Edge detection is a fundamental technique in digital image processing, crucial for identifying object boundaries. However, detecting edges in low-intensity and noisy images remains a significant challenge. This study proposes an optimal edge detection method by integrating the Canny algorithm, Contrast Limited Adaptive Histogram Equalization (CLAHE), and Perona-Malik Diffusion Filter, with automatic kappa (k) value determination using the Fractional Order Sobel Mask. The process begins with noise reduction through the Perona-Malik Diffusion Filter, followed by local contrast enhancement using CLAHE, and concludes with edge detection via the Canny algorithm. Experimental results demonstrate that the proposed method significantly enhances edge clarity and robustness against noise compared to the conventional Canny algorithm, particularly for low-intensity images and images with noise. Tests on leaf and medical images confirm the effectiveness of this method in improving edge detection quality in digital images.