二重曝光高动态的图像预处理方法

Image preprocessing method of double exposure to get high dynamic range images

为提高图像识别算法抵抗环境中高低光区域的能力,提出一种使用二重曝光获取高动态图像的图像预处理技术。使用摄像机交替长短曝光,并在图层级上设置权值将这二帧进行图像合成,通过灰度映射、直方图均匀化将16位图像处理成新的8位图像;并利用经验公式确定权值大小,分析直方图信息调整长短曝光大小。实验结果表明,该算法简单高效,能有效提高输出高动态图像的灰度范围,具有抵抗高低光能力。

An image preprocessing method with double exposure technique was proposed to get High Dynamic Range （HDR） image to improve the capacity of image recognition algorithm to resist high and low light from environment. First, two frames of 16-bit image were obtained with long and short exposure parameters by camera ahemately and synthesized with an appropriate layer weight. Then, the synthesized 16-bit image was transmitted to an 8-bit image by gray-scale mapping and histogram equalization. At last, the image layer weight was adjusted according to an empirical formula and the length of exposure size was adjusted by the histogram information. Experimental results show that this algorithm was simple and efficient, and could improve the output illumination range of the HDR image effectively.