Enhancement of circle hough transform to be applied in circular road traffic sign recognition

ABSTRACT: The Enhanced Circle Hough Transform (ECHT) is an optimized algorithm designed to improve the accuracy and efficiency of circular object detection in digital images. This study focuses on minimizing computational load while enhancing detection precision and reliability. ECHT incorporates three core enhancements: the Kasa method for exact mathematical circle fitting, Contrast Limited Adaptive Histogram Equalization (CLAHE) for refining potential circle center locations by improving local contrast, and adaptive parameters that dynamically adjust in response to input data. These improvements eliminate the need for manual tuning, ensuring optimal performance under varying conditions and increasing overall productivity. The objective of this research are: (1) to lessen the computational complexity using mathematical equations such as the Kasa method, (2) to improve the algorithm’s capability to accurately recognize circles while minimizing false detections, and (3) to implement adaptive parameters for automatic system optimization. The methodology includes applying the Kasa method for efficient circle fitting, utilizing CLAHE to enhance image contrast and assist in identifying likely circle centers, and integrating adaptive parameters to allow real-time adjustments based on image variability. Performance evaluations demonstrate ECHT’s superiority over the traditional Circle Hough Transform (CHT). ECHT achieves higher precision (81% vs. 29%), greater F1-score (75% vs. 33%) and faster execution time (0.3113 sec vs. 0.4464 sec). Although CHT reports a slightly higher recall (84% vs 75%), it results in more false positives, lowering overall detection reliability. ECHT, by contrast, offers a better balance of precision and recall, making it a more robust and computationally efficient solution for practical circle detection tasks.

Filipiniana