Multiscale feature detection of multispectral remotely sensed imagery in wavelet domain

This paper develops a new representation to detect the edge features on multispectral remotely sensed imagery which is based on the concept of combining dyadic wavelet transform of Mallat with first fundamental form. The dyadic wavelet transform can be applied to produce the multiscale edge detection, and making use of the first fundamental form is for edge detection of multispectral images. The representation provides a local measure for the contrast of a multispectral image at different scales. A variety of experiments by using the IKONOS images and Quick Bird images of Nanjing area are presented to demonstrate the representation efficient. It is shown from the results that the edge information of the objects, i.e. road, factory, paddy, can be detected clearly from coarse to fine at different scale levels. The technique of detecting edge information on the multispectral image is demonstrated to be beneficial to using the numerous data of multispectral remotely sensed imagery. And the multiscale edge detection method paves the way for developing the applications of object-oriented method in remote sensing.