A Detection Model for Cucumber Root-Knot Nematodes Based on Modified YOLOv5-CMS

The development of resistant cucumber varieties is a great importance for reducing the production loss caused by root-knot nematodes. After plants are infected with nematodes, their roots will swell into spherical bumps. Rapid and accurate detection sites assessment disease severity play key role in selecting varieties. Because locations sizes bumps formed after different degrees infection random, currently available counting methods based on manual operation extremely time-consuming labor-intensive, prone to human error. In response these problems, this paper proposes nematode model modified YOLOv5s (i.e., YOLOv5-CMS) order support breeding proposed model, dual attention module (CBAM-CA) was adopted enhance model’s ability extracting features, K-means++ clustering algorithm applied optimize selection initial cluster center, which effectively improved performance, novel bounding box regression function (SIoU) used fuse direction information between ground-truth predicted so as improve precision. experiment results show that recall (R) mAP YOLOv5s-CMS were 3% 3.1%, respectively, compared original means it can achieve better performance detection. This study provides an effective method obtaining more intuitive data sources during nematode.