基于自适应引导滤波的全景图像增强算法研究

虚拟实景漫游系统中的全景图像细节比较多，因天气、拍摄技术等原因导致图像场景不清晰而需要增强，增强的目的是为了突出全景图像的美景细节信息，提高场景视觉效果。为了增强全景图像的场景细节，本文提出了自适应引导滤波算法AGIF（Adaptive Guided Image Filter）。该算法利用具有较强细节提取能力的加权最小二乘滤波器WLS （the Weighted Least Squares）构建引导图像，对已有的引导滤波GIF（Guided Image Filter）进行改进，通过引入梯度值偏移量进行像素值自适应，突出全景图的图像边缘，使得全景图像的细节进一步增强。实践表明，增强后的全景图用于虚拟漫游系统，提高了旅游场景的用户体验。

ADAPTIVE GUIDED IMAGE FILTERING FOR THE PANORAMIC IMAGE ENHANCEMENT

Aiming at the problem of image with low contrast cause by weather conditions, capture device or the technological level of photographer and the problem of indistinct mosaic caused by late panorama mosaic algorithm, the paper takes the advanced adaptive guided image filter algorithm for image detail enhancement. This paper using the weighted least squares filter for the construction of guided image, improving the existing adaptive filtering method, so as to better enhance the shape and surface details through introducing the gradient offset. The algorithm can improve the visual effect of the tourism scene from both the subjective result and object quality evaluation. At the same time, the algorithm can also compress the file size to achieve the transmission requirement of lightweight travel scene, so as to improve the user experience of virtual tourism.