Image processing techniques to estimate weight and morphological parameters for selected wheat refractions

Automatic Measurement of Seed Geometric Parameters Using a Handheld Scanner

Seed geometric parameters are important in yielding trait scorers, quantitative trait loci, and species recognition and classification. A novel method for automatic measurement of three-dimensional seed phenotypes is proposed. First, a handheld three-dimensional (3D) laser scanner is employed to obtain the seed point cloud data in batches. Second, a novel point cloud-based phenotyping method is proposed to obtain a single-seed 3D model and extract 33 phenotypes. It is connected by an automatic pipeline, including single-seed segmentation, pose normalization, point cloud completion by an ellipse fitting method, Poisson surface reconstruction, and automatic trait estimation. Finally, two statistical models (one using 11 size-related phenotypes and the other using 22 shape-related phenotypes) based on the principal component analysis method are built. A total of 3400 samples of eight kinds of seeds with different geometrical shapes are tested. Experiments show: (1) a single-seed 3D model can be automatically obtained with 0.017 mm point cloud completion error; (2) 33 phenotypes can be automatically extracted with high correlation compared with manual measurements (correlation coefficient (R2) above 0.9981 for size-related phenotypes and R2 above 0.8421 for shape-related phenotypes); and (3) two statistical models are successfully built to achieve seed shape description and quantification.

Determination of the melanin and anthocyanin content in barley grains by digital image analysis using machine learning methods

The pigment composition of plant seed coat affects important properties such as resistance to pathogens, pre-harvest sprouting, and mechanical hardness. The dark color of barley (Hordeum vulgare L.) grain can be attributed to the synthesis and accumulation of two groups of pigments. Blue and purple grain color is associated with the biosynthesis of anthocyanins. Gray and black grain color is caused by melanin. These pigments may accumulate in the grain shells both individually and together. Therefore, it is difficult to visually distinguish which pigments are responsible for the dark color of the grain. Chemical methods are used to accurately determine the presence/ absence of pigments; however, they are expensive and labor-intensive. Therefore, the development of a new method for quickly assessing the presence of pigments in the grain would help in investigating the mechanisms of genetic control of the pigment composition of barley grains. In this work, we developed a method for assessing the presence or absence of anthocyanins and melanin in the barley grain shell based on digital image analysis using computer vision and machine learning algorithms. A protocol was developed to obtain digital RGB images of barley grains. Using this protocol, a total of 972 images were acquired for 108 barley accessions. Seed coat from these accessions may contain anthocyanins, melanins, or pigments of both types. Chemical methods were used to accurately determine the pigment content of the grains. Four models based on computer vision techniques and convolutional neural networks of different architectures were developed to predict grain pigment composition from images. The U-Net network model based on the EfficientNetB0 topology showed the best performance in the holdout set (the value of the "accuracy" parameter was 0.821).

Quantitative characterizations of visual fibrousness in meat analogues using automated image analysis

A desirable quality of plant-based meat analogues is to resemble the fibrous structure of cooked muscle meat. While texture analysis can characterize fibrous structures mechanically, assessment of visual fibrous structures remains subjective. Quantitative assessment of visual fibrous structures of meat analogues relies on expert knowledge, is resource-intensive, and time-consuming. In this study, a novel image-based method (Fiberlyzer) is developed to provide automated, quantitative, and standardized assessment of visual fibrousness of meat analogues. The Fiberlyzer method segments fibrous regions from 2D images and extracts fiber shape features to characterize the fibrous structure of meat analogues made from mung bean, soy, and pea protein. The computed fiber scores (the ratio between fiber length and width) demonstrate a strong correlation with expert panel evaluations, particularly on a per-formulation basis (r2  = 0.93). Additionally, the Fiberlyzer method generates fiber shape features including fiber score, fiber area, and the number of fiber branches, facilitating comparisons of structural similarity between meat analogue samples and cooked chicken meat as a benchmark. With a simple measurement setup and user-friendly interface, the Fiberlyzer method can become a standard tool integrated into formulation development, quality control, and production routines of plant-based meat analogue. This method offers rapid, cheap, and standardized quantification of visual fibrousness, minimizing the need for expert knowledge in the process of quality control.

Detection of Green Asparagus Using Improved Mask R-CNN for Automatic Harvesting

Advancements in deep learning and computer vision have led to the discovery of numerous effective solutions to challenging problems in the field of agricultural automation. With the aim to improve the detection precision in the autonomous harvesting process of green asparagus, in this article, we proposed the DA-Mask RCNN model, which utilizes the depth information in the region proposal network. Firstly, the deep residual network and feature pyramid network were combined to form the backbone network. Secondly, the DA-Mask RCNN model added a depth filter to aid the softmax function in anchor classification. Afterwards, the region proposals were further processed by the detection head unit. The training and test images were mainly acquired from different regions in the basin of the Yangtze River. During the capturing process, various weather and illumination conditions were taken into account, including sunny weather, sunny but overshadowed conditions, cloudy weather, and daytime greenhouse conditions as well as nighttime greenhouse conditions. Performance experiments, comparison experiments, and ablation experiments were carried out using the five constructed datasets to verify the effectiveness of the proposed model. Precision, recall, and F1-score values were applied to evaluate the performances of different approaches. The overall experimental results demonstrate that the balance of the precision and speed of the proposed DA-Mask RCNN model outperform those of existing algorithms.