Deep Learning-Based Wheat Stem Rust Disease Recognition Using Mask Scoring RCNN

Over the past 25 years, climate change has made it easier for fungal infections to spread, particularly Puccinia graminis f.sp. tritici (Pgt), which leads to wheat stem rust and offers a grave risk to global wheat output and food security. In this paper, the usage of image processing techniques and the mask-scoring R-CNN model for independent wheat stem rust recognition. With a precision of 82.92% and a Mean Intersection over Union (MIoU) of 95.32%, the model exhibits outstanding performance metrics after its training on a dataset of one thousand high-resolution photos. Using preprocessing techniques like normalization and scaling along with deep learning, the model correctly identifies wheat stem rust lesions on individual wheat leaves. Later studies combining advanced AI techniques such as meta-learning and reinforcement learning show interest in increasing the model’s adaptability to various contexts. Operating together with partners in the food system can encourage innovation, improve crop resilience, fight world hunger, and advance sustainable agricultural methods.