In computer vision, edge detection is the characterization of boundaries and thus is paramount to the problem of automated feature extraction in images. This research project presents a comparative analysis of edge detection and image enhancement techniques implemented on the Raspberry Pi 5, a low-cost, resource-constrained platform. Similar comparative analyses focused on industrial settings utilizing top-tier hardware. Instead, this paper addresses the practical implementation and performance of edge detection and image enhancement techniques on affordable hardware, contributing to applications in fields such as surveillance, agriculture, and robotics. Here the performance of various common edge detection algorithms is evaluated, including the Prewitt, Sobel, and Canny operators. The goal is to determine the suitability of these techniques for real-time applications on the Raspberry Pi 5 by assessing their effectiveness in improving image quality, computational efficiency, and adaptability. While advanced techniques like Canny edge detection and Laplacian-of-Gaussian operators offer superior image quality, they are computationally demanding and therefore limiting applications requiring real-time processing on the Raspberry Pi 5. Meanwhile, simpler methods like the Prewitt operator and mean filtering provided a more efficient solution for real-time applications, striking a balance between performance and resource consumption despite less accuracy. The results highlight the need for optimization of algorithms tailored to the limitations of low-cost platforms.

Faculty Mentor: Dr. Nuri Yilmazer

Department of Electrical Engineering and Computer Science