On-spot Citrus Canker Disease Detection using YOLOv7

Citrus lemon plants are widely cultivated in India and have a variety of uses, including in horticulture, cuisine, medicine, and industry. However, they are vulnerable to various diseases that may significantly reduce their yield and quality. Hence, early detection is crucial for timely treatment and prevention of further spread. Convolutional neural networks (CNN) have been widely used in plant disease detection. However, these methods may not be suitable for real-world images with multiple leaves and multiple diseases. Instead, object detection models, such as the YOLOv7 model can better handle the challenges present in in-field images of citrus lemon plants with multiple diseases. The results show that the proposed method can detect multiple diseases on citrus lemon plants with significant accuracy for use in real-time conditions. The proposed method obtained a mean average precision (mAP) of 0.698 with an IoU threshold of 50%.