降低无人机影像冗余方法实验研究

EXPERIMENTAL STUDY ON REDUCING IMAGE REDUNDANCY OF UAV

针对无人机原始影像重叠度过高,影像幅数过多导致处理过程耗时太大,在不影响数据质量的情况下,如何降低无人机影像冗余是实验研究的问题,利用固定间隔抽稀和基于空域无参考图像质量评估的降冗,提出先利用固定间隔抽稀的方法,发现虽然能够大幅度减少影像处理时间,减少了近半的处理时间,但是最后处理结果相比于全部影像参与处理平面位置的中误差偏大,因此提出在固定间隔抽稀的基础上,进一步应用基于空域无参考图像质量评估的降冗方法,能在提高处理效率节省时间的同时处理结果的误差得到良好的控制。

In view of the high overlap of UAV’s original image and the excessive number of images,the processing time is too long.How to reduce the UAV redundancy without affecting the data quality is the problem.This Experiment introduces the fixed interval thinning and the method of reducing redundancy based on the evaluation of UAV’s reference image quality in the airspace.First,the method of fixed interval thinning is used,and it is found that although it can greatly reduce the processing time.Reduce the image processing time,reduce nearly half of the processing time,but the final processing results are more error than all the images involved in processing plane position.Then,based on the fixed interval thinning,a method of reducing redundancy based on the spatial domain no reference image quality evaluation is proposed,which can improve the processing efficiency and save time,while the error of processing results is well controlled.