Font Size: a A A

Research On Detection Of Infrared Small Target Under Complex Backgrounds

Posted on:2011-09-05Degree:MasterType:Thesis
Country:ChinaCandidate:J YaoFull Text:PDF
GTID:2178330338476193Subject:Measuring and Testing Technology and Instruments
Abstract/Summary:PDF Full Text Request
Nowadays, with the rapid development of high-tech weapons, kinds of weapons systems based on infrared detection have became one of the key studying and developing projects around the world. Detecting and tracking infrared target has been a technical bottleneck problem to be solved. Therefore, it's significant for improving the performance of weapon systems as soon as possible to obtain information about incoming targets.Firstly, the paper analyzes the existing multi-spectral imaging principle, the advantages and disadvantages of existing imaging equipments. A novel design, the imaging focal plane which can obtain image of multiple frequency bands simultaneously, is proposed. After the processing of a multispectral image, a higher contrast target image is obtained to facilitate detection and tracking in the next stage.Secondly, the paper studies the pre-processing algorithms of infrared small target image in the space and frequency domain. Due to the lack of a perfect method which can preprocess image in different background under the same framework, this paper studies the characteristics of singular value decomposition(SVD) of infrared small target and analyzes the relation among the point target's singular values, location, size and intensity. It is derived that small target in infrared image is a composition of part of the singular values and the corresponding U and V matrix. In order to extract the small target's singular values for reconstructing image, we proposed a method of reconstructing infrared small target which based on differential percentage curve of the difference of singular value prediction curve and the original of the singular value curves. Experiments show that the method can improve image signal noise ratio and suppress background clutter.Lastly, considering the small target extraction algorithm of wavelet energy combination can only carry on the horizontal and vertical direction, a new small target detection method, which is based on NSCT mutual energy combination, is presented. The results show that the new method can detect small targets and improve the detection rate.
Keywords/Search Tags:infrared small target, image preprocessing, singular value decomposition and reconstruction, target detection, NSCT mutual energy combination
PDF Full Text Request
Related items