A novel approach of orbit identification from level-0 data based on level-1B convention for scatterometer using fuzzy logic

This paper presents a novel Fuzzy Logic (FL)-based Orbit Identification (OI) approach for scatterometer data products which is ready to be used for ISRO's upcoming earth observation satellite's Data Quality Evaluation (DQE). A specifically designed system packages the data into standard format containing two types of information, i.e. Imaging information and Non-imaging Information. The proposed approach is related to Non-Imaging Information, more specifically for Orbit Attitude data. As the data have been prepared into the specific and scientific spatial data format for evaluation, attributes which are to be selected for preparation of fuzzy sets have been kept fixed but dynamic. An effective fuzzy system is proposed for generating rule set(RS) from OAT-data-set. Afterwords, a simplification or fuzzyfication method and a rule evaluation method are utilized to carry out useful information from the assigned value to each of the useful rules, respectively. A point dynamic fuzzy-set-classifier is developed, in which an optimal rule selection method is employed for selecting the points on orbit change accurately by diagnosis of a test OAT record. To compare the proposed approach traditional change detection(CD) and fault detection(FD) methods are also tested with same OAT-data-sets. Results show that the proposed approach of using fuzzy logic is capable of generating a number of meaningful analytical standards which are mandatory for the quality evaluation of the scatterometer multi-spectral data-products. This approach has been implemented and tested successfully to identify the orbit-cuts(OC) in OCEANSAT-I Level-0 data. Also, Performance comparison of those traditional and this novel approach is shown.