摘要
为了将红外图像中所包含的信息更加友好、直观地呈现给用户,改善用户对于红外图像的理解效果,针对车载红外图像的特点,提出了一种基于随机森林分类器和超像素分割算法相结合的车载红外图像彩色化算法。首先对原图的各个像素点进行特征提取,然后训练随机森林分类器,使它能够对待测试图像的各个像素点进行正确的分类。再使用超像素分割与直方图统计相结合的方法对分类结果图像进行优化。最后将优化后的分类结果图像转换到HSV色彩空间进行对应的色彩传递。通过实验证明该方法能够在很好的对红外图像进行彩色化处理的同时,保证色彩传递的正确性和实时性。
In order to improve the effect of the information contained in the infrared image which makes the infrared image much more friendly and intuitive to users. According to the characteristics of the vehicular infrared image, this paper proposes a vehicular infrared image colorization algorithm, which is combined of random forest classifier and superpixel segmentation algorithm. Firstly this method extracts the original characteristic of each pixel, and then trains the random forest classifier which can make sure that each test image classified correctly. Secondly it can use the combination of superpixel segmentation and histogram statistics to optimize the classification results. Finally it can convert the optimization of classification result images to HSV color space and do the corresponding color transfer. The experiments prove that this method can be very good in dealing with infrared image colorization, and at the same time, it can ensure the accuracy and timeliness of color transfer.
出处
《红外技术》
CSCD
北大核心
2015年第12期1041-1046,共6页
Infrared Technology
基金
国家自然科学基金资助项目
编号:61072090
61205017
61375007
关键词
车载红外图像
图像彩色化
随机森林
超像素分割
vehicular infrared image, image colorization, random forest, superpixel segmentation
