Extracting the centroid from the sun-moon image in astronomy orientation

In astronomical orientation, we usually choose the sun, the moon as observation objects, using the theodolite to obtain the horizontal and vertical scale readings as the crosshair theodolite and astral edge are tangent. This operation is subjective to human disturbance, affecting the orientation precision afterwards. Comparing to the previous method, the method of image processing is much simple in operation, and reduces the manual intervention as long as the sun and the moon are within the scope of the CCD imaging. After imaging, we first extract the object centroid coordinates from the embedded image module, and then according to the imaging model, we get the angle of visual axis interacts with the center line of the stars. This article analyzes the main characteristic of the sun-moon image and discusses the various circumstances image centroid algorithm for fitting. The algorithm is able to separate the sun and moon from complex background environment. Besides, even in the cases of the poor quality of the images, such as stars appear some incompleteness, low star brightness, or fuzzy boundaries, the algorithm can still locate the edge effectively with better fitting.