Measuring dispersed spot of positioning CMOS camera from star image quantitative interpretation based on a bivariate-error least squares curve fitting algorithm

Abstract

For a positioning CMOS camera, we put forward a system which can measure quantitatively dispersed spot parameters and the degree of energy concentration of certain optical system. Based on this method, the detection capability of the positioning CMOS camera can be verified. The measuring method contains some key instruments, such as 550mm collimator, 0.2mm star point, turntable and a positioning CMOS camera. Firstly, the definition of dispersed spot parameters is introduced. Then, the steps of measuring dispersed spot parameters are listed. The energy center of dispersed spot is calculated using centroid algorithm, and then a bivariate-error least squares curve Gaussian fitting method is presented to fit dispersion spot energy distribution curve. Finally, the connected region shaped by the energy contour of the defocused spots is analyzed. The diameter equal to the area which is 80% of the total energy of defocused spots and the energy percentage to the 3×3 central area of the image size are both calculated. The experimental results show that 80% of the total energy of defocused spots is concentrated in the diameter of the inner circle of 15μm, and the percentage to the 3×3 pixels central area can achieve 80% and even higher. Therefore, the method meets the needs of the optical systems in positioning CMOS camera for the imaging quality control. © 2017 COPYRIGHT SPIE.