Search

• YOLOv8-GP based on key point detection is proposed. • C2f in the backbone network is replaced by C2f-Faster-EMA. • Multi-scale feature fusion is achieved by using BiFPN in Neck network. • While reducing the calculation, the AP of grape is improved by 3.3 %. • The positioning error of the picking point is kept within 30 pixels. Precise positioning of fruit and picking point is crucial for harvesting table grapes using automated picking robots in an unstructured agricultural environment. Most studies employ multi-step methods for locating picking points based on fruit detection, leading to slow detection speed, cumbersome models, and algorithmic fragmentation. This study proposes an improved YOLOv8-GP (YOLOv8-Grape and picking point) model based on YOLOv8n-Pose to solve the problem of simultaneous detection of grape clusters and picking points. YOLOv8-GP is an end-to-end network that integrates object detection and key point detection. Specifically, the Bottleneck in C2f is replaced with FasterNet Block that incorporates EMA (Efficient Multi-Scale Attention), resulting in C2f-Faster-EMA. BiFPN is applied to substitute the original PAN as Neck network. The FasterNet Block, designed based on partial convolution (PConv), reduces redundant computation and memory access, thereby extracting spatial features more efficiently. The EMA attention mechanism achieves performance gains with lower computational overhead. Furthermore, BiFPN is employed to enhance the effect of feature fusion. Experimental results demonstrate that YOLOv8-GP achieves AP of 89.7 % for grape cluster detection and a Euclidean distance error of less than 30 pixels for picking point detection. Additionally, the number of Params is reduced by 47.73 %, and the model complexity GFlops is 6.1G. In summary, YOLOv8-GP offers excellent detection performance, while the reduced number of parameters and model complexity contribute to lower deployment costs and easier implementation on mobile robots. [ABSTRACT FROM AUTHOR]

• A detection model of sesame and weeds based on YOLOv4, YOLO-sesame, is proposed. • Using local importance pooling to add attention mechanism to SPP structure. • Use SE block to improve the logic module of local importance pooling. • The ASFF structure is integrated to solve the problem of missing detection. • Effectively improve the detection accuracy while maintaining a fast detection speed. Weeds have a significant impact on sesame throughout its early stages of development, thus they must be rigorously controlled. However, the shape of sesame seedlings and weeds are similar, and the size specifications are not defined, making reliable weed detection difficult. To achieve the goal of weed recognition, the majority of solutions now use a deep learning model to learn the weed image. Weed targets with big variances in size and specification are easy to overlook with the current deep learning algorithm. As a result, standard deep learning models have room for improvement when it comes to sesame and weed recognition rates. The YOLO-sesame model is proposed to improve the efficiency and accuracy of sesame weed identification. Based on the YOLOv4 model, an attention mechanism is introduced. Local importance pooling is added to the SPP layer, on which the SE module is used as a logical module. To address the issue of large differences in target size and specifications, an adaptive spatial feature fusion structure is included at the feature fusion level. The experimental results suggest that the YOLO-sesame model proposed in this study outperforms mainstream models such as Fast R-CNN, SSD, YOLOv3, YOLOv4, and YOLOv4-tiny in terms of detection performance. Sesame crops and weeds received F1 scores of 0.91 and 0.92, respectively, while the mAP was 96.16%. The detecting frame rate was 36.8 per second. In conclusion, the YOLO-sesame model successfully meets the needs for accurate sesame weed detection. [ABSTRACT FROM AUTHOR]

• An image restoration method based on convolutional autoencoder is proposed. • Use the area growth method to divide and code the obstruction according to the contour. • Predict the overall outline of citrus based on the unobstructed citrus outline. • Fill the area to be repaired with the selection method of similar pixels. • The repair rate and image recognition rate have been effectively improved. At present, machine vision and deep learning theories have been widely used in fruit recognition and picking. However, in the process of identification and picking, there are often situations where the target is blocked, existing methods cannot accurately identify, or the identification accuracy rate is low. For improving the recognition rate of fruits under occlusion, this paper proposes an image restoration method based on a convolutional auto-encoder. This method first encodes the obstruction and then determines the general shape of the occluded fruit and compares the general shape with the coding part is combined to determine the area to be repaired. Finally, the area to be repaired is filled with pixels to realize the image repair and recognition of the occlusion map. The average repair rate of the method proposed in this paper is 95.96%, the restoration rate is 3.52% higher than the traditional convolution method, the L2 loss value is 0.63% lower than the traditional convolution method, the average detection accuracy of the restored fruits is 94.77%. [ABSTRACT FROM AUTHOR]

• A multi-scale feature extraction module (MFE) is designed by using multi-layer dilated convolution to obtain the multi-scale feature information of pest image with a large receptive field without increasing the amount of model parameters. • A deep feature extraction module (DFE) is proposed. Firstly, the deep feature information of pest image is obtained by multiple convolution operations, and then the obtained deep feature image is restored to the size of the original image by deconvolution. At the same time, the deep and shallow feature information are fused by multiple skip-connection structure (Zhang et al., 2019) to reduce the loss of feature information. • We propose a network model based on multi-scale feature fusion to accurately identify and classify crop pests. Experiments on 12 types of pest datasets show that our method has better performance in pest identification and classification than other advanced methods, its classification accuracy (ACC) reaches 98.2%, and the model training time is only 197 min. Crop diseases and insect pests are a serious natural disaster, which needs to be predicted and monitored in time to ensure the output of crops. Due to the wide variety of pests and the similar morphology of crops in the early stages of growth, it is difficult for agricultural workers to accurately identify various types of pests. Crop insects have brought huge challenges to the prevention and control of plant diseases and insect pests. In response to this problem, we propose a way of classification of crop pests based on multi-scale feature fusion(MFFNet) to accurately recognizes and classifies crop pests. First, the multi-scale feature extraction module (MFE) is designed by using dilated convolution to obtain the multi-scale feature map of the pest image. At the same time, extracted the deep feature information of the image by the feature extraction module (DFE). Finally, the features extracted separately by the multi-scale feature extraction module (MFE) and the feature extraction module (DFE) were fused thus achieving accurately classified and identified the crops insects by the way of end-to-end. Experiments show that our proposed method has obtained excellent classification performance on the dataset of 12 types of pests, its classification accuracy rate (ACC) reached 98.2%. [ABSTRACT FROM AUTHOR]

• An algorithm for detecting ripe fruit under natural illumination and occlusion conditions is proposed. • A graying factor K is proposed to remove the complex background of colorful ripe fruits. • Improving the Hough Transform by selecting equidistant points and increasing judgement threshold. • A real contour Hough Transform was proposed to detect ripe fruits or occluded circles. For an accurate detection of ripe fruits under uneven illumination and ubiquitous occlusion conditions, this paper proposes a method to detect and locate ripe fruits based on machine vision. There are four key steps in this method including image graying and background removal, binary image optimization, true contour fragments extraction, and fruit fitting. For testing the proposed method, field experiments were conducted with tomato and citrus, and the ripe fruits in complex environments were successfully detected and located. From the detection experiments, it showed that the recognition rate for ripe fruits in the near zone of the proposed method was higher than 97.44%, the average time consumption was 0.2966 s, and the positioning error was less than 4.41%. In addition, it can be concluded from the comparative experiment that the proposed method is superior to conventional Hough transform, random Hough transform, and other methods based on deep learning in terms of detection rate, time performance and positioning accuracy. Therefore, it can be applied to picking robots for real-time detecting and locating ripe fruits. [ABSTRACT FROM AUTHOR]