论文部分内容阅读
结合航拍视频序列的特点与视觉导航系统的需求,提出了一种基于图像区域相似性和禁忌搜索算法的地平线拟合视觉导航算法.从航拍视频序列中提取单帧图像,并做下采样去噪等相应预处理;初始化禁忌搜索算法的初始解等参数,结合适配值函数,拟合图像中的地平线;用禁忌搜索算法的特赦准则和收敛准则评判拟合的准确性,直到适配值函数值满足收敛准则为止,得出最优解;分解出飞行器的横滚角和俯仰角,并由横滚角和俯仰角的大小决定下次参数校正的间隔,输出结果给控制系统.实验结果表明,与传统视觉导航方法和基于Hough变换的方法相比,此算法具有无参考、实时性好和抗噪性强等特点.
Combining with the characteristics of aerial video sequences and the demand of visual navigation system, a horizon-fitting visual navigation algorithm based on similarity of image regions and tabu search algorithm is proposed.A single frame image is extracted from aerial video sequence and downsampled denoising Such as the initial solution of the tabu search algorithm and other parameters, and fitting the fitness function to fit the horizon in the image; using the tabu search algorithm’s amnesty criteria and convergence criteria to judge the accuracy of fitting until the fitness function The value of the convergence to meet the criteria until the optimal solution; Decomposition of the aircraft roll angle and pitch angle, and the size of the roll angle and pitch angle determines the next parameter correction interval, the output results to the control system .Experimental results show Compared with traditional visual navigation methods and Hough transform-based methods, this algorithm has the characteristics of no reference, good real-time performance and strong anti-noise.