Research On Key Visual Technology Of Cable-driven Parallel Robot Environment

Cable-driven parallel robot is a new type of parallel robot,which is driven by cables instead of rigid links.There are many advantages about the robot such as simple structure,easy reconfiguration,light mechanism,large translational workspace,lower cost of manufacture and so on.Because of those advantages so that it can be applied to many complex environment scenes.How to get an effective and accurate robot environment image in this complex environment is an image denoising problem that must be solved first in the visual technology.In the process of obtaining the working environment image,due to additive white Gaussian noise generated by Camera sensor heat and impluse noise generated by the transmission channel factor,the actual image is usually corrupted by the mixed noise in practical applications.In this paper,the mixed noise is additive white Gaussian noise(AWGN)coupled with impluse noise(IN).An algorithm is proposed for the mixed noise removal.Firstly,the impluse noise selective method of two-threshold is used in it.Secondly,the image priors about sparsity and nonlocal similarity are integrated into the existing weighted coding algorithm.Finally the denoising image is obtained.The detailded work is as follows:1 ? I carry out in-depth analysis and experimentation of the noise model(Gaussian noise,impulse noise and mixed noise)and simulate the mixed noise.2 ? The principle and algorithm of image denoising based on non-local similarity and sparse representation are analyzed respectively,and the correlation index of the denoising performance is obtained experimentally.3?According to the denoising model,a more effective detection method for impulse noise detection is proposed.4?For the mixed noise in this paper,the non-local similarity and sparse representation of the prior knowledge are integrated into the denoising model effectively,and the optimization design of denoising model is obtained through the regularization.By solving the denoising model,the ideal denoising image is got.5?The image denoising experiment is carried out on the algorithm of the paper and several typical denoising algorithms,and the superiority of the algorithm is proved from the objective data and the effect graph.