【摘要】 传统的K-means聚类算法在进行图像分割时只考虑图像的特定灰度值，初始聚类中心的随机选取将导致分割结果存在很多干扰，在沥青路面这种高噪音的复杂背景下，裂缝的聚类提取效果不理想。本文提出了基于GSO-Kmeans算法来进行沥青路面裂缝分割。该算法首先使用GSO算法对沥青路面裂缝图像进行搜索，确定初始聚类中心，然后利用K-means聚类算法对沥青路面裂缝图像进行分割。结果表明，GSO-Kmeans算法在沥青路面裂缝提取方面有着很好的精准度，具有收敛速度快、分割结果准确等优势。更多还原

【Abstract】 The traditional K-means clustering algorithm only considers the specific gray value of the image in image segmentation, and the random selection of the initial clustering center will lead to a lot of interference in the segmentation results. In the complex background of asphalt pavement with high noise, the clustering extraction effect of cracks is not ideal. In this paper, GSO-Kmeans algorithm is proposed to segment asphalt pavement cracks. The algorithm firstly uses GSO algorithm to search the asphalt pavement crack image and determine the initial clustering center, and then uses K-means clustering algorithm to segment the asphalt pavement crack image. The results show that GSO-Kmeans algorithm has good precision in asphalt pavement crack extraction, and has the advantages of fast convergence speed and accurate segmentation results.更多还原