一种利用大气散射模型实现图像去雾的方法

Abstract

针对雾天图像清晰度不高的问题，我们采用大气散射模型进行图像去雾处理，利用Curvelet变换提取图像边缘，根据边缘相交的直线先计算灭点值，再由灭点计算景深值，得到含雾图像中入射光的辐射系数，即图像真实的颜色值，达到图像去雾的目的。在景深值计算中，分别对含雾图像的每个像素点都做一次景深计算，在一定程度上可以解决由于单一景深导致部分区域颜色失真的问题，使去雾后图像的颜色更自然。最后，通过实验验证了方法的有效性。 In view of the problem that the image under foggy days is not clear, the atmospheric scattering model is used to process the image dehazing. First, the Curvelet transform extracts the image edge and calculates the vanishing point on the basis of the intersection of straight line edges. Then it calculates the depth of field value according to the vanishing point. The radiation coefficient of incident light in the foggy image is obtained, which means that the real color value of the image is obtained and the image dehazing is finished. When calculating the depth of field value, it respectively calculates the depth of field value of every foggy image’s pixel at the same time, which solves the partial area color distortion problem that is caused by using single depth of field, and makes the processed image look more natural. Finally, the validity of the method is verified by experiments.