Research On ROI Image Processing Technology Of Construction Robot Based On Trinocular Stereo Vision

With the development of remote control robot technology,information technology, the clearer of digital image will lead to a large quantity of image data and too long time for transport,consequently affecting the real-time of remote control system. Hence it is needed for an image processing technology which can store image in the smaller and less space and transport with less bit-rate under the premise of image quality,the image processing tasks were completed by coding technology,the image of the non-ROI region hierarchical fuzzy,do feature extraction and measurement for target.Simulation of the human eye ROI image processing technology not only allows images to be compressed,but also reduce the image transmission time and improve the real-time control of robotic systems.Therefore,the simulation of the human eye ROI image processing techniques will play an important role in the remote control robot or the field of remote sensing.In this paper,on the basis of the research of robot system works with stereo vision both at home and abroad,we analyze and compare the traditional image processing methods,such as image segmentation,image coding.According to the basic principles of the human eye to identify objects and image fuzzy approach,describe a new type of simulated human eye ROI image processing technology.The technology is based on the basic principles of Gaussian blur to study the Gaussian radius select of a various parts of the image.Hierarchical Gaussian blur displace the human eye about the object recognition process for image,and verify the feasibility of the algorithm by test.This article relies on four degrees of freedom construction robot test platform in machinery and electronics engineering laboratory,Jilin University,solved the parameters of stereo vision system,and done the camera calibration.Using the new matching algorithm for stereo matching,and the simulated human eye ROI image processing method was tested.My research includes the following five aspects:1.Through collecting,reading,comparing,analyzing about stereo vision both at home and abroad,based on the research conditions and trends in this field,combined with the development of stereo vision technology,the topic "Research on ROI Image Processing Technology of Construction Robot Based on trinocular Stereo Vision"was determined,and described the research purpose,meaning and the main contents of the study.2.Based on the trinocular stereo vision system,construct a camera model about the pinhole imaging.Thinking the problem that the linear camera calibration accuracy is not good and the work of calculation is hard,the impact of tangential distortion and radial distortion,We improved calibration algorithm for 3D stereo target.Through the solution of calibration algorithm with MATLAB,we obtained the parameters of the camera,and optimized the parameter with non-linear optimization algorithm. Calibration test data show that the algorithm improved the precision of camera calibration and simplifies the calibration process.3.By comparing a variety of matching algorithms,we described the trinocular stereo matching algorithm based on the edge.By increasing the binding and the line-point corresponding method at different levels,realized the stereo matching;Parallax measurement method achieve the object distance measurement and got the stereo map and the depth map of object.The algorithm has better matching accuracy and location accuracy verify the feasibility of the algorithm by experiment.4.Do the research about simulated the human eye ROI image processing technology. Analysis the advantages and disadvantage of traditional image processing,the basic principles of human eye to identify objects,as well as a variety of image blur approach,investigate a new type of simulated human eye image processing method of ROI.Due to the dynamic Gaussian blur similar with the principle of human eye to identify objects,combined with Gaussian blur algorithm analysis,MATLAB simulation and the discussions about Gaussian fuzzy processing achieve the simulation of human eye identifing objects,in theory,the feasibility and superiority of algorithm were analysised.5.We complete the study of simulated human eye image processing method of ROI.Based on trinocular cameras and DIGI-HIBW-60 stereo imaging system,using VC++ compile the algorithm of simulated human eye ROI image processing method, the test of the simulated human eye ROI image processing method was carried out by Using black and white photographs,color photographs.The test results show that the algorithm can achieve higher image compression and the simulation process of the human eye to identify objects.That the algorithm reduces the image transmission time was verify by the tests of image transmission use the IP Message software.In this paper,Image decompression was't essential for the simulation of the human eye ROI image processing technology,image transmission time shortened,so that real-time of remote control robotic systems has been ensured.In this paper,the ROI image processing technology has complete the initial ROI image processing to simulate the human eye,that lay the foundation for the research of the human eye-level ROI image processing and image transmission technology in the future.