Please use this identifier to cite or link to this item:
|Title:||Class separability and features selection: GB-MIMO SAR MELISSA cast of study|
|Authors:||MARINO GIOVANNI; TARCHI DARIO; FIGUEIREDO MORGADO JORGE; KYOVTOROV VLADIMIR; SAMMARTINO PIER FRANCESCO|
|Citation:||INTERNATIONAL GEOSCIENCE AND REMOTE SENSING SYMPOSIUM (IGARSS) p. 377-380|
|Publisher:||INSTITUTE OF ELECTRICAL AND ELECTRONICS ENGINEERS|
|Type:||Articles in periodicals and books|
|Abstract:||MELISSA is a new Ground Based (GB) MIMO SAR system developed by JRC . The processing is based on a modified 2-D FFT algorithm  and is able to create an image in azimuth and range, termed polar image by acquiring range profiles of scatters with multiple antennas. The application of this kind of system has succeeded in many security and surveillance tasks, such as avalanche, volcano eruption and landslip monitoring. Despite the successes of MELISSA, GB MIMO SAR sensors could be particularly suitable for many other applications such as MIMO MTI, 3-D ISAR systems and Automatic Target Recognition (ATR) which are still under investigation. This document is addressed for investigating the ability of MELISSA as ATR sensor. In order to understand the capability of MELISSA as ATR system, a set of experiments has been designed for testing the suit- ability of the features which have been usually adopted in SAR research community for classifying potential targets by adopting a criterion, based on Kolmogorov Appropriate Prediction of Separability (KAPS)  coefficient, for the selection of the most suitable ones. The performances of MELISSA have been assessed by employing a K-Nearest Neighbors (KNN) classifier . The first results have demonstrated the ability of MELISSA at detecting, discriminating and recognizing potential targets pre- senting in the scene and have confirmed that KAPS is a valid criterion for the feature selection. In conclusion firstly a new criterion for designing classifier algorithms based on (KAPS) has been validated by processing real data. Moreover the adopted design method has showed the ability of conceiving a reliable as well as fast classification algorithm and finally the ability of MELISSA, more generally GB MIMO SAR systems, of detecting, discriminating and recognizing po- tential targets presenting in the area of interest|
|JRC Directorate:||Space, Security and Migration|
Files in This Item:
There are no files associated with this item.
Items in repository are protected by copyright, with all rights reserved, unless otherwise indicated.