The Algorithm Study Of Filter Optimization Based On Target Spectrum Infrared Feature Extraction

As long as the temperature of object is higher than absolute zero, it will emit infrared radiation all the time which carries many characteristic information of the object. This provides an effective method for the detection and identification of target. In this paper, characteristics of target are calculated by inversing infrared spectrum radiation characteristics of target. The main research contents are as follows:(1) The infrared radiation characteristic analysis of target. Here we introduce the infrared radiation characteristics of the target. The infrared radiation characteristics of the target mainly include two aspects: First is the infrared radiation of target, which is mainly decided by the target’s own characteristics of area and temperature. Second is the radiation of other object that target reflects, which includes the radiation of sun, the radiation of earth and the reflective radiation of the earth.(2) The spectral signal detection and labeling of target. Firstly, the infrared radiation characteristics of target vary with the target’s wavelength. The relationship between the infrared radiation characteristics of target and wavelength is analyzed. And the best band for feature extraction is found. Secondly, the infrared radiation signal of target not only contains the effective signal of target, but also contains a lot of noise. So we need to detect the signal, and mark the effective signal. In this paper, the threshold processing algorithm and region growing algorithm are used to detect and mark the signal of target.(3) The infrared feature extraction of target. Firstly, the mathematical model of target feature extraction is established, and the mathematical parameters which are necessary for feature extraction are calculated. Secondly, the theoretical basis of optimization(OPT) algorithm is analyzed, which is used to extract the equivalent characteristic of target. And the noise information is analyzed. Thirdly, the Calman filter(KF) theory is researched, which is used to extract the equivalent characteristic of target. And the noise information is also analyzed. In the end, two methods are contrasted, and their advantages and disadvantages are analyzed.