改进的自适应Canny边缘检测算法

Improved adaptive Canny edge detection algorithm

针对传统Canny算法需要设定滤波参数且对椒盐噪声滤波效果不佳,以及需要预先设定高、低阈值等问题,提出一种改进的自适应Canny边缘检测算法。利用分治法思想和相邻窗口排序信息相关性的原理,改进自适应中值滤波器,以此来替代高斯滤波器;增加45°和135°方向的梯度模板计算像素梯度;利用二分法查找原理改进Otsu算法计算梯度幅值的高、低阈值。Matlab仿真结果表明,该算法在椒盐噪声和高斯噪声环境下都有较好的边缘检测效果,具有自适应性强、抗噪性好、计算机耗时少等优点。

For the problem that the traditional Canny algorithm has poor effects on salt and pepper noise and needs to set the filtering parameters,high and low threshold in advance,an improved adaptive Canny edge detection algorithm was proposed.Based on the method of partition,and the adjacent windows ordering information relevance principles,the adaptive median filter was improved to replace the traditional Gaussian filter.The gradient template in the 45°and 135°directions was increased to calculate the pixel gradient.Based on the dichotomy search principle,the Otsu algorithm was improved to calculate the high and low thresholds of gradient amplitude.The Matlab simulation results show that the proposed algorithm has good edge detection effects for the salt and pepper noise and Gaussian noise environment,and has the advantages of stronger adaptability and good anti-noise as well as less computer time.