具有深度视觉感的车载红外图像彩色化方法

Color Stereo Vision Method of Vehicular Infrared Images with Depth Perception

提出了一种实现具有深度视觉感的车载红外图像彩色化的方法,对车载红外图像采用区域生长分割算法进行景物分类,分为天空、树木和道路,再分别对每一景物估计深度,然后根据得出的深度信息再结合景物的色彩先验知识,实现具有深度视觉感的红外图像彩色化。实验结果显示该方法能够得到较接近真实的深度和色彩视觉效果。

Colorization with depth perception for vehicular infrared image is implemented. We segment the vehicular infrared image into sky, tree and road classes using region growing method and then estimate the depth of each kind of scene. According to the depth information and prior knowledge of scenery colors, colors with depth perception are transferred to the infrared images. The experimental results show that the proposed method can colorize the infrared image with natural depth perception.