Forest classification by means of pattern recognition method applied to scatterometer data

Presents vegetation classification results using C-band helicopterborne scatterometer data acquired over different forested areas during two experimental campaigns in France (l991) and in French Guiana (in the framework of the ESA SAREX campaign in 1992). The scatterometer involved in the two campaigns was ERASME, a small, low power, high resolution (1 m), ranging radar. It operated as a sounder of the trees, in a nadir looking mode. A new pattern recognition method, which the authors call morphodecomposition, has been used to process the data and is also presented. The results of this statistical method of analysis applied to four classes of forests are promising