STRETCH-CORRELATION AS A REAL-TIME ALTERNATIVE TO FEATURE-BASED STEREO MATCHING ALGORITHMS

We have analysed the requirements for a robust stereo vision algorithm for use in typical industrial applications. For such applications the views obtained in the two cameras have large differences in visual ap pearance due to the orientation difference between the two cameras and the close proximity of illumination sources. We have concluded that f or this category of problem, feature-based methods should be more robu st than conventional, area-based approaches, and this conclusion appea rs to be borne out in the published literature. However, correlation-b ased approaches are more suited to efficient implementation on availab le hardware. The technique which we have developed, called Stretch-Cor relation, is based on the cross-correlation of warped image blocks whi ch have been preprocessed to maximize the useful information content. Our new method models the severe warping effects encountered in diffic ult stereo problems and effectively relaxes the front-o-parallel const raint which is normally imposed in area-based disparity calculation. T his algorithm imposes effectively most of the local constraints presen t in feature-based algorithms, and can be efficiently implemented on a vailable hardware.