Color Register In Camera-Based Multi-Projector Large Scale Displays

In the recent years, along with the calculation ability and saving ability of calculator, the science calculating scale is more and more big in the meteorologic analysis, biology techniques, reservoir emulation and nucleus techniques. The meshes to calculate is up to million. How to deal with so many data quickly and exactly is a problem in need of solution. However, large-scale systems now available adopt expensive monolithic rendering engines and specialized light projectors. This makes the system an expensive "luxury" for large-scale visualization. However, camera-based multi-projector displays are the most effective method to solve the problem.The research on large-scale display system develops fast in the abroad. This commonly involves three parts: geometric correction, color registration and intensity blending. Geometric calibration is needed to remove the discontinuity caused by mis-aligned projectors in the overlapped regions;Due to photometric imbalance within and among the projectors, color register is needed to achieve uniform effectiveness of such a display;In order to seamlessly merge the images without any obvious overlapped region, intensity blending is needed.Famous large, high-resolution display systems, such as Princeton Scalable Display Wall, require aligning projectors precisely and take up large construction room. The installation and operation of such systems is extremely expensive.Raskar et. al. proposed to create a low-cost but easy-to-use multi-projectors mosaic system by exploiting computer vision. The cost of the system is simply the cost of the projector in addition to the inexpensive camera and software. But color calibration and photometric correction were neglected, thus the system can be improved greatly. Based on these researches, the paper proposes a new approach to realize the color registration of Camera-based Multi-projector Large Scale Displays.We make use of three-order B-spline to fit the color space of projectors. First, some sample colors are projected from each projector. Second, a camera is used to catch the response of each color from each projector. As a result, we can obtain all other color response by B-spline function. At last, a compensate function is calculated to correct the color of each projector.The method we present here is entirely different from the others, which are being used. The technique based on B-spline can low down the complexity of calculation. And we need not to calculate the gamma function, which is necessary in the photometric uniformity. Meanwhile, we present a whole new approach to solve this kind of problem.