层次聚类算法和基于图的分割算法相融合的图像分割算法

基于图的分割算法（Graph-Based Segmentation，GBS）算法）是由Felzenszwalb和Huttenlocher提出的经典的图像分割算法之一，但其分割结果中存在明显的欠分割现象。为此，在GBS算法的基础上引入层次聚类（Hierarchical Clustering，HC）算法，构造出一种解决GBS算法欠分割的方法，同时采用多线程并行处理数据的方式，有效改善了传统层次聚类算法的处理速度。该方法在RGB彩色空间中使用GBS算法得到图像中每个像素点的初始分割结果，并提取出每一类区域中的像素值，对其进行层次聚类，得到每一类区域中像素值的类别标签，根据层次聚类所得到的类别标签和预设的类别范围，修改每个像素点的初始分割结果。最后根据区域合并准则，生成一个新的分割图。经实验表明，该方法与Kmeans-SLIC(simple linear iterative clustering)算法和GBS算法等相比，很好地解决了欠分割现象，并产生了分割精度较高的语义分割图。

Image segmentation algorithm combining hierarchical clustering algorithm and graph-based segmentation algorithm

Graph based segmentation (GBS) is one of the classical image segmentation algorithms proposed by felzenszwalb and huttenlocher, but there is obvious under segmentation in the segmentation results. Therefore, based on the GBS algorithm, this paper introduces the hierarchical clustering (HC) algorithm, constructs a method to solve the under segmentation of GBS algorithm, and uses the way of multi-threaded parallel processing of data, which effectively improves the processing speed of the traditional hierarchical clustering algorithm. Firstly, the GBS algorithm is used to obtain the initial segmentation result of each pixel in the image in RGB color space, and then the pixel value in each type of region is extracted and hierarchical clustering is carried out to obtain the category label of pixel value in each type of region. According to the category label obtained by hierarchical clustering and the preset category range, the initial segmentation result of each pixel is modified. Finally, a new segmentation graph is generated according to the region merging criterion. Experiments show that compare with kmeans SLIC (simple linear iterative clustering) algorithm and GBS algorithm, this method solves the phenomenon of under segmentation, and produces a semantic segmentation graph with high segmentation accuracy.