摘要
针对激光束干扰情况下,高光谱图像存储过程中存在的丢帧问题,提出基于稀疏表示与队列式缓存相结构相结合的高光谱图像数据存储方法研究。首先利用Gabor滤波器组,滤除激光束对高光谱图像的干扰,提高光谱图像的可识别度;基于稀疏表示原理构造训练样本集合,利用获取的基函数字典对高光谱图像数据进行稀疏编码,完成对图像的压缩处理;最后基于队列式缓存结构消除存储速度波动带来的不良影响,实现高光谱图像数据的实时存储。实验证明了提出高光谱图像数据存储方法,能够有效提高存储速度,并解决了图像存储过程中的数据丢失问题。
For the frame losing in the storing process of the high spectral image, this paper presented a research of the high spectral image data storage based on the sparse representation combined with queued buffering. First, filter the interference of the laser beam by using Gabor filter group to improve the recognition of the high spectral image. Based on the principle and construction of sparse representation, and make the ample set, Sparse code the high spec- tral image by using the acquired basis function dictionary to finish the image compressing. Finally, the side effect from velocity fluctuation of storage was eliminated by the queued buffering, it achieved real time storing of the high spectral image data. The experimental results show that the proposed method improved the storage speed efficiently and solve the frame losing in image storing.
作者
路芳
LU Fang(School of Mathatical science, Baotou Normal University,Baotou 014030,China)
出处
《激光杂志》
北大核心
2017年第4期71-74,共4页
Laser Journal
基金
内蒙古教育厅项目(NJSG201282)
关键词
激光束
高光谱图像
数据存储
laser beam
hyperspectral image
data storage
