基于改进Canny算法的合作靶标边缘检测

Cooperative targets edge detection based on improved Canny algorithm

合作靶标在机器视觉中应用广泛,对合作靶标图像进行采集处理并获取目标的精确定位是目前研究的主流。边缘检测是对靶标图像进行处理的关键一步。该文针对传统Canny算法中存在的缺陷进行了改进,使用自适应中值滤波-导向滤波代替传统高斯滤波,去除噪声的同时保留了边缘信息;计算梯度时增加45°和135°方向,防止边缘细节丢失;采用Otsu算法结合梯度直方图自适应获取阈值,提高阈值的准确性。实验和数据表明,该文算法在边缘完整和细节保留方面更有优势,较传统Canny算法提升约15%~20%。

Cooperative targets are widely used in the field of machine vision,and achieving precise localization of targets is currently a focus of research.Edge detection is a critical step in processing target images.This paper proposes an improved approach to address the limitations of the traditional Canny algorithm.Specifically,this algorithm replaces traditional Gaussian Filter with Adaptive Median Filter and Guided Filter,which effectively suppresses noise while preserving edge information.Additionally,we introduce 45°and 135°directions to gradient calculation to prevent the loss of edge details.Finally we employs the Otsu algorithm in combination with a gradient histogram for adaptive thresholding,which enhances threshold accuracy.Extensive experiments and data analysis demonstrate the superior performance of our approach in terms of edge integrity and detail preservation,which improves by about 15%~20%compared with the traditional Canny algorithm.