提出了基于双重约束函数的无重叠图像期间的颤振估计算法。建立了视差图像与低频颤振相结合的双重约束函数,以评价无重叠图像期间颤振曲线的平滑度;采用共轭梯度法筛选颤振曲线平滑度最优解,并以此作为无重叠图像期间颤振的最佳估计;建立了无重叠图像期间与重叠图像期间颤振之间的函数关系式。以我国XX-1号可见光详查相机为例进行实验,颤振探测结果在被测颤振的频率既不等于也不接近特征频率的整数倍处具备有效性,证明了所提算法能够对无重叠图像期间非特征频率整数倍及其附近处的颤振进行有效探测。

In this study, a jitter estimation algorithm during the acquisition period of non-overlapping images was proposed based on a double-constraint function. The double-constraint function was derived by combining the parallax-observation images with the low-frequency jitter to evaluate the smoothness of the jitter curve during the acquisition period of non-overlapping images. The conjugate-gradient method was used to find the smoothest curve, and the corresponding solution was used as the optimum estimation of jitter during the acquisition period of non-overlapping images. The quantitative relation between the jitter during the acquisition period of non-overlapping images and that during the acquisition period of overlapping images was established. The experimental results using the Chinese XX-1 space camera show that the jitter detection results are only valid when the measured jitter frequencies are neither equal nor close to any integer multiple of the characteristic frequency. This in turn proves that in such cases, the proposed algorithm can effectively detect the jitter during the acquisition period of non-overlapping images.