KEY WORDS3 D recognition, shape from contours, perspective projection.
ABSTRACTA method to find the analytical solutions of the inverse perspective problem for the determination of the 3D object attitude in space from a single perspective image is presented.Its principle is based on the interpretation of a triplet of any image lines as the perspective projection of triplet of linear ridges of the object model. The geometrical transformations to apply to the model to bring it into the corresponding location are obtained by the resolution of an eight degree equation. The number of admissible solutions can still be reduced, using simple pruning rules. This approach leads to very strong results useful for both location and recognition of 3D-objects. Because few admissible hypotheses are retained, the line matching procedure by prediction-verification is thus less complex.
INTRODUCTIONThe human vision system is a beautiful but complex machine that many researchers in Psychophysics, Neurophysiology and Artificial Vision try to understand and even to copy in order to build efficient vision machines. However it is clear that man does not use the same processes for vision at close range or large distances. In the first case, the matching (stereo-vision principle) of the two different images perceived by each retina allows man to estimate distances and to keep directly tridimensional information on the world which is essential for hand-catching. In the second case, when man looks at far scenes or pictures, it is obvious that there is no significant differences between single or double eye vision. Nevertheless, he is often able to estimate the shape and the location of the different objects present in the scene. This implies some mental processes which permit both the recognition of 3D object from one of its 2D shapes, and the estimation of the spatial orientation of this object. To compensate the loss of information on the 3D world, man uses properties on image formation and on the shape of objects viewed under perspective projection. We believe that the enormous capacities of the human vision system proceed from the fact that these knowledge are continuously acquired and improved since childhood.The same classification can be found among the different vision machines developed by the researchers in Artificial Vision. Some work from tridimensional data obtained either by a multiple camera device using stereo-vision or by special range sensors based on triangulation [2,9]. Other systems try to CH2605-4/88/0000/0061$01.00 0 1988 IEEE understand the content of a scene from a single grey level image [l I]. For this, various techniques have been designed which can be classified in different families , depending on the knowledge used : "Shape from Shading", "Shape from Texture", "Shape from Contours'' ...In this paper, we present a method for computing the spatial location of an object from a single image. The principle is based on the interpretation of image lines as the projections of linear ridges of the object model, and on ...
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.