Algorithm of global image motion detection based on remote images and engineering parameters

Identical scenes in images taken by two adjacent imaging sensors' splicing areas are essential for detecting image motion. However, images with identical scenes are not always present throughout an imaging mission, and it is still a question to detect the image motion for the period with images without identical scenes. In this study, an algorithm based on remote images and engineering parameters is proposed to detect image motion throughout an imaging mission, including the period with images without identical scenes. According to the imaging characteristics of the splicing areas, an imaging mission is divided into two periods: the period with images without identical scenes and that with identical scenes. For the first period, low-frequency image motion is first calculated by using engineering parameters, and an objective function based on the low-frequency image motion and images from slicing areas is then established to measure the deviation between image motion curve and the low-frequency component during the period without identical scenes. Steepest descent method is adopted to find the minimal solution which is taken as the optimal estimation of the image motion for the period without identical scenes. A formula is derived to calculate the image motion for the period with identical scenes by the use of the estimated image motion of the period without identical scenes. Experiments performed on XX-1 space optical remote sensor show that an image motion of about 7 pixels at 0.133 Hz is captured, and its value during the initial 226 ms, the period with images without identical scene, is detectable. Besides, the deviation at blind frequencies is reduced significantly, proving that the proposed algorithm can effectively measure the global image motion, including that during the period with images without identical scenes.