Research On Color Correction And Instrument-side Calibration For Moon Monitor Camera Based On Meachine Vision

In view of the imaging conditions and mission characteristics of the lander’s monitor camera in the sampling package and other key processes of lunar exploration,we carried out in-depth research on method of color correction of monitor color camera,and the binocular vision system calibration about the internal and external parameters,to meet the requirements of imaging quality and three-dimensional reconstruction accuracy about moon topography model.The specific research work and results obtained in this paper are shown in the following aspects:Firstly,the present technology development situation of color monitor camera was summarized according to the current lunar exploration projects all over the world.We introduced the operating characteristics and imaging task of the monitor camera of CE-5lander.The research status of the color camera’s color correction and instrument-side calibration of binocular vision system about the internal and external parameters were analyzed in detail,and the overall structure of this paper was given also.Secondly,the color correction methods for the situation of illumination imitating lunar surface were studied.The conversion model of camera response RGB values to standard XYZ values was respectively established by least square regression,Lasso regression and Ridge regression,using linear polynomial and quadratic polynomial as initial fitting models.The color correction experiment was conducted under D65 standard light source.The correction efficience of each algorithm were tested by CIE 1976 Lab color difference formula.The results show that Lasso regression and Ridge regression with regularization item are feasible and have higher accuracy.Thirdly,the color correction methods under different illumination conditions were discussed.For the case of variable illumination,a high order-homogeneous mixed polynomial model were proposed in basis of homogeneous polynomial.And we also proposed an algorithm that owes variable exponent regularization constraint item.The color correction experiment was conducted in different illuminations by change exposure time.The experimental results verifies the better correction effect and adaptability to variable illumination intensity of the algorithms.Fourthly,aiming at the instrument-side calibration requirement of monitor camera,a two-steps geometric parameters calibration scheme was put forward based on Zhang Zhengyou calibration method.The scheme includes extracting harris corners of the images,establishing linear equations to initially calibrate the internal and external parameters of the camera,and adding nonlinear model to calibrate the camera’s distortion and orther parameters exactly.With the calibration result,the relative position of binocular cameras can be established and the binocular vision system calibration is completed.According to the scheme,calibration experiment was carried out using chessboard template.The result manifests that the scheme is simple,convenient and can improve the calibration accuracy.Fifthly,the work done in this paper was summarized,and the development of camera’s color correction and instrument-side calibration technology was prospected.