Multiscale merging by a mixed approach for edge detection

We present an edge detection algorithm from a multiscale analysis, for the real images with which we are confronted in our laboratory. These images are often 3D and because of the fields of depth of the lenses, the edges to be detected are blur. Our work is based on a generalization of the Canny-Deriche (1986, 1987) filter, modelized by a more realistic edge than the traditional ramp shape edge. This filter is used to generate a frame of wavelets. We propose a new criterium which allows to choose in a deterministic way the intrinsic parameters of the filters. This criterium introduces the concept of detection resolution which is a new concept in edge detection. For the merging of the data from the projection we use a new method which is based on a scanning of the scales in both directions: coarse-to-fine and fine-to-coarse, in order to ensure the best tradeoff between detection and localization. An example of the result obtained is presented with a β-version of the method (monodimensional, without detection of the local maximums and without taking into account all our propositions)