To overcome the dramatically increasing data amount of space remote sensing image, an improved back-propagation(BP) neural network was put forward to compress it. The algorithm used the Levenberg-Marquart algorithm to improve the convergence speed of neural network and used algorithm to improve the generalization ability of neural network. We compared and analyzed the compression result and error performance function of the improved algorithm and the standard BP algorithm to the same image. The experimental results show that, when the image compression ratio is 1/2, the mean square error (MSE) of standard BP algorithm is 343.3750; for improved BP algorithm, the MSE is 69.5796 when the image compression ratio is 1/16, the MSE is 20.9561 when the image compression ratio is 1/8, and the MSE is 5.5123 when the ratio is 1/4. Moreover, the peak signal-to-noise ratio (PSNR) of the improved algorithm is always in the range from 30 dB to 40 dB. The improved algorithm has been applied in practical engineering, which meets the need of practical work.
LI Ji.
Image compression used improved error back-propagation neural network. Journal of Applied Optics. 2013, 34(6): 974-979
{{custom_sec.title}}
{{custom_sec.title}}
{{custom_sec.content}}
参考文献
[1]霍承富.超光谱遥感图像压缩技术研究[D].安徽:中国科技大学,2012. HUO Cheng-fu. Research on hyperspectral remote sensing image compression technique[D]. Hefei:University of Science and Technology of China, 2012.(in Chinese) [2]SAID A, PEARLMAN W A. An image multire solution representation for lossless and lossy compression[J]. IEEE Trans Image Processing, 1996,5(9):1303-1310. [3]赵米阳,陈卫东,卢晓燕. 基于SPIHT的改进图像压缩算法[J]. 应用光学,2007,28(4):388-391. ZHAO Mi-yang, CHEN Wei-dong, LU Xiao-yan. Improvement of image compressing algorithm based on SPIHT[J]. Journal of Applied Optics, 2007, 28(4): 388-391. (in Chinese with an English abstract) [4]刘海英,李云松,吴成柯,等. 一种高重构质量低复杂度的高光谱图像压缩感知[J]. 西安电子科技大学学报,2011,38(3):37-41. LIU Hai-ying, LI Yun-song, WU Cheng-ke, et al. Compressed hyperspectral image sensing based on inte