中文摘要：

为了进一步提高遥感图像配准精度，提出了尺度不变特征变换（SIFT）结合区域互信息优化的遥感图像配准方法．首先利用混沌序列的随机性和遍历性，提出一种混沌量子粒子群优化（CQPSO）算法，在量子粒子群优化（QPS01算法迭代陷入早熟收敛时，采用一种新的机理引入混沌序列，进化粒子克服早熟．图像配准算法分为预配准和精配准两个过程．基于SIFT算法提取特征点，经匹配和有效地外点排除完成预配准，然后对匹配特征点坐标进行亚像素级微调，通过最小二乘法求得一系列匹配参数构造初始粒子群，最后利用混沌量子粒子群优化区域互信息完成精配准，得到最优匹配参数．用一些标准测试函数对所提出的CQPSO和QPS0及粒子群优化（PSO）算法进行了实验比较，另外，对SIFT，SIFT结合PSO算法优化区域互信息，SIFT结合QPSO算法优化区域互信息和SIFT结合CQpSO算法优化区域互信息（SRC）等四种算法进行了不同分辨率遥感图像配准实验比较和不同时相遥感图像配准实验比较，实验结果验证了所提出的CQPSO算法的优越性和SRC配准方法的有效性．

英文摘要：

In order to further improve the precision of remote sensing image registration, we propose a new registration scheme by combining the scale-invariant feature transform （SIFT） and the optimization of regional mutual information in this paper. Firstly, taking advantage of the randomness and ergodicity of chaotic sequence, we present a new chaos quantum- behaved particle swarm optimization （CQPSO） algorithm to solve the premature convergence problem of the quantum particle swarm optimization （QPSO） algorithm. By taking full account of the quantity differences among the values of different dimensions for the particle location information, small disturbances are generated as the Hadamard product of chaotic sequence and the particle location information. Before being added to the particle location information, the small disturbances are adjusted by an evolutionary parameter to ensure that each new particle location information is within the scope of reasonable evolution. The image registration scheme consists of two processes, namely the pre-registration process and fine coregistration process. The pre-registration process is implemented by the SIFT approach with a reliable outlier removal procedure. By the repetitive fine-tuning of several selected matched feature point coordinates, a series of registration parameters is estimated by a least square method and used to construct initial particle swarms. Next, the fine coregistration process is implemented to obtain the optimal match parameters by maximizing regional mutual information based on CQPSO. The proposed CQPSO algorithm is tested on several benchmark functions and compared with QPSO as well as standard PSO experimentally. Furthermore, comparative experiments are carried out on the registration of remote sensing images with different ground resolutions and the registration of remote sensing images at different phases by using four algorithms： the SIFT algorithm, SIFT combined with PSO algorithm, SIFT combined with QPSO algorithm, and SIFT combined