基于颜色特征的胸环靶图分割算法

Chest Silhouette Segmentation Algorithm Based on Color Feature

胸环靶图的识别是自动报靶系统的重要研究内容之一。在分析胸环靶颜色特征的基础上,提出了一种基于颜色特征的胸环靶分割方法。首先从RGB颜色模型颜色空间分析了靶环绿色恒量的存在性;并基于绿色恒量特征初提取胸环靶图像。然后在HSI颜色空间中基于色调和饱和度分量进一步提取靶环图像。最后对二值化后的图像作投影变换,确定胸环靶在图中的位置。该算法其他的胸环靶分割算法,避免了阈值选择不当的问题;并且能够有效克服光线、背景变化和天气条件的影响,具有很好的鲁棒性和自适应性,定位精度高。

Chest silhouette recognition is a principal issue in automatic target-scoring system,on the basis of analysis of Chest silhouette color characteristics,a new segmentation method was proposed based on color feature First,the color constant in RGB space was found to segment the green Chest silhouette from the background; then the further extracting target ring image based on hue and saturate component in HSI color space; Finally,a projection transformation was taken to binarized image to determine the position of Chest silhouette in the figure. Different with other segmentation algorithm,the algorithm avoids the problem of improper threshold selection,and can effectively overcome the influence of changes of light,the background and weather conditions,with high adaptation and robustness and positioning accuracy.