The Edge Detection Technology of Gray-Scale Image Based on Dynamic Fuzzy Control

Abstract

Edge detection is a key step in image processing, which is significant for image analysis and understanding. Traditional edge detection algorithms have limitations when dealing with grayscale images with blurred boundaries or low contrast features. To optimize the edge detection quality in grayscale images, relevant detection algorithms are designed based on fuzzy control methods in fuzzy mathematics. Then, a dynamic fuzzy control method is proposed to further optimize the algorithm. The study compares the detection effects of various commonly used image edge detection methods, including Canny, Roberts, Sobel, Fuzzy Control algorithm, and the research method before optimization. The results showed that the number of edge points detected by the research method increased by 32.78% compared with before optimization. The average detection time was decreased by 4.76%. The peak signal-to-noise ratio at maximum noise level was 4.43dB, and the structural similarity index was 0.0367, which was 0.35 dB and 0.0016 higher compared with the unoptimized method. The highest true positive rate and true negative rate were 47.79% and 98.87%, respectively, and the highest sensitivity was 7.75% higher compared with the method before optimization. The proportion of total hardware resources in the system was 23.71%, and the optimized detection method reduced the proportion of total hardware resources by 8.61% compared with before optimization. All relevant evaluation indexes were better than other similar algorithms. When the noise level increased from 0.5% to 1%, the fluctuation amplitude of peak signal-tonoise ratio and structural similarity index of the research method was lower than 0.12dB and 8×10-3 , respectively. The true negative rate value was stable from 93.26% to 98.87%. The research method can efficiently and accurately process edge images with low system resource utilization, improving the adaptability and robustness of the detection method.