A NEW METHOD FOR THE CLASSIFICATION OF THE SATELLITE ALTIMETRY WAVEFORMS

Satellite radar altimetry has been utilized to study inland waters for more than two decades. The satellite onboard processing unit, flags the data referring to these waters as unacceptable observations. In order to employ this data, waveform re-tracking implementation is necessary. Firstly, the aforementioned correction is calculated based on the waveform shape and secondly, the satellite processing unit and the re-tracking methods are established based on the Brown model. Since, the more similar the waveforms are to the Brown model, the more suitable they are. Therefore, such classification based on the waveform shape is needed. This paper proposes an alternative method for the waveform classification based on echo shapes. This assertion is confirmed by the application of statistical concepts, Correlation Coefficient and Cross Correlation. The numerical results, from the Louisiana wetland and the Lake Michigan as the test fields, express accuracy increase. The best result of the first case study which shows ۵.۶ times accuracy increase attributes to the Threshold re-tracking method with ۳۰% as the threshold value by the use of ۸۰% as the classifying threshold. Also applying re-tracking method, Threshold (۴۵%) and ۸۰% as the classifying threshold value has achieved the best result with ۲۷% accuracy increase at the mentioned lake. Thus the credibility and efficiency of this method can be concluded.

Keywords:

classification , satellite altimetry , waveform , correlation coefficient , cross correlation