Edge Detection Circuit Based on Memristive Crossbar Array with Adaptive Thresholding and Morphological Processing

This paper presents a novel edge detection method using a memristor-based array, which integrates image feature extraction, adaptive binarization, and morphological processing into a unified hardware system. The proposed approach leverages the memristor array to simultaneously perform horizontal and vertical feature extraction and apply an adaptive thresholding technique based on row and column averages. This in-situ processing minimizes data transfer and computation overhead, significantly reducing power consumption while enhancing edge detection robustness. Furthermore, morphological operations such as erosion and dilation are incorporated to refine edge continuity and noise suppression.