Reserch On Dynamic Range Improvement And Image Optimization Algorithm Of Rocket Visible Light Imaging System

The visible light imaging system has been widely used in the field of domestic and foreign launch vehicles,which is used to observe the rocket booster separation,separation between stages,engine ignition,fairing separation,star(ship)arrow separation,fault location and so on.However,the particularities of rocket camera environment,such as wide dynamic range,high reliability and wide range of environmental temperature,etc.These characteristics lead great difficulties to the design of the visible light imaging system.As the main part of a research project,this dissertation is mainly concerned with dynamic range improvement and image optimization algorithm of rocket visible light imaging.The author's major contributions are outlined as follows:(1)The dynamic range model of imaging system is proposed.The model is applied to analyze the related factors that affect the dynamic range of the imaging system,such as the sensor noise,AD conversion noise,analog gain,digital gain,and mapping.The related mathematic model is established.The model can guide researchers to design high-performance dynamic range imaging systems.(2)The hardware and software extension method of dynamic range is studied deeply.In view of software,a fast image fusion algorithm based on BAYER is presented.For establishing the database of image processing,compared with the traditional RGB algorithm,this algorithm improves the standard deviation of 1.5%,the spatial frequency is increased by 0.33%,and the information entropy is increased by 0.40%.In the aspect of hardware,a photoelectric system platform is designed.The dynamic range of the platform is extended using multiple slope integration and single frame exposure multi frame readout.(3)For color image restoration,a hybrid algorithm is proposed.A new interpolation operator based on color difference gradient,which has good signal-to-noise ratio,is employed for the hybrid algorithm.The average signal tonoise ratio of the hybrid algorithm is 38.76 dB,while the average signal-to-noise ratio of ACPI is 36.88 dB,and the average signal-to-noise ratio of BI is 30.06 dB.The experimental results show that the algorithm has good color recovery effect for the BAYER image of the missile borne image pickup device.(4)According to the experimental system,the noise characteristic of the platform in the high temperature environment is analysed.Through histogram analysis,the noise is more close to the Rayleigh distribution.Thus,a wavelet domain maximum a postrerior image denoising algorithm based on Rayleigh noise is proposed.For "Lena" and "Peppers" denoising,the proposed algorithms for PSNR are 30.11 dB and31.80 dB respectively,which are better than bilateral filtering algorithms,PSNR of27.38 dB and 28.44 dB.Finally,the algorithm is applied to the actual verification platform,and a better denoising effect is achieved.