An Apple Detection Method Based on Des-YOLO v4 Algorithm for Harvesting Robots in Complex Environment

Abstract

Real-time detection of apples in natural environment is a necessary condition for robots to pick apples automatically, and it is also a key technique for orchard yield prediction and fine management. To make the harvesting robots detect apples quickly and accurately in complex environment, a Des-YOLO v4 algorithm and a detection method of apples are proposed. Compared with the current mainstream detection algorithms, YOLO v4 has better detection performance. However, the complex network structure of YOLO v4 will reduce the picking efficiency of the robot. Therefore, a Des-YOLO structure is proposed, which reduces network parameters and improves the detection speed of the algorithm. In the training phase, the imbalance of positive and negative samples will cause false detection of apples. To solve the above problem, a class loss function based on AP-Loss (Average Precision Loss) is proposed to improve the accuracy of apple recognition. Traditional YOLO algorithm uses NMS (Nonmaximum Suppression) method to filter the prediction boxes, but NMS cannot detect the adjacent apples when they overlap each other. Therefore, Soft-NMS is used instead of NMS to solve the problem of missing detection, so as to improve the generalization of the algorithm. The proposed algorithm is tested on the self-made apple image data set. The results show that Des-YOLO v4 network has ideal features with a mAP (mean Average Precision) of apple detection of 97.13%, a recall rate of 90%, and a detection speed of 51 f/s. Compared with traditional network models such as YOLO v4 and Faster R-CNN, the Des-YOLO v4 can meet the accuracy and speed requirements of apple detection at the same time. Finally, the self-designed apple-harvesting robot is used to carry out the harvesting experiment. The experiment shows that the harvesting time is 8.7 seconds and the successful harvesting rate of the robot is 92.9%. Therefore, the proposed apple detection method has the advantages of higher recognition accuracy and faster recognition speed. It can provide new solutions for apple-harvesting robots and new ideas for smart agriculture.

Precision detection of crop diseases based on improved YOLOv5 model

“…A class loss function based on AP-Loss (Average Precision loss) is proposed to deal with the positive and negative sample imbalance problem during training. The final experimental results show that the mAP of the improved algorithm is 97.13% ( Chen et al., 2021 ). An improved YOLOv5 model was proposed to improve the detection of kiwi defects by adding a small target layer, introducing SE attention mechanism and CIoU loss function, and the mAP of the improved model was improved by nearly 9% compared with the original model ( Yao et al., 2021 ).…”

“…A multi-scale parallel algorithm MP-YOLOv3 was proposed to improve the detection of tomato gray mold based on the MobileNetv2-YOLOv3 model (Wang and Liu, 2021a), and the experimental results showed that the improved model has strong robustness in real natural environments. An algorithm based on super-resolution image enhancement (Zhu et al, 2021) and an algorithm combining Inception and an improved Softmax classifier are proposed to detect grape and apple diseases, respectively (Li et al, 2022). Enhancing feature extraction by incorporating DenseNet interlayer density in the YOLOv4 model (Gai et al, 2021).…”

CSFN-YOLOv5s: A Rapid Apple Detection Algorithm in the Natural Environment

“…Sun et al [20] constructed a new apple leaf disease detection model called MEAN-SSD, which achieved 83.12 % mAP detection performance and operated at a detection speed of 12.53 FPS. The complex network structure of YOLOv4 will reduce the efficiency of the apple picking robot, Chen et al [21] proposed a Des-YOLO structure to reduce network parameters and improve the detection speed of the algorithm. Wang et al [22] made lightweight improvements to YOLOv5s model channel pruning, the improved model has an accuracy of 95.8% and a size of only 1.4 MB.…”

Video based oil palm ripeness detection model using deep learning

“…YOLOv4 is state of the art in object detection which is the result of the development of YOLOv3 research [ 18 ] and has been tested in agriculture with satisfactory results such as automating pear counting [ 12 ], apple detection [ 19 ] and citrus fruit ripeness detection [ 20 ]. YOLOv4 employs a DarkNet framework based on the low-level programming language to do fast computing, which is ideal for implementing conditions in real time.…”