摘要
针对经典Canny算子在噪声去除、高低阈值选取和边缘断裂等问题,提出了1种改进的Canny算子边缘检测算法。首先用自适应中值滤波取代高斯滤波;然后采用改进的Otsu算法根据图像梯度幅值自适应地生成Canny算子的高、低双阈值;最后采用多特征融合技术修复断裂的边缘。实验结果表明,改进的算法具有较好的检测精度和抗噪声能力,能有效地修复图像断裂边缘,很好地控制断裂边缘的错误连接。
To solve the problem of the noise removal, high-low threshold and edge breaking in the traditional canny algorithm, an edge detection algorithm is proposed based on improved Canny algorithm. Firstly, noise is removed with the adaptive median fil-ter instead of Gaussian filter. Then, the high-low threshold of canny algorithm is adaptive generated by the improved Otsu algo-rithm based on image gradient magnitude. Finally, multi-feature fusion technique is used to repair the broken edge. The experi-mental results show that the improved algorithm not only has better detection precision and the ability of anti-noise in the edge de-tection, but also can effectively repair the broken edge and control the errors of the edge connection.
出处
《中国科技论文》
北大核心
2017年第8期910-915,共6页
China Sciencepaper
基金
国家自然科学基金资助项目(61403060)
淮安市科技项目(HAR2015007)
