基于边缘重构图像的边缘检测算法优选研究

Edge detection algorithm optimization based on edge reconstruction image

边缘检测是视觉定位、零件缺陷识别、视觉测量等图像处理中的关键环节,其检测质量直接影响后续的图像目标识别与定位的精度。针对具体图像处理需选取合适的边缘检测算法的问题,提出了一种基于多方向滑动窗口线性插值重构法的图像边缘检测质量评价算法。首先,采用待处理边缘图像的最大连续非边缘矩形区域作为重构搜索窗口,设计了多方向滑动窗口线性插值图像重构法;然后,根据图像边缘的特征,建立了集成图像结构相似度和边缘错检率为性能指标的边缘检测算法的优选评价方法;最后,进行了实例应用,通过实验对重构算法的性能、优选方法的可行性和正确性进行了验证。研究结果表明:该边缘图像重构算法准确率最高,实验中重构图像的结构相似度可达到0.7087,耗时320.9021 s;采用该优选方法能快速地筛选出最佳的边缘处理算法,与人体视觉评价一致,便于实现边缘检测算法的智能优选和图像的智能处理目标。

Edge detection is a key link in image processing,such as visual positioning,part defect identification and visual measurement,and its detection quality directly affects the accuracy of subsequent image target identification and positioning.Aiming at the problem of selecting the appropriate edge detection algorithms for specific image processing applications,based on the improved image reconstruction method,an image edge detection quality evaluation algorithm based on multi-directional sliding window linear interpolation reconstruction method was proposed.Firstly,using the maximum continuous non-edge rectangular region of the edge image to be processed as the reconstruction search window,a linear interpolation image reconstruction method with multi-directional sliding window was designed.Then,based on the characteristics of image edges,an optimal evaluation method for edge detection algorithms was established,which integrates image structural similarity and edge error detection rate as performance indicators.Finally,the examples were applied to verify the performance of the reconstruction algorithm and the feasibility and correctness of the optimization method through experiments.The results show that the accuracy of the edge image reconstruction algorithm is the highest.In the experiment,the structural similarity of the reconstructed image can reach 0.7087,and the time consuming is 320.9021 s.Using the proposed optimization method can quickly screen out the best edge processing algorithm,and the evaluation result is consistent with the human visual evaluation,which is convenient to realize the intelligent optimization of the edge detection algorithm and the intelligent image processing.