Design Of Radar Target Recognition Software Based On Subspace And Neural Network

Due to the advantages of the penetrating in cloud, working in day and night and high resolution, the synthetic aperture radar(SAR) technique has great prospect in the fields of the earth observation and militaryreconnaissance.In the national economy community, the change of the ground targets or the environmental change can be monitored by usingSAR technique. Therefore, the SAR techniquecan play an important role in the agricultural production, environmental monitoring, resource prospection and disaster prevention and mitigation. In the national defense buildups, because SAR can provide the high resolution image of the ground moving targets or military targets, it has great applicationprospects in the battle reconnaissance and military strike assess.The paper aims to analyze and investigate the recognition and classification methods of the typical targets on ground based on SAR image and build the data of the typical targets samples. Based on the above, we can achieve the feature extraction and automatic target recognition(ATR) of the SAR image.The results of our research can be applied to the satellite and airborne SAR systems. Its successful development will improve the application level of these systems and the performance of the information acquisition and signal processing of the satellite and airborne SAR image. In addition, this research outputs can increase the reliability of target information, decrease the ambiguity of target or incident and improve target or incident detectivity. The real time and complete assess results of battlefield situation and threat can be achieved by using the outcomes of this study. The results of this paper are also applied to the national economy field, such as urban planning, resource administration, crop monitoring and so on.