生成式对抗网络下路面裂缝图像检测识别仿真

为了有效提高生成式网络下路面裂缝检测精度,针对路面裂缝图像的复杂多样性,提出一种基于分水岭算法的路面裂缝图像检测识别方法。通过对路面裂缝图像进行分析,采用邻近区域相关知识对路面裂缝图像进行虚假裂缝去除,提取出路面真实裂缝纹理图像,引用分水岭算法对裂缝图像进行分割,获取细化后的检测图像,并将相同路面的两种图像进行融合,完成路面裂缝图像检测识别。实验结果表明,所提出基于分水岭算法的路面裂缝图像检测识别方法检测时间较短、检测准确率较高、误报率较低。

Simulation and Detection of Pavement Crack Image Detection Based on Generated Confrontation Network

In order to effectively improve the detection accuracy of pavement crack based on generative adversarial network,a method to detect and recognize pavement crack image based on watershed algorithm was proposed.Through the analysis of pavement crack image,the relevant knowledge of adjacent area was used to remove the false crack of pavement crack,so that the real crack texture image was extracted.Then,the watershed algorithm was used to segment pavement crack image,so as to obtain the refined detection image.Moreover,two images of the same road surface were fused to complete the detection and recognition of pavement crack image.Simulation results show that the proposed method has shorter detection time,higher detection accuracy and lower false positive rate.