采用基于梯度的算法，通过matlab实现高动态范围成像。高动态范围图像，相比普通的图像，可以提供更多的动态范围和图像细节，根据不同的曝光时间的低动态范围图像，利用每个曝光时间相对应最佳细节的低动态范围图像来合成最终高动态范围图像，能够更好的反映人真实环境中的视觉效果。这里采用的高斯模糊展示了高动态范围图像中超出取值范围的数值也是有用的，即使它们在转换成低动态范围图像的时候通常都要被裁掉。在原始的高动态范围图像中，这些像素都有非常大的亮度值。当图像模糊的时候，周围的像素亮度被“拉高”并且在色度映射时裁剪成最大值，当然高亮像素的亮度也会被周围像素“拉低”，但是它们的亮度非常高所以在色度映射的时候仍然要超出最大值。这样的结果就是更大的区域变成了白色。但是对于低动态范围模糊来说，高亮区域的像素在模糊处理之前已经裁剪成了最大值，这就从总体上大幅度地降低了亮度值。因此，在模糊处理之后，光亮区域周围的像素将不再有非常高的亮度，即使是高亮区域内的像素也由于周围像素的影响而被拉低。这样高亮区域就变得比较模糊，看起来不再非常明亮。(Using a gradient-based algorithm , through matlab to achieve high dynamic range imaging . High dynamic range image , compared to ordinary images can provide more dynamic range and image detail , according to the low dynamic range images of different exposure times , using the corresponding optimal exposure time for each detail of the low dynamic range images to synthesize final high dynamic range image , to better reflect the human visual real environment . Gaussian blur used here show high dynamic range image beyond the range of values ​　​　is also useful , even if they are converted into low dynamic range images usually have the time to be cut off. In the original high dynamic range image, those pixels have a very large brightness value. When the image is blurred when the brightness of the surrounding pixels are " pulled " to the maximum value , and when chroma clipping mapping , of course, the brightness of highlight pixels surrounding the pixel will be " low " , but they a)