Corners and vertexes are strong and useful features in computer vision for scene analysis, stereo matching, and motion analysis. Here, we deal with the development of a computational approach to these important features. We consider first a corner model and study analytically its behavior once it has been smoothed using the wellknown Gaussian filter. This allows us to clarify the behavior of some well-known cornerness measure based approaches used to detect these points of interest. Most of these classical approaches appear to detect points that do not correspond to the exact position of the corner. A new scale-space based approach that combines useful properties from the Laplacian and Beaudet's measure (Beaudet 1978) is then proposed in order to correct and detect exactly the comer position. An extension of this approach is then developed to solve the problem of trihedral vertex characterization and detection. In particular, it is shown that a trihedral vertex has two elliptic maxima on extremal contrast surfaces if the contrast is sufficient, and this allows us to classify trihedral vertexes in 2 classes: "vertex," and "vertex as corner." The corner-detection approach developed is applied to accurately detect trihedral vertexes using an additional test in order to make a distinction between trihedral vertexes and corners. Many experiments have been carried out using noisy synthetic data and real images containing comers and vertexes. Most of the promising results obtained are used to illustrate the experimental section of this paper.

No abstract

This paper presents a scene interpretation system in multisensor fusion context whose application involves the interpretation of remote sensed images. First we discuss how multisensor fusion is achieved, and we derive the modeling problems for objects, scene, and strategy. The proposed multispecialist architecture generalizes the ideas of our previous works by taking into account the knowledge about sensors, the multiple viewing notion (shot), and the uncertainty and imprecision of models and data modeled with the possibility theory. Especially, generic models of objects are represented by concepts independent of sensors (geometry, material, and spatial context). Three kinds of specialists are present in the architecture: generic specialists (scene and conflict), semantic object specialists, and low level specialists. A blackboard structure with a centralized control is used. The interpreted scene is implemented as a matrix of pointers allowing easy conflict detection and easy spatial context verification. Under the control of the scene specialist, the conflict specialist solves conflicts using the spatial context knowledge of objects. Finally, an interpretation system with SAR/SPOT sensor images is described, and an example of a scene interpretation involving rivers, bridges, urban areas, and roads detection is shown.