结构光条纹图像分割方法

摘要

结构光视觉三维重建中，结构光条纹分割与提取是关键。提出一种结构光条纹分割方法，该方法首先采用多尺度 Ret-inex 算法对条纹图像暗部细节进行增强，然后使用 K 均值聚类算法对图像的高光区域进行检测，并在图像高光区域约束下提出基于最小偏态分布指标的图像分割技术，计算出高光区域和非高光区域两个分割阈值；最后对处理后的图像进行双阈值分割，获得最终的条纹图像分割结果。实验结果表明该方法能够有效处理结构光条纹图像的分割。%Segmenting and extracting structured light stripe is the key in 3D reconstruction of structured light vision.In this paper,we present a segmentation method for structured light stripe image.The method first uses multi-scale Retinex algorithm to enhance the details of darkness part of the stripe image,then uses k-means clustering algorithm for detecting the high-light areas of the image.We also proposed an image segmentation technique which is based on minimum bias-normal distribution index,at the constraints of image’s high-light area it calculates two segmentation thresholds for high-light areas and non-high-light areas.Finally,it makes dual thresholds segmentation on processed image and extracts the final segmentation result of stripe images.Experimental results show that this method can effectively deal with the segmentation of structured light stripe image.