An Online Method for Detecting Seeding Performance Based on Improved YOLOv5s Model

Prior to dispatch from manufacturing facilities, seeders require rigorous performance evaluations for their seeding capabilities. Conventional manual inspection methods are notably less efficient. This study introduces a wheat detection approach anchored in an enhanced YOLOv5s image-processing technique. Building upon the framework, we integrated four CBAM attention mechanism modules into its model. Furthermore, traditional upsampling technique neck layer was superseded by CARAFE method. The augmented model achieved mAP of 97.14%, illustrating ability elevate both recognition precision and processing speed seeds while ensuring that remains lightweight. Leveraging this advanced model, can effectively count locate seed images, enabling precise calculation assessment sowing uniformity, accuracy, dispersion. We established test bench conducted experiments validate our results showed after improved, average accuracy above 97.55% under different rates travel speeds. indicates method has high total number particles. rate were consistent with measurements did not significantly affect or