Research on improved diffusion algorithm based on gradient quantization in target localization algorithm

As the industrial automation field rapidly advances, the application of image processing in practical engineering becomes increasingly sophisticated. This study focuses on the target localization problem in industrial production lines, utilizing the Histogram of Oriented Gradient (HOG) algorithm to extract the gradient direction features from images. Building upon this, we propose an improved gradient diffusion algorithm. The algorithm enhances the precision and detail of each pixel's gradient direction by identifying primary gradient directions within the pixel's neighborhood and adjusting the central pixel's gradient information. This algorithm does not require a timeconsuming training phase, enhancing the robustness of image processing. It not only addresses interference from various image acquisition environments but also improves the reliability of the localization algorithm when dealing with irregularly shaped workpieces. Our experimental results have confirmed the effectiveness of the proposed method in the precise detection and localization within industrial production. When faced with various scenarios, the accuracy of its positioning exceeds 99%, and the execution time does not exceed 0.05 seconds.