نوع مقاله : مقاله پژوهشی
نویسندگان
1 دانشجوی دکتری، دانشگاه خواجه نصیر الدین طوسی، تهران، ایران
2 دانشیار،دانشگاه خواجه نصیر الدین طوسی، تهران، ایران.
3 دانشیار، دانشگاه امام حسین علیه السلام، تهران، ایران
چکیده
کلیدواژهها
موضوعات
عنوان مقاله [English]
نویسندگان [English]
Target finding and pattern recognition systems are systems that have defense and security applications in military fields. The most important advantage of using these systems is eliminating the role of humans in identification processes, such as tanks, cars, ships, helicopters, etc. In the pattern recognition system, the input image is obtained by one of the imaging sensors such as millimeter wave radar, laser radar, video camera or infrared camera, and after initial preprocessing, feature extraction and feature selection and finally classification are done. In this article, an effective method for identifying the noise pattern of helicopters based on the area feature vector and weighted sparse representation classification is introduced. The proposed method includes three steps: preprocessing, identification and classification. In the preprocessing stage, changes are made by applying processing algorithms in order to improve the quality of received images and remove irrelevant data (noise). Then, in the identification stage, a 32-component feature vector is considered based on shape, surface and length features, which in the method presented in this article, only surface features are used and the shape and length features are discarded due to lack of efficiency, and finally, in the third stage, thin weighted representation is used for classification. Applying the above three steps leads to reducing the time of the algorithm and increasing the accuracy of the method in identifying helicopters. To check the performance of the proposed method compared to other methods, the database of 60 different images of helicopters was examined and the proposed method achieved the highest recognition rate of 96.3%. On the other hand, the presented method has the least time complexity among the methods, which indicates its high speed.
کلیدواژهها [English]