基于LOG算子的无人机模糊影像快速剔除算法研究

Research on fast elimination algorithm of UAV fuzzy image based on LOG operator

无人机在复杂地形区域获取影像的时候,容易受到气流的影响造成飞行平台的抖动,获得的影像出现模糊、不清晰等现象。模糊影像传统的剔除方法是采用人工检索的办法,工作效率较低,耗费时间长。影像的清晰度与影像的边缘特征具有很大的相关性,论文利用高斯-拉普拉斯算子(LOG算子)提取影像的边缘特征,通过计算影像的方差,设置阈值可以自动的进行模糊影像的剔除。结合工程实例进行验证,利用该方法进行的模糊影像识别准确率可以达到100%。

When the UAV acquires images in complex terrain,it is easy to be affected by the air flow and cause the flight platform to shake.Thus the obtained images appear blurred,unclear and so on.The traditional method of eliminating fuzzy images is manual retrieval,which is inefficient and time-consuming.The definition of the image has great correlation with the edge feature of the image.In this paper,the Gao Si-Laplace operator(LOG operator)was used to extract the edge feature of images,and the variance of images was calculated.The fuzzy image can be eliminated automatically by setting threshold,which was verified by engineering examples.The accuracy of fuzzy image recognition can reach 100%by using this method.