Passive Target Detection And The Research Of The Automatic Judgement Expert System

Automatic judgement of underwater targets is one of the key techniques in modern passive sonar systems. The covert characteristic of passive sonar makes the submarines can get the target information without being detected easily by the enemies' detection equipment. The information includes target azimuth, distance, speed, type, etc. For judging the target automatically, it needs to build up the expert system for passive target judgement, which can replace the artificial judgment, at the base of extracting much judgement feature of the target. Then put the target information from feature extraction into the system for judgement, and will as a result give manifestation.The method of expert system judgement is building up the inference judgement system according to the field expert's experience knowledge. The knowledge base that set up has certain catholicity and representative. In judgement system, send the target characteristic that is from feature extraction into inference engine, then the inference engine analyzes according to the person's thinking and combine with the condition in knowledge base to get the judgement result.The passive sonar target judgement is mainly realized according to the dissimilarity of target radiation noise. How to extract the effective judgement feature from the target radiation noise is the key of the passive sonar target judgement. There is much information about the targets in the envelope spectrum of the signals received by the passive sonar system. In this paper, through analyzing the envelope spectrum, separate the line spectrum, from which detect the target base-frequency related with the rotate speed of ship screws. At the same time can get the target azimuth, etc. Then send the information into the expert system for judging automatically, and reasonably display the judgement result.