Automated Detection of Wheat Powdery Mildew Using YOLACT Instance Segmentation

There is a 5% annual decline in wheat results worldwide due to Blumeria graminis f. sp. tritici (Bgt), which makes an accurate identification methodology important. This study addressed the challenges encountered in early wheat powdery mildew detection through the development of an innovative technique that makes use of the YOLACT model, a real-time object recognition system. The review of the literature covers a variety of approaches, such as deep learning and hyperspectral imaging, with an emphasis on YOLACT incorporation for improved early detection. The study describes dataset collection, YOLACT integration, and preparation methods. Having a 95.6 ^% recognition rate, a mean Intersection over Union (MIoU) of 0.85, a precision of 0.81, and a mean average precision (mAP) of 0.56, the results illustrate the efficiency of the YOLACT model. Comparative assessments illustrate how much stronger the suggested YOLACT methodology is. The study concludes by highlighting the accuracy of the YOLACT model in detecting wheat powdery mildew and stressing the value of texture analysis and the best possible combination of vegetative indices (VIs) for early diagnosis. By fusing technology with realistic farming situations, the YOLACT model not only increases disease identification but also can entirely reinvent agriculture and increase crop cultivation's sustainability and efficiency practices