Fan Zhang Longgang Zhao Dongwei Wang Jiasheng Wang Igor Smirnov Juan Li ^*

• Qingdao Agricultural University, Qingdao, Shandong Province, China

The emergence rate is an important indicator for variety selection, variety evaluation, field management, and yield prediction. To solve the low recognition accuracy problem caused by uneven size and different growth conditions of crop seedlings under salt-alkali stress, this research proposes a peanut seedling recognition model MS-YOLOv8 to fast recognize and count for peanut seedlings by using closerange remote sensing from unmanned aerial vehicles. Specifically, this paper first proposes and constructs a lightweight adaptive feature fusion module (called MSModule), which can group the channels of input feature maps and input these feature maps into different convolutional layers with different kernel sizes for multi-scale feature extraction. Furthermore, the module can automatically adjust the channel weights of each group of feature maps based on their contribution, which improves the feature fusion effect. Secondly, the structure of the neck network is reconstructed to enhance the recognition ability for small objects based on YOLOv8. Moreover, the MPDIoU loss function is introduced to solve the problem that the original loss function could not effectively optimize the detection box of seedlings with scattered branch growth. Experimental results demonstrate that the proposed MS-YOLOv8 model achieves an AP50 of 97.5% for peanut seedling detection, which is respectively 12.9%, 9.8%, 4.7%, 5.0%, 11.2%, 5.0%, and 3.6% higher than Faster R-CNN, EfficientDet, YOLOv5, YOLOv6, YOLOv7, YOLOv8, and RT-DETR. This research provides valuable insights for crop recognition under extreme environmental stress and lays a theoretical foundation for the development of intelligent production equipment.