Radar waveform classification method and system based on effective area identification

Abstract

The invention provides a radar waveform classification method and system based on effective area identification, and the method comprises the steps: inputting a radar radiation source time-frequency image into a trained preset SSD network, and obtaining an effective area where a single intra-pulse signal outputted by the preset SSD network is located and a corresponding radar waveform type. The preset SSD network is trained by taking a plurality of radar radiation source time-frequency images as training samples and taking an effective area where a single intra-pulse signal in each radar radiation source time-frequency image is located and a corresponding radar waveform category as labels. According to the method, the effective area where the intra-pulse signal of the time-frequency image of the radar radiation source is located is identified by using the SSD network, and only the types of the waveforms in the effective area are classified subsequently, so that the influence of signal features in an invalid area on identification is avoided, and the identification accuracy of the radar waveforms is effectively improved.

Description

Technical field [0001] The present invention relates to the field of image processing, and more particularly to a radar waveform classification method and system based on an effective area identification. Background technique [0002] With the wide use of radars in military and civilian fields, radar confrontation has become a major reconnaissance form in electromagnetic spectrum war. The radar waveform classification is an important part of the radar against information processing. It is one of the key technologies of radar against reconnaissance systems. The result is an important basis for judging the extent of enemy weapons threats. [0003] The current image classification method is currently used in the radar waveform classification, and the radar radiation source signal is used to time-frequency image, and the waveform classification is completed by using the CNN to extract the arranging signal characteristics. However, the electromagnetic signal intensive in the actual sp...

AI Technical Summary

Problems solved by technology

However, electromagnetic signals are dense in the actual space, especially the radar signals of the new system are difficult to intercept, and it cannot be guaranteed that the identified object must be a complete intra-vein signal, and the intra-vein signal may be incomplete.

If features are directly extracted from the entire time-frequency image, it is susceptible to interference from invalid signal features, resulting in insufficient focus of the recognition area, which greatly increases the number of image features to be recognized, which in turn affects the accuracy of radar waveform classification, and may affect the commander's subsequent decision-making deployment. Therefore, the above problems have greatly restricted the radar waveform classification performance, and further research is needed on the actual space radar waveform classification method to improve the recognition accuracy

Images