无人直升机自主着舰视觉特征提取算法研究

摘要

针对无人直升机自主着舰计算机视觉系统图像特征提取困难,在图像处理方面提出一种改进的拉普拉斯算子边缘检测算法,在抑制无关信息的同时可以有效提取设计图形中的边缘信息,克服了拉普拉斯算子对噪声敏感的不良特性.对无人直升机着舰图像处理过程进行了数字仿真,建立了整套系统(含摄像头成像模型)的数字仿真模型.仿真结果表明,所提出的解决方案和算法有效抑制了无关信息,正确提取到预先设计的图形边缘.%An improved Laplace algorithm for image edge detection was designed to overcome the difficulty in a computer vision system for unmanned helicopter's landing on a moving ship. The traditional Laplace algorithm is too sensitive to noise or irrelative edges. The improved algorithm restrains useless information and extracts the edges of pre - designed pattern. The computer vision process of unmanned helicopter's landing on a ship was modeled and simulated. The simulation proves that the edges of the designed pattern can be detected correctly and other useless edges can be restrained.