论文部分内容阅读
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。
In order to solve the problem of frame loss in hyperspectral image storage under the condition of laser beam interference, a new method of storing hyperspectral image data based on sparse representation and queue-based buffer phase structure is proposed. Firstly, the Gabor filter bank is used to filter the interference of the laser beam on the hyperspectral image and improve the recognizability of the spectral image. The training sample set is constructed based on the sparse representation principle, and the hyperspectral image data is sparsely coded by using the basis function dictionary. Complete the image compression processing; Finally, based on the queue cache structure to eliminate the adverse effects of fluctuations in storage speed to achieve real-time storage of hyperspectral image data. Experiments show that the proposed method of storing hyperspectral image data can effectively improve the storage speed and solve the problem of data loss in image storage.