基于GPU的单幅图像去雾的实现及优化
|
摘要
基于暗通道先验规律的去雾算法已取得了良好的去雾效果,但算法所需要的计算时间过长,无法达到实时去雾的要求。使用GPU初步并行实现了去雾算法,并确定了算法中需要优化的部分。在优化过程中,一方面将数据存储到高速内存中以实现对数据的快速读取,另一方面设计新的算法实现方式以减少算法的计算量,最终提高了加速比。优化后的加速算法处理768×1024的图像仅需21 ms,达到了实时去雾的要求。
The defogging algorithm which based on dark channel prior had achieved good results,but the time spent on computing was too long to meet the requirements of real-time defogging. With parallel GPU,this paper implemented the defogging algorithm and it determined the portion of algorithm which need to optimized. During the optimization process,on the one hand,it stored the data in the high-speed memory to achieve rapid data read. On the other hand,it designed a new algorithm implementation to reduce the amount of calculation,it improved the acceleration rate ultimately. The acceleration algorithm just needed 21 ms when dealing with images of 768 × 1024 after optimized,so it reached real-time defogging implementation.
The defogging algorithm which based on dark channel prior had achieved good results,but the time spent on computing was too long to meet the requirements of real-time defogging. With parallel GPU,this paper implemented the defogging algorithm and it determined the portion of algorithm which need to optimized. During the optimization process,on the one hand,it stored the data in the high-speed memory to achieve rapid data read. On the other hand,it designed a new algorithm implementation to reduce the amount of calculation,it improved the acceleration rate ultimately. The acceleration algorithm just needed 21 ms when dealing with images of 768 × 1024 after optimized,so it reached real-time defogging implementation.
引文
[1] Narasimhan S G,Nayar S K. Contrast restoration of weather degraded images[J]. IEEE Trans on Pattern Analysis Machine Intelligence,2003,25(6):713-724.
[2] Liu Qi,Gao Xinbo,He Lihuo,et al. Haze removal for a single visible remote sensing image[J]. Signal Processing,2017,137(8):33-43.
[3] Hung C L,Ma Zhaohui,Lin Chunyuan,et al. Image haze removal of optimized contrast enhancement based on GPU[J]. Frontier Computing,2016,375(4):53-63
[4]郭璠,蔡自兴,谢斌,等.图像去雾技术研究综述与展望[J].计算机应用,2010,30(9):2417-2421.(Guo Fan,Cai Zixing,Xie Bin,et al. Review and prospect of image dehazing techniques[J]. Journal of Computer Applications,2010,30(9):2417-2421.)
[5] He Kaiming,Sun Jian,Tang Xiao’ou. Single image haze removal using dark channel prior[J]. IEEE Trans on Pattern Analysis and Machine Intelligence,2011,33(12):2341-2353.
[6] He Kaiming,Sun Jian,Tang Xiao’ou. Guided image filtering[J].IEEE Trans on Pattern Analysis and Machine Intelligence,2013,35(6):1397-1409.
[7] Lyu Xingyong,Chen Wenbin,Shen I F. Real-time dehazing for image and video[C]//Proc of the 18th IEEE Conference on Pacific Computer Graphics and Applications. Washington DC:IEEE Computer Society,2010:62-69.
[8] Xue Yungang,Ren Ju,Su Huayou,et al. Parallel implementation and optimization of haze removal using dark channel prior based on CUDA[C]//Proc of High Performance Computing. Berlin:Springer,2013:99-109.
[9]李佳童,章毓晋.图像去雾算法的改进和主客观性能评价[J].光学精密工程,2017,25(3):735-741.(Li Jiatong,Zhang Yujin. Improvements of image haze removal algorithm and its subjective and objective performance evaluation[J]. Optics and Precision Engineering,2017,25(3):735-741.)
[10]魏颖慧,张彦娥,梅树立,等.基于暗通道先验和区间插值小波变换的图像去雾霾方法[J].农业工程学报,2017,33(S1):281-287.(Wei Yinghui,Zhang Yan’e,Mei Shuli,et al. Image dehazing method based on dark channel prior and interval interpolation wavelet transform[J]. Trans of the Chinese Society of Agricultural Engineering,2017,33(S1):281-287.)
