Research On Key Techniques Of Vision Based On Line Structured Light

In the3D measurement system, structured light visual system was widely used in thereconstruction of the3D model, measurement of surface profile and other fields, because ofits non-contact, high controllability,real-time, high precision,wide measurement range,andsimple structure.Firstly, the characters of the structure light images and their pre-processing are studiedand do some pre-processing about the structure light image. After an analysis of the commonmethods in structured light center extraction, an improved algorithm which based on sub pixelwas proposed. The algorithm should calculate the maximum grey value in every row of thestructured light line and get the normal direction of this point. According to the set of thosepoints, a quadratic curve can be fitted. After that, the center of the structured light center canbe calculated according to the quadratic curve. The result shows that this method improvingthe speed and accuracy because it avoiding a large number of convolution filtering calculationwhich existing in Hessan matrix mothod. It also ensures that the structured light center has ahigh precision and robustness. It has more practical value.Image mosaic is a technology which uses space matching alignment to the mutualbetween the overlapping portions of the image sequence to get the complete image whichincludes more information. It includes image matching and image fusion. On the basis ofabove theory, line structure light as specific artificial point was placed in the surface of thedetected object, so that we can narrow the search scope of feature points to be extracted,alsofeature points were easy to match.Thus shortened the corresponding calculation and improvedthe speed and accuracy of the match,had robustness in a way. The dissertation introduce acoding scheme which base on the theory of channel coding in the information theory. Thiscoding scheme can correct the error automatically when appeared an error. Then we used thisscheme image as the mark point on the surface of the object and get the images from different angle. After got the images, taking corresponding action and finished the task of imagestitching. This method can improve the precision and robustness about image stiching. Itprovided a new idea for image stitching.