Multi-Resolution Weed Classification via Convolutional Neural Network and Superpixel Based Local Binary Pattern Using Remote Sensing Images

Assessment of Invasive and Weed Species by Hyperspectral Imagery in Agrocenoses Ecosystem

The present study aimed to investigate the possibility of using hyperspectral imaging data to identify the invasive and weed species in agrocenoses ecosystem. The most common weeds in grain agrocenoses, i.e., Ambrosia artemisiifolia L., Euphorbia seguieriana Neck., Atriplex tatarica L., Glycyrrhiza glabra L., Setaria pumila (Poir.) Roem. and Schult, served as objects. The population of weeds, especially Ambrosia artemisiifolia is invasive for the selected region of study. Therefore, the shooting of objects was carried out with a hyperspectral camera, Cubert UHD185, and the values of 100 spectral channels were obtained from hyperspectral images. The values of 80 vegetation indices (VIs) were calculated. The material was processed using mathematical statistics (analysis of variance, t-test) and search methods of data analysis (principal component analysis, decision tree, and random forest). Using statistical methods, the simultaneous use of several VIs differentiated between species more deliberately and precisely. The combination of VIs Derivative index (D1), Chlorophyll content index (Datt3), and Pigment specific normalized difference (PSND) can be used for weeds identification. Using the decision tree method, VIs established a good division of weeds into groups; (1) perennial rhizomatous weeds (Euphorbia seguieriana, and Glycyrrhiza glabra), and (2) annual weeds (A. artemisiifolia, A. tatarica, and S. pumila); These Vis are Chlorophyll index (CI), D1, and Datt3. Using the random forest method, the VIs that have the greatest impact on Mean Decrease Accuracy and Mean Decrease Gini are D1, Datt3, PSND, and Double Peak Index (DPI). The use of spectral channel values for the identification of plant species using the principal component analysis, decision tree, and random forest methods showed worse results than when using VIs. A great similarity of the results was obtained with the help of statistical and search methods of data analysis.

Weed Classification from Natural Corn Field-Multi-Plant Images Based on Shallow and Deep Learning

Crop and weed discrimination in natural field environments is still challenging for implementing automatic agricultural practices, such as weed control. Some weed control methods have been proposed. However, these methods are still restricted as they are implemented under controlled conditions. The development of a sound weed control system begins by recognizing the crop and the different weed plants presented in the field. In this work, a classification approach of Zea mays L. (Crop), narrow-leaf weeds (NLW), and broadleaf weeds (BLW) from multi-plant images are presented. Moreover, a large image dataset was generated. Images were captured in natural field conditions, in different locations, and growing stages of the plants. The extraction of regions of interest (ROI) is carried out employing connected component analysis (CCA), whereas the classification of ROIs is based on Convolutional Neural Networks (CNN) and compared with a shallow learning approach. To measure the classification performance of both methods, accuracy, precision, recall, and F1-score metrics were used. The best alternative for the weed classification task at early stages of growth and in natural corn field environments was the CNN-based approach, as indicated by the 97% accuracy value obtained.

Superpixel-Based Regional-Scale Grassland Community Classification Using Genetic Programming with Sentinel-1 SAR and Sentinel-2 Multispectral Images

Grasslands are one of the most important terrestrial ecosystems on the planet and have significant economic and ecological value. Accurate and rapid discrimination of grassland communities is critical to the conservation and utilization of grassland resources. Previous studies that explored grassland communities were mainly based on field surveys or airborne hyperspectral and high-resolution imagery. Limited by workload and cost, these methods are typically suitable for small areas. Spaceborne mid-resolution RS images (e.g., Sentinel, Landsat) have been widely used for large-scale vegetation observations owing to their large swath width. However, there still keep challenges in accurately distinguishing between different grassland communities using these images because of the strong spectral similarity of different communities and the suboptimal performance of models used for classification. To address this issue, this paper proposed a superpixel-based grassland community classification method using Genetic Programming (GP)-optimized classification model with Sentinel-2 multispectral bands, their derived vegetation indices (VIs) and textural features, and Sentinel-1 Synthetic Aperture Radar (SAR) bands and the derived textural features. The proposed method was evaluated in the Siziwang grassland of China. Our results showed that the addition of VIs and textures, as well as the use of GP-optimized classification models, can significantly contribute to distinguishing grassland communities, and the proposed approach classified the seven communities in Siziwang grassland with an overall accuracy of 84.21% and a kappa coefficient of 0.81. We concluded that the classification method proposed in this paper is capable of distinguishing grassland communities with high accuracy at a regional scale.

Hyperspectral Image Classification Based on Superpixel Pooling Convolutional Neural Network with Transfer Learning

Deep learning-based hyperspectral image (HSI) classification has attracted more and more attention because of its excellent classification ability. Generally, the outstanding performance of these methods mainly depends on a large number of labeled samples. Therefore, it still remains an ongoing challenge how to integrate spatial structure information into these frameworks to classify the HSI with limited training samples. In this study, an effective spectral-spatial HSI classification scheme is proposed based on superpixel pooling convolutional neural network with transfer learning (SP-CNN). The suggested method includes three stages. The first part consists of convolution and pooling operation, which is a down-sampling process to extract the main spectral features of an HSI. The second part is composed of up-sampling and superpixel (homogeneous regions with adaptive shape and size) pooling to explore the spatial structure information of an HSI. Finally, the hyperspectral data with each superpixel as a basic input rather than a pixel are fed to fully connected neural network. In this method, the spectral and spatial information is effectively fused by using superpixel pooling technique. The use of popular transfer learning technology in the proposed classification framework significantly improves the training efficiency of SP-CNN. To evaluate the effectiveness of the SP-CNN, extensive experiments were conducted on three common real HSI datasets acquired from different sensors. With 30 labeled pixels per class, the overall classification accuracy provided by this method on three benchmarks all exceeded 93%, which was at least 4.55% higher than that of several state-of-the-art approaches. Experimental and comparative results prove that the proposed algorithm can effectively classify the HSI with limited training labels.

Identification of Weeds Based on Hyperspectral Imaging and Machine Learning

Weeds can be major environmental and economic burdens in New Zealand. Traditional methods of weed control including manual and chemical approaches can be time consuming and costly. Some chemical herbicides may have negative environmental and human health impacts. One of the proposed important steps for providing alternatives to these traditional approaches is the automated identification and mapping of weeds. We used hyperspectral imaging data and machine learning to explore the possibility of fast, accurate and automated discrimination of weeds in pastures where ryegrass and clovers are the sown species. Hyperspectral images from two grasses (Setaria pumila [yellow bristle grass] and Stipa arundinacea [wind grass]) and two broad leaf weed species (Ranunculus acris [giant buttercup] and Cirsium arvense [Californian thistle]) were acquired and pre-processed using the standard normal variate method. We trained three classification models, namely partial least squares-discriminant analysis, support vector machine, and Multilayer Perceptron (MLP) using whole plant averaged (Av) spectra and superpixels (Sp) averaged spectra from each weed sample. All three classification models showed repeatable identification of four weeds using both Av and Sp spectra with a range of overall accuracy of 70-100%. However, MLP based on the Sp method produced the most reliable and robust prediction result (89.1% accuracy). Four significant spectral regions were found as highly informative for characterizing the four weed species and could form the basis for a rapid and efficient methodology for identifying weeds in ryegrass/clover pastures.