Multi-Touch Recognition Related Technology Research Based On Projection Display

Based on the projection display, multi-touch technology is a hot research topic in the field of human-computer interaction, aiming at achieving multi-touch locating and recognition of contact information in the touch interface. Multi-touch locating is generally based on computer visual information processing technology through efficient image processing algorithms to achieve the denoising, the touch area extraction and contour extraction of the images collected from camera, thus the accuracy of subsequent contact coordinate positioning and mobile information recognition is improved. As an emerging technology, it is widely used in hotels, restaurants, large-scale exhibition, entertainment and other places. At present, the multi-touch technology based on the projection display has deficiencies in the robustness and efficiency of the algorithm in image recognition. Therefore, improving the efficiency of image recognition has important theoretical significance and practical value in multi-touch effective information acquisition.This paper, based on the National Science and Technology Support Project "The integration and application of holographic virtual image technology cooperated real scene system in the culture of Tang dynasty", explores the technology of finger-touch image pre-processing, touch area extraction and contour information extraction through computer vision.For the problem that the traditional pre-processing algorithm can not effectively eliminate the interference of sudden noise and sensitive to the sunlight, image denoising algorithm based on robust principal component analysis is proposed. Simulation results show that this method can effectively eliminate the influence of sudden noise information and robust to lighting condition.A finger touch region extraction algorithm based on quadtree decomposition model is proposed to deal with the inefficiency of global threshold in separating the fingers touch from the real touch areas. On the base of image segmentation of quadtree decomposition, it can solve the partitioning problem of traditional local threshold segmentation algorithm,meanwhile, eliminate the "block effect" phenomenon in traditional local thresholding algorithm when effective extract from the touch area.About the classical Canny algorithm’s poor adaptivity in the lag thresholding processing, a improved Canny algorithm based on Blob contour extraction algorithm is proposed. Simulation results show that this method is more continuous and may improve the accuracy of subsequent contact positioning to some extent.