Weeds are one of the main factors affecting the yield and quality of agricultural products. Accurate evaluation of weed density is of great significance for field management, especially precision weeding. In this paper, a weed density calculating and mapping method in the field is proposed. An unmanned aerial vehicle (UAV) was used to capture field images. The excess green minus excess red index, combined with the minimum error threshold segmentation method, was used to segment green plants and bare land. A modified U-net was used to segment crops from images. After removing the bare land and crops from the field, images of weeds were obtained. The weed density was evaluated by the ratio of weed area to total area on the segmented image. The accuracy of the green plant segmentation was 93.5%. In terms of crop segmentation, the intersection over union (IoU) was 93.40%, and the segmentation time of a single image was 35.90 ms. Finally, the determination coefficient of the UAV evaluated weed density and the manually observed weed density was 0.94, and the root mean square error was 0.03. With the proposed method, the weed density of a field can be effectively evaluated from UAV images, hence providing critical information for precision weeding.

Apples are one of the most kind of important fruit in the world. China has been the largest apple producing country. Yield estimating, robot harvesting, precise spraying are important processes for precise planting apples. Image segmentation is an important step in machine vision systems for precision apple planting. In this paper, an apple fruit segmentation algorithm applied in the orchard was studied. The effect of many color features in classifying apple fruit pixels from other pixels was evaluated. Three color features were selected. This color features could effectively distinguish the apple fruit pixels from other pixels. The GLCM (Grey-Level Co-occurrence Matrix) was used to extract texture features. The best distance and orientation parameters for GLCM were found. Nine machine learning algorithms had been used to develop pixel classifiers. The classifier was trained with 100 pixels and tested with 100 pixels. The accuracy of the classifier based on Random Forest reached 0.94. One hundred images of an apple orchard were artificially labeled with apple fruit pixels and other pixels. At the same time, a classifier was used to segment these images. Regression analysis was performed on the results of artificial labeling and classifier classification. The average values of Af (segmentation error), FPR (false positive rate) and FNR (false negative rate) were 0.07, 0.13 and 0.15, respectively. This result showed that this algorithm could segment apple fruit in orchard images effectively. It could provide a reference for precise apple planting management.

The segmentation of broccoli seedlings in the crops and weeds co-exist field environment is of great significance for weeding and herbicide spraying. This paper constructed a crop segmentation algorithm with a small training set for discriminating broccoli seedlings from weeds and soil. This algorithm was based on a support vector machine (SVM) combined with color-texture features. Correlation analysis and chi-square tests were used to select 6 features from the 21 color features. Gray-level co-occurrence matrix (GLCM) was used to extract 5 texture features. And each parameter of GLCM had been assessed and optimized by the chi-square test. Linear Discriminant Analysis (LDA) was used to decompose the original dataset in a set of 3 successive orthogonal components. This method selected features more reasonable and gained higher plant segmentation accuracy. When the training sample is greater than 50, the accuracy of the test set could reach 90%. The coefficient of determination (R2) between the ground truth broccoli seedling area and the segmentation broccoli area was 0.91, and the root-mean-square error (σ) was 0.10. Results demonstrated that the color-texture features were able to effectively segment broccoli seedlings even when there was a significant amount of weeds.

Vision-based smart agriculture is an important way to improve the efficiency of agricultural production. Labeling images for deep learning in complex field photos is a difficult task. In this paper, an image augmentation method that can help reduce the workload of image labeling based on synthetic images is proposed. The synthetic images consisted of three parts: crop, weed, and soil. The crop and weeds were obtained automatically by Excess Green (ExG) and minimum error threshold segmentation. The data augmentation method was tested on image classification, objection detection, and semantic segmentation tasks by Resnet, YOLOV5, and DeeplabV3. The accuracy of the classification model reached 0.99. The IoU of objection detection and semantic segmentation were 0.98 and 0.96, respectively. The results showed that the method in this paper was acceptable despite slight overfitting. This method was proposed based on the characteristics of field images, and it was meant for reducing the workload of labeling images.

Efficient and intelligent surface defect detection of photovoltaic modules is crucial for improving the quality of photovoltaic modules and ensuring the reliable operation of large-scale infrastructure. However, the scenario characteristics of data distribution deviation make the construction of defect detection models for open world scenarios such as photovoltaic manufacturing and power plant inspections a challenge. Therefore, we propose the Gather and Distribute Domain shift Suppression Network. It adopts a single domain generalized method that is completely independent of the test samples to address the problem of distribution shift. Using a one-stage network as the baseline network breaks through the limitations of traditional domain generalization methods that typically use two-stage networks. It not only balances detection accuracy and speed but also simplifies the model deployment and application process. The network first employs the DeepSpine module to capture a wider range of contextual information. By concatenating and aligning multi-scale channel features, it effectively suppresses background style shifts. Building upon this, the Gather and Distribute Module performs cross layer interactive learning on multi-scale channel features. The multi-level features and semantic dependencies learned enhance the localization and recognition ability of target defects, thereby achieving the suppression of defect instance shift. Furthermore, we utilizes normalized Wasserstein distance for similarity measurement, reducing measurement errors caused by bounding box position deviations. We conducted a comprehensive evaluation of our network on the Electroluminescence Endogenous Shift Dataset and Photovoltaic Inspection Infrared Dataset. In scenarios with three production lines and four heights on two datasets, the detection accuracy of GDDS reached 91.2%, 82.3%, 79.9%, and 92.8%, 82.7%, 77.2%, and 69.2%, respectively. The experimental results showed that our method can adapt to defect detection in open world scenarios faster and better than other state-of-the-art methods.