Accuracy Comparison of YOLOv7 and YOLOv4 Regarding Image Annotation Quality for Apple Flower Bud Classification

Object detection is one of the most promising research topics currently, whose application in agriculture, however, can be challenged by difficulty annotating complex and crowded scenes. This study presents a brief performance assessment YOLOv7, state-of-the-art object detector, comparison to YOLOv4 for apple flower bud classification using datasets with artificially manipulated image annotation qualities from 100% 5%. Seven YOLOv7 models were developed compared corresponding terms average precisions (APs) four growth stages mean APs (mAPs). Based on same test dataset, outperformed all at training quality levels. A 0.80 mAP was achieved quality, meanwhile 0.63 only 5% quality. improved 1.52% 166.48% mAPs 3.43% 53.45%, depending stage Fewer instances required than achieve levels accuracies. The AP increase observed instance number range roughly 0 2000. It concluded that undoubtedly superior classifier YOLOv4, especially when suboptimal.