| The detection and recognition of target are the key problems that various countries have been trying to solve. With the rapid development of space exploration technology, many countries have been able to achieve the goal of accurate identification. However, the recognition efficiency is not high. So, in order to shorten the recognition time, they all expect that the real target could be identified as much as possible in the far distance. At this time, due to far from the detector, goals are dots with no shape information. The traditional image detection, recognition and tracking technology is no longer valid. And in the actual cases, the targets are accompanied by a lot of decoys with similar characteristics around. For the point targets and decoys far away, no matter visible light detection or radar detection has lost its advantage. But using infrared detection technology can effectively distinguish them. With the development of photoelectric technology, decoys can be made very similar to the real targets, so distinguishing between the target and decoy is becoming more and more difficult. As a result, only using radiation information acquired from single-band infrared detector can not accurately distinguish the two. Therefore, in order to improve the detection and recognition capability of point target, countries begin the study of multi-band target detection and recognition. By the comprehensive utilization of multiple wavelengths radiation information, the target and decoy can be distinguished.Based on this, the infrared radiation characteristics of the point target and accompanying decoy are analyzed in this paper. Then radiation intensity and radiation spectrum difference between them are discussed. Finally, useful recognition characteristics are extracted and classification identification algorithms are designed. The research work of this paper mainly includes the following aspects:(1) The temperature characteristic of target: the models of target and decoy are established; the time-space relationship and heating of target are analyzed; the heat balance equation of target is built and solved; the temperature characteristics of target and three types of decoys are compared.(2) The infrared radiation characteristics of target: the target radiation received by detector is calculated in the ideal case; the influence of outside radiation reflected by target to total target radiation received by detector is discussed; the atmospheric attenuation of target infrared radiation is analyzed; the influence of different material emissivity to target infrared radiation is studied.(3) The band characteristics analysis of target: the target and decoy radiation received by detector are compared; the control variable method is used to discuss the influence of sun position, orbital shape, orbital inclination, case thickness ratio, surface coating, shell material and internal heating to target and decoy radiation characteristics and difference characteristics; the band studied is divided based on the difference characteristics of target and decoy.(4) The multi-band processing of target: based on the band radiation characteristics and band radiation variation rate characteristics of target and decoy, two kinds of recognition algorithms based on artificial statistics are put forward; the target and decoy are effectively identified by using Fisher discriminant analysis method and back propagation neural network method; the recognition algorithms of single band, dual bands and three bands are compared. |
PDF Full Text Request