Observations of forests using a helicopter-borne nadir-looking radar: a new method for morphometric purposes

The authors present an original pattern recognition method for the observation of forests using a nadir-looking ranging scatterometer. The radar configuration provides vertical profiles of the vegetation. Radar data were not processed in a classical way, i.e., the calculation of a radar cross section, but through pattern recognition methods applied to the raw data. The aim of the study was to search for typical patterns related to three types of forests. A small subsample of the data was determined for methodology validation purposes. Some results are also reported for a much larger data set. Morphodecomposition, a new methodology presented for the first time in these types of applications, is able to recognize the main features of the data, not only the types of forests involved, but also subtle morphological patterns of the signals. Morphodecomposition provides a morphometric methodology applicable to other geophysical univariate sampled signals.