Noise Evaluation Study Of Fusion Imaging Systems

Image fusion technology is widely used in military,medical and other fields,it can effectively combine the characteristics of multi-spectral images to achieve the goal of scene information complementation.The development of the study on the algorithm and fusion image systemin recent years is rapid,the objective evaluation of fusion image quality is an important branch of the field,which can effectively promote the improvement and inspiration of related algorithms,but there is little research of fusion system evaluation which focus on noise.We put forward a kind of fusion test of the characteristics of noise before and after.According to the infrared and LLL image fusion process,which is important to the fusion image evaluation.Firstly,the FPGA board is designed combining the functional requirements.Noise theoretical research in the fusion process of infrared and LLL image was studied.Mathematical modeling was established according to different types of noise such as thermal noise,quantum noise,non-uniform noise.Secondly,several typical noises were abstracted to random sequence of a certain distribution or the coupling of several random number.An improved scheme of uniform random sequence based on FPGA was proposed based on the study of statistical characteristics of several random sequences and the existing simulation method.The simulation results were good compared with the actual image noise.Thirdly,the auxiliary modules such as USB3.0 communication and PAL/Camera-link output were designed to make the system function more complete.Then a set of image evaluation index calculation software was designed by the GUI tool in MATLAB including the calculation of the peak signal to noise ratio,mean,etc.A complete test system of the noise characteristics of the fusion board was built by system debugging.Finally,the results of the noise test and the calculation of related index were analyzed to verify the reliability and practicability of the system.The exploration of objective evaluation method in the design of hardware system was completed through the experiment according to the proposed method.