Edge detection is crucial for computer vision tasks like face and pedestrian detection. A new algorithm called B-Edge addresses issues with existing methods by using multiple thresholds to improve edge connectivity and thickness. By selecting simulated triple thresholds, the algorithm enhances image contrast, selects edge pixels effectively, handles errors, and matches ground truth. Experimental results show that B-Edge outperforms other algorithms for both grayscale and colored images, providing better edge detection for various applications.