Research Of Double Views Motion Deblurring Method And 3D Non-contract Measurement

The requirement of automation is increase In modern industrial production, In the actual production, some parts like quality inspection and information response necessarily involves the noncontact 3D measurement of product.But because the work piece is already moving a distance in the production line in the camera shutter exposure time, will bring motion blur. motion blur will loss the image information, and make follow steps like 3D reconstruction, noncontact measurement more difficult, so we need to restore the collected images first. The existing image restoration method has some limitations in dealing with such problems, previous blind deblurring method do not considerate the corresponding relationship in the imaging plane, and difficult to reserved space information of each point in the deblurring result, this can not competent the need of the following 3D reconstruction. To solve this problem, A double view motion deblurring method is proposed.The double view motion deblurring method proposed in this paper based on binocular camera model to analyze epipolar geometry relationship, establishing the relationship of the object trajectories onto the imaging plane according to the camera model and geometric relation, retain the relationship of the PSF of each view point,and then deblurring together. We achieve good result in deblurring the image, and provide basis for the following 3D reconstruction works. Firstly we establish the relationship of the double view image PSF based on the camera model and epipolar geometry, then estimate the PSF by single image blind deblurring method, and elaboration the PSF of single image, then derivation the PSF of the other image by the geometry relationship between double view image. In the premise of guarantee the corresponding relation of the double view image PSF, We are optimizing the motion blur PSF, and deblurring the double view image together.A series of experiments verified The effectiveness and universality of the idea and method that proposed in this paper. firstly Simulation experiment have been carried out to verify the correctness of the mathematical model of corresponding relationship, simulation experiments are carried out respectively on the horizontal linear motion(single parameters) and sloping linear motion(double parameters). After verify the correctness of mathematical model, we verify the correctness in solving practical problem by experiment respectively on actual motion image( horizontal and sloping motion). A series of experimental results show that the double view motion image deblurring method that proposed in this paper can achieve good deblurring results in deal with linear motion blur on the premise of the geometry relationship between the double view images...