. YOLOv4 detected the largest number of positive samples, and its values for R, AP, and F1 were the highest among the three methods. Although the training time was the slowest, the detection efficiency was better than EfficientDet. (3) With an increase in the overlap ratios, both positive and negative samples tended to increase, and when the threshold value was 0.3, all three methods had better detection results. With an increase in the confidence value, the number of positive and negative samples significantly decreased, and when the threshold value was 0.3, it balanced the numbers for sample detection and detection accuracy. An increase in IoU was accompanied by a gradual decrease in the number of positive samples and a gradual increase in the number of negative samples. When the threshold value was 0.3, better detection was achieved. The research findings can provide a methodological basis for accurately detecting and counting sorghum heads using UAV.

Detection of Dense Citrus Fruits by Combining Coordinated Attention and Cross-Scale Connection with Weighted Feature Fusion

The accuracy detection of individual citrus fruits in a citrus orchard environments is one of the key steps in realizing precision agriculture applications such as yield estimation, fruit thinning, and mechanical harvesting. This study proposes an improved object detection YOLOv5 model to achieve accurate the identification and counting of citrus fruits in an orchard environment. First, the latest visual attention mechanism coordinated attention module (CA) was inserted into an improved backbone network to focus on fruit-dense regions to recognize small target fruits. Second, an efficient two-way cross-scale connection and weighted feature fusion BiFPN in the neck network were used to replace the PANet multiscale feature fusion network, giving effective feature corresponding weights to fully fuse the high-level and bottom-level features. Finally, the varifocal loss function was used to calculate the model loss for better model training results. The results of the experiments on four varieties of citrus trees showed that our improved model proposed to this study could effectively identify dense small citrus fruits. Specifically, the recognized AP (average precision) reached 98.4%, and the average recognition time was 0.019 s per image. Compared with the original YOLOv5 (including deferent variants of n, s, m, l, and x), the increase in the average accuracy precision of the improved YOLOv5 ranged from 7.5% to 0.8% while maintaining similar average inference time. Four different citrus varieties were also tested to evaluate the generalization performance of the improved model. The method can be further used as a part in a vision system to provide technical support for the real-time and accurate detection of multiple fruit targets during mechanical picking in citrus orchards.

Application of Convolutional Neural Network-Based Detection Methods in Fresh Fruit Production: A Comprehensive Review

As one of the representative algorithms of deep learning, a convolutional neural network (CNN) with the advantage of local perception and parameter sharing has been rapidly developed. CNN-based detection technology has been widely used in computer vision, natural language processing, and other fields. Fresh fruit production is an important socioeconomic activity, where CNN-based deep learning detection technology has been successfully applied to its important links. To the best of our knowledge, this review is the first on the whole production process of fresh fruit. We first introduced the network architecture and implementation principle of CNN and described the training process of a CNN-based deep learning model in detail. A large number of articles were investigated, which have made breakthroughs in response to challenges using CNN-based deep learning detection technology in important links of fresh fruit production including fruit flower detection, fruit detection, fruit harvesting, and fruit grading. Object detection based on CNN deep learning was elaborated from data acquisition to model training, and different detection methods based on CNN deep learning were compared in each link of the fresh fruit production. The investigation results of this review show that improved CNN deep learning models can give full play to detection potential by combining with the characteristics of each link of fruit production. The investigation results also imply that CNN-based detection may penetrate the challenges created by environmental issues, new area exploration, and multiple task execution of fresh fruit production in the future.

Object Detection Algorithm for Wheeled Mobile Robot Based on an Improved YOLOv4

In practical applications, the intelligence of wheeled mobile robots is the trend of future development. Object detection for wheeled mobile robots requires not only the recognition of complex surroundings, but also the deployment of algorithms on resource-limited devices. However, the current state of basic vision technology is insufficient to meet demand. Based on this practical problem, in order to balance detection accuracy and detection efficiency, we propose an object detection algorithm based on a combination of improved YOLOv4 and improved GhostNet in this paper. Firstly, the backbone feature extraction network of original YOLOv4 is replaced with the trimmed GhostNet network. Secondly, enhanced feature extraction network in the YOLOv4, ordinary convolution is supplanted with a combination of depth-separable and ordinary convolution. Finally, the hyperparameter optimization was carried out. The experimental results show that the improved YOLOv4 network proposed in this paper has better object detection performance. Specifically, the precision, recall, F1, mAP (0.5) values, and mAP (0.75) values are 88.89%, 87.12%, 88.00%, 86.84%, and 50.91%, respectively. Although the mAP (0.5) value is only 2.23% less than the original YOLOv4, it is higher than YOLOv4_tiny, Eifficientdet-d0, YOLOv5n, and YOLOv5 compared to 29.34%, 28.99%, 20.36%, and 18.64%, respectively. In addition, it outperformed YOLOv4 in terms of mAP (0.75) value and precision, and its model size is only 42.5 MB, a reduction of 82.58% when compared to YOLOv4’s model size.

