耦合光度配准的双边全变差正则化MAP超分辨率重建算法

BILATERAL TOTAL VARIATION REGULARISED MAP SUPER-RESOLUTION RECONSTRUCTION ALGORITHM WITH COUPLING PHOTOMETRIC REGISTRATION

针对曝光量不同的序列图像超分辨率重建问题,提出一种耦合光度配准的双边全变差正则化MAP超分辨率重建算法。首先计算感兴趣区域图像的方向梯度直方图(HOG)并以此为匹配的特征,其次采用直方图匹配的方法估算序列图像之间的光度映射函数,最后采用双边全变差正则化的超分辨率重建算法进行重建。通过计算重建图像的平均梯度、标准差及图像对比度发现,该算法能够有效提高重建图像的细节信息,并能有效减小重建图像间的亮度差异,提高图像对比度。

In order to realise super-resolution reconstruction of the sequence images with different exposure times,we proposed a bilateral total variation regularised MAP super-resolution reconstruction algorithm with coupling photometric registration. First,it calculates the histogram of oriented gradients（ HOG） of the images in interested regions and uses it as the matching features. Secondly,it employs histogram matching method to estimate the photometric mapping function between images. Finally,it uses bilateral total variation regularised super-resolution reconstruction algorithm to accomplish the reconstruction. It is found by calculating the average gradient,standard deviation and image contrast of the reconstructed images that this algorithm is able to improve the detailed information of the reconstructed images effectively,and can reduce the difference in brightness between the reconstructed images and raise the contrast of images effectively as well.