作　　者： (孙砚飞）; (常晓刚）; (李东兴）; (张华强）; (马良慧）;

机构地区： 山东理工大学机械工程学院,山东淄博255049

出　　处： 《山东理工大学学报（自然科学版）》 2017年第6期18-21,共4页

摘　　要： 针对Canny算法需要人工设定高斯方差值和双阈值,红外图像存在噪声大、边缘模糊等缺点,提出一种基于自适应Canny的红外图像边缘检测算法.该算法采用自适应中值滤波代替高斯滤波计算梯度的幅值和方向,对梯度的幅值在3×3邻域内进行非极大值抑制,并根据图像灰度使用Otsu算法,自适应获取高低阈值,用高低阈值算法检测和连接边缘.实验结果表明,该算法减小了均方误差,提高了峰值信噪比和平均结构相似度,能有效提取红外图像边缘. The gaussian variance and double-threshold should be determined artificially, mean- while, aiming at defects of infrared image with high noise and fuzzy edge, an infrared image edge detection algorithm is proposed based on adaptive Canny. Firstly, adaptive median filter is used to replace gaussian filter. Secondly, the amplitude and direction of the gradient is computed. And then, non-maximum suppression is used to process the gradient magnitude in the 3 M 3 neighborhood. Finally, Otsu algorithm is adopted to get high and low thresholds adaptively according to the gray, and image edge is detected and connected by high and low thresholds algorithm. The experimental results show that the mean square error （MSE） is decreased, the peak signal to noise ratio （PSNR） and the mean structural similarity （MSSIM） are simultaneously increased, and then edge of the infrared image is extracted effectively.

关 键 词： 红外图像 边缘检测 自适应中值滤波 算法 算法