GPU Accelerated Code Optimization: Leaf Disease Detection

Modern GPU's highly parallel structure makes manipulating computer graphics and image processing, more efficient and faster than general. Our project is based on Code Optimization which makes use of GPU acceleration. Code optimization can be achieved by parallel execution using GPU. GPUaccelerated computing offloads compute-intensive portions of the application to the GPU, while the rest of the code still runs on the CPU. We propose and experimentally evaluate a solution for faster detection and classification of plant leaf diseases using GPUs. Our project solution is an improvement to the existing system as it gives more faster and accurate solutions. To demonstrate the effects of optimizations over time, we at first experimented on CPU (using Matlab) and then we will experiment it on GPUs using OpenCV to understand the improvement in performance. Identification of plant diseases is beneficial in monitoring large fields of crops, and automatically detect the symptoms of diseases at early states only. Therefore, for great realistic significance one need to look for fast, automatic, less expensive methods to detect plant diseases. By making use of OpenCV libraries in mobile's GPU through an Application, it helps to decrease the execution time of detection of disease level and also provides prevention method (name of disease and pesticides to be used) in English and Marathi and the percentage of the infected area. The language flexibility is provided for better understanding of farmers. The dataset of leaf disease has taken from Kaggle site. The mobile application identifies four diseases: Rust, Tar spot, Linden Leaf and Phyllosticta.