Weed Identification by Single-Stage and Two-Stage Neural Networks: A Study on the Impact of Image Resizers and Weights Optimization Algorithms

The accurate identification of weeds is an essential step for a site-specific weed management system. In recent years, deep learning (DL) has got rapid advancements to perform complex agricultural tasks. The previous studies emphasized the evaluation of advanced training techniques or modifying the well-known DL models to improve the overall accuracy. In contrast, this research attempted to improve the mean average precision (mAP) for the detection and classification of eight classes of weeds by proposing a novel DL-based methodology. First, a comprehensive analysis of single-stage and two-stage neural networks including Single-shot MultiBox Detector (SSD), You look only Once (YOLO-v4), EfficientDet, CenterNet, RetinaNet, Faster Region-based Convolutional Neural Network (RCNN), and Region-based Fully Convolutional Network (RFCN), has been performed. Next, the effects of image resizing techniques along with four image interpolation methods have been studied. It led to the final stage of the research through optimization of the weights of the best-acquired model by initialization techniques, batch normalization, and DL optimization algorithms. The effectiveness of the proposed work is proven due to a high mAP of 93.44% and validated by the stratified k-fold cross-validation technique. It was 5.8% improved as compared to the results obtained by the default settings of the best-suited DL architecture (Faster RCNN ResNet-101). The presented pipeline would be a baseline study for the research community to explore several tasks such as real-time detection and reducing the computation/training time. All the relevant data including the annotated dataset, configuration files, and inference graph of the final model are provided with this article. Furthermore, the selection of the DeepWeeds dataset shows the robustness/practicality of the study because it contains images collected in a real/complex agricultural environment. Therefore, this research would be a considerable step toward an efficient and automatic weed control system.

Study on Pear Flowers Detection Performance of YOLO-PEFL Model Trained With Synthetic Target Images

Accurate detection of pear flowers is an important measure for pear orchard yield estimation, which plays a vital role in improving pear yield and predicting pear price trends. This study proposed an improved YOLOv4 model called YOLO-PEFL model for accurate pear flower detection in the natural environment. Pear flower targets were artificially synthesized with pear flower's surface features. The synthetic pear flower targets and the backgrounds of the original pear flower images were used as the inputs of the YOLO-PEFL model. ShuffleNetv2 embedded by the SENet (Squeeze-and-Excitation Networks) module replacing the original backbone network of the YOLOv4 model formed the backbone of the YOLO-PEFL model. The parameters of the YOLO-PEFL model were fine-tuned to change the size of the initial anchor frame. The experimental results showed that the average precision of the YOLO-PEFL model was 96.71%, the model size was reduced by about 80%, and the average detection speed was 0.027s. Compared with the YOLOv4 model and the YOLOv4-tiny model, the YOLO-PEFL model had better performance in model size, detection accuracy, and detection speed, which effectively reduced the model deployment cost and improved the model efficiency. It implied the proposed YOLO-PEFL model could accurately detect pear flowers with high efficiency in the natural environment.

Real Time Pear Fruit Detection and Counting Using YOLOv4 Models and Deep SORT

This study aimed to produce a robust real-time pear fruit counter for mobile applications using only RGB data, the variants of the state-of-the-art object detection model YOLOv4, and the multiple object-tracking algorithm Deep SORT. This study also provided a systematic and pragmatic methodology for choosing the most suitable model for a desired application in agricultural sciences. In terms of accuracy, YOLOv4-CSP was observed as the optimal model, with an AP@0.50 of 98%. In terms of speed and computational cost, YOLOv4-tiny was found to be the ideal model, with a speed of more than 50 FPS and FLOPS of 6.8-14.5. If considering the balance in terms of accuracy, speed and computational cost, YOLOv4 was found to be most suitable and had the highest accuracy metrics while satisfying a real time speed of greater than or equal to 24 FPS. Between the two methods of counting with Deep SORT, the unique ID method was found to be more reliable, with an F1count of 87.85%. This was because YOLOv4 had a very low false negative in detecting pear fruits. The ROI line is more reliable because of its more restrictive nature, but due to flickering in detection it was not able to count some pears despite their being detected.