Research On Jitter Processing Methods For High Resolution Optical Satellite Imagery

High-resolution optical satellite (HROS) is one of the most important components of the Earth Observing System. High accuracy geometric processing is the key and foundation of HROS imagery for the national development strategy and social economy construction applications. Satellite platform jitter, a microvibration of a satellite platform, is commonly induced by the periodic motion of dynamic structures on board, the changes in the body temperature, the operation of the orbital control, etc. Considering that most of the optical imaging sensors onboard satellites adopt the pushbroom linear array CCD to capture images, platform jitter will directly induce the distortion to the images, decreasing the image accuracy and affecting the following high accuracy remote sensing applications, such as stereo mapping and image fusion. Currently, China has dozens of in-flight high-resolution satellites, such as ZiYuan series, GaoFen series, TianHui series and YaoGan series. In the future, China will launch tens or hundreds of high-resolution satellites, and the ground sample distance (GSD) is expected to reach 0.1m, which means the influence on image geometric accuracy caused by satellite jitter becomes more and more obvious. Therefore, it is urgent to carry out the research on high-resolution satellite platform jitter processing method to solve the problems caused by platform jitter so that HROSs can be used widely and efficiently. To solve the high-resolution image geometric accuracy problem caused by satellitejitter, this paper investigates the mechanism how geometric accuracy of HROS images is influenced by platform jitter. Then, satellite jitter detection and modeling method based on parallax imaging and jitter distortion correction method based on virtual steady reimaging (VSRI) are proposed. Both of the two methods are verified by simulation data and real data. The research content and innovation of this paper is summarized as follows: (1) The mechanism of influences on geometric accuracy of HROS imagescaused by platform jitter is investigated, which lays the theoretical foundation for the platform jitter detection and compensation. The mechanism of influence on imaing geometry caused by satellite jitter isanalyzed through the theory analysis and simulation experiment. First, the characteristic analysis of satellite jitter and the camera design and imaging feature of HROS are introduced. The rule of imaging motion of cameras using TDI CCD as core component caused by satellite jitter is analyzed, which establishes the theoretical foundation for the precise jitter recovery with TDI CCD imaging. For imaging sensors with parallax observation, such as multispectral camera, non-collinear TDI CCD camera and stereo cameras, the influence on imaging parallax caused by satellite jitter is analyzed to establish the theoretical foundation for the following platform jitter detection and modeling method based on remote sensing images.(2) Platform jitter detection and modeling method based on parallax imaging is proposed, which provides a new approach for the measuremnt of platform jitter.For pushbroom sensors with parallax observation, such as multispectral camera and non-collinear CCD camera, the coordinate differences of corresponding points among parallax imaging images show linear error when platform jitter does not exist. Otherwise, the coordinate differences show nonlinear error. So satellite jitter detection and modeling method based on parallax observation is proposed. First, the relative error curves of parallax images are obtained by using high accuracy dense matching. The dense corresponding points of parallax images are obtained accurately by using least squared sub-pixel matching. The coordinate differences of corresponding points are calculated, and the average of the coordinate differences in the same line is taken as the error of this line, which is then used to obtain the relative error curves. Then the quantitative relation between the relative registration error of parallax imaging images and the absolute displacement caused by platform jitter is determined based on sine function by the strict formula derivation. So the absolute error curves caused by platform jitter are determined with the derivation model when the relative error curves are obtained. Finally, the jitter component of satellite attitude is obtained according to the camera parameters, which provides an important jitter observation data for the following processing.(3) Jitter distortion correction model and method based on virtual steady reimaging (VSRI) is proposed, which can effectively improve the internal geometric accuracy of optical remote sensing images.Based on the precise recovery of jitter attitude, attitude interpolation methods under the condition of platform jitter are analyzed. Five different attitude interpolation models, including the cubic polynomial model (CPM), linear piecewise polynomial (LPPM), quadratic piecewise polynomial model (QPPM), cubic piecewise polynomial model (CPPM) and Lagrange polynomial model (LPM), are compared and analyzed with experiments. The VSRI based jitter distortion correction method is further proposed. The reimaging mode is applied to translate images that imaging under the condition of platform jitter into images that imaging under steady state. What’s more, the corresponding rational function model (RFM) can be also obtained accurately, which provides images with high accuracy for the following applications.To verify the proposed methods, simulation images and real images including ZY-3 three-line and multispectral imagery, GF-8 panchromatic imagery and GF-9 panchromatic and multispectral imagery are utilized. The results show that the proposed jitter detection and modeling methods can accurately recover jitter component of satellite attitude by using multispectral images. By using the proposed jitter distortion correction method, the geometric accuracy of high-resolution remote sensing images is effectively improved, and the internal distortion caused by platform jitter has been eliminated.