Abstract:

The noise and interference are the critical influences on the analysis of electromagnetic radiation in ionosphere based on the satellite electromagnetic observation. An autodetection method is presented which is combined with the threshold segmentation and morphological filter to solve this problem. Due to the electromagnetic radiation having certain time width and bandwidth characteristics, the timefrequency spectrogram is converted to binary images by threshold segmentation, then the noise and interference can be deleted by setting the appropriate structural elements. The excellent performance of this method is verified by detecting and statistically analyzing the electromagnetic emission with lefthand polarization, and it is very effective for data mining from massive data.