Research Of Ghost-Free High Dynamic Range Imaging Technology

High Dynamic Range Imaging (HDRI) was developed into a mature technology. It is one of the hottest research directions in graphic image processing field. HDRI has many advantages, such as rich layers, strong realism and large dynamic range, you can restore a better scene lighting effects image which more approximation to the real reality. However, when use multiple exposure photographs to recovery HDR radiance maps, we must ensure the absolutely still of the shooting scene, otherwise, the moving object will cause ghosting in the final HDRI. For the sake of remove the HDRI’s ghosting, ghost-free HDRI technology was developed. Ghost-free HDRI can apply into many fields, such as special effects in the film and television, medical, military, satellite meteorology, augmented reality and so on. In addition, ghost-free HDRI also has a very important significance for the construction of digital campus. What’s more, ghost-free HDRI technology can unrestrict by the moving objects in the scene, this feature greatly increasing the applicability of HDRI technology. In a word, ghost-free HDRI technology is a creative improvement of HDRI technology. The main works of the thesis are as follows:1. Generate an ordinary HDRI. This paper using multi-exposure images method to recovery HDRI. First, using the least squares method to recover the camera response function, so that derived from the actual scene brightness to the pixel values of the image mappings. And then weighted average the resulting irradiance value. At last, synthesized high dynamic range radiance maps of the original scene.2. Generate a ghost-free HDRI. The main advantage of this algorithm is to select the reference frame and the technical for division images into patch. By divided the original image stack into the same size patches, and put the patches match with the reference image. So that, we can remove the inconsistent patches. Finally, generate a ghost-free HDRI.3. Optimization of ghost-free HDRI. Because the application of the patch blocks and the inaccuracies of the camera response function, leading to the presence of discontinuous boundaries blur in the de-ghosting HDRI. In order to eliminate the discontinuity boundary, this paper using gradient transform and Laplacian transform methods, to optimize the ghost-free HDRI.4. This paper research and realized De-ghosting algorithm’s use in image noise and panoramic roaming.By comparative and analysis the test results, they show that, paper proposed De-ghosting algorithm can solve below problems and its advantages are as follows:①Paper proposed algorithm can successfully remove all types of ghost sources;②Algorithm can still apply into scenes which contains a large area of moving objects;③HDRI generation algorithm could maximize the use of different exposure images’dynamic range information.