Kirsch联合高低双阈值的多色彩空间图像边缘检测算法

Kirsch combined high and low double threshold adaptive image edge detection algorithm in multi-color space

摘要:

边缘检测是计算机视觉中非常重要且实用的图像处理方法，被应用在各个领域。然而在图像采集或传输过程中，由于外界环境的干扰，容易出现结果边缘检测率较低或者伪边缘现象，学者们为此提出了很多改进方法。但是通用的边缘检测方法确很少，现有的算法都是以处理特定场景或特定情况下的问题为目的。Kirsch联合高低双阈值的RGB图像边缘检测算法正是针对上述问题提出的。首先，提取原图RGB色彩空间下的不同分量图，对每个分量图利用改进的Kirsch算子求取边缘强度；然后利用高低双阈值划分图像的边缘点和背景点，得到不同色彩空间的边缘结果；最后对不同分量的边缘检测结果进行融合，得到最终的边缘结果。利用基准数据集BSDS500数据集中的200张测试图像对算法进行验证评估，实验结果表明，本文算法相比于其他算法检测到的边缘更加清晰，细节更加完整，边缘连贯性更好，检测率更高，适用范围更广。

Abstract:

Edge detection is a very important and practical image processing method in computer vision, which is used in various fields. However, in the process of image acquisition or transmission, due to the interference of the external environment, it is easy to have a low detection rate of the edge of the result or the phenomenon of pseudo-edge, and scholars have proposed many improvement methods for this. However, there are few general edge detection methods, and existing algorithms are aimed at dealing with problems in specific scenarios or situations. Kirsch's RGB image edge detection algorithm combined with high and low double thresholds is proposed to solve the above problems. Firstly, the different component maps under the RGB color space of the original image are extracted, and the edge intensity is obtained by using the improved Kirsch operator for each component map. Then, the high and low double thresholds are used to divide the edge points and background points of the image to obtain the edge results of different color spaces. Finally, the edge detection results of different components are fused to obtain the final edge results. The experimental results show that compared with other algorithms, the proposed algorithm has clearer edges, more complete details, better edge coherence, higher detection rate and wider application range than other algorithms.