Radio Frequency Interference Detection in Microwave Radiometry Using Density Based Spatial Clustering

In this paper, the problem of detecting radio frequency interference (RFI) in microwave radiometry is addressed. Due to the significant increase in the RFI in natural emissions, an accurate detection method is crucial to retrieve the actual passive sensing measurements. To address this problem, we have proposed a density-based clustering approach that considers both RFI-free and RFI-contaminated radiometer measurements are from different distributions. Further, we have automated the proposed algorithm such that the parameters required for this approach can be extracted from any given data. The evaluation results on simulated data demonstrate improvement over existing RFI detection techniques even in low interference to noise ratio (INR) and low duty cycle (DC) RFI cases.