[11]宋颖超,罗海波,惠斌,等.尺度自适应暗通道先验去雾方法[J].红外与激光工程,2016,45(9):286-297.(Song Yingchao,Luo Haibo,Hui Bin,et al. Haze removal using scale adaptive dark channel prior[J]. Infrared and Laser Engineering,2016,45(9):286-297.)
[12]陈书贞,任占广,练秋生.基于改进暗通道和导向滤波的单幅图像去雾算法[J].自动化学报,2016,42(3):455-465.(Chen Shuzhen,Ren Zhanguang,Lian Qiusheng. Single image dehazing algorithm based on improved dark channel prior and guided filter[J]. Acta Automatica Sinica,2016,42(3):455-465.)
[13]Tarel J P,Hautiere N. Fast visibility restoration from a single color or gray level image[C]//Proc of the 12th IEEE International Conference on Computer Vision. Piscataway,NJ:IEEE Press,2009:2201-2208.
[14]Wilt N. CUDA专家手册-GPU权威编程指南[M].北京:机械工业出版社,2014.(Wilt N. CUDA expert manual-GPU authoritative programming guide[M]. Beijing:Machinery Industry Press,2014.)
[2] Liu Qi,Gao Xinbo,He Lihuo,et al. Haze removal for a single visible remote sensing image[J]. Signal Processing,2017,137(8):33-43.
[3] Hung C L,Ma Zhaohui,Lin Chunyuan,et al. Image haze removal of optimized contrast enhancement based on GPU[J]. Frontier Computing,2016,375(4):53-63
[4]郭璠,蔡自兴,谢斌,等.图像去雾技术研究综述与展望[J].计算机应用,2010,30(9):2417-2421.(Guo Fan,Cai Zixing,Xie Bin,et al. Review and prospect of image dehazing techniques[J]. Journal of Computer Applications,2010,30(9):2417-2421.)
[5] He Kaiming,Sun Jian,Tang Xiao’ou. Single image haze removal using dark channel prior[J]. IEEE Trans on Pattern Analysis and Machine Intelligence,2011,33(12):2341-2353.
[6] He Kaiming,Sun Jian,Tang Xiao’ou. Guided image filtering[J].IEEE Trans on Pattern Analysis and Machine Intelligence,2013,35(6):1397-1409.
[7] Lyu Xingyong,Chen Wenbin,Shen I F. Real-time dehazing for image and video[C]//Proc of the 18th IEEE Conference on Pacific Computer Graphics and Applications. Washington DC:IEEE Computer Society,2010:62-69.
[8] Xue Yungang,Ren Ju,Su Huayou,et al. Parallel implementation and optimization of haze removal using dark channel prior based on CUDA[C]//Proc of High Performance Computing. Berlin:Springer,2013:99-109.
[9]李佳童,章毓晋.图像去雾算法的改进和主客观性能评价[J].光学精密工程,2017,25(3):735-741.(Li Jiatong,Zhang Yujin. Improvements of image haze removal algorithm and its subjective and objective performance evaluation[J]. Optics and Precision Engineering,2017,25(3):735-741.)
[10]魏颖慧,张彦娥,梅树立,等.基于暗通道先验和区间插值小波变换的图像去雾霾方法[J].农业工程学报,2017,33(S1):281-287.(Wei Yinghui,Zhang Yan’e,Mei Shuli,et al. Image dehazing method based on dark channel prior and interval interpolation wavelet transform[J]. Trans of the Chinese Society of Agricultural Engineering,2017,33(S1):281-287.)
[11]宋颖超,罗海波,惠斌,等.尺度自适应暗通道先验去雾方法[J].红外与激光工程,2016,45(9):286-297.(Song Yingchao,Luo Haibo,Hui Bin,et al. Haze removal using scale adaptive dark channel prior[J]. Infrared and Laser Engineering,2016,45(9):286-297.)
[12]陈书贞,任占广,练秋生.基于改进暗通道和导向滤波的单幅图像去雾算法[J].自动化学报,2016,42(3):455-465.(Chen Shuzhen,Ren Zhanguang,Lian Qiusheng. Single image dehazing algorithm based on improved dark channel prior and guided filter[J]. Acta Automatica Sinica,2016,42(3):455-465.)
[13]Tarel J P,Hautiere N. Fast visibility restoration from a single color or gray level image[C]//Proc of the 12th IEEE International Conference on Computer Vision. Piscataway,NJ:IEEE Press,2009:2201-2208.
[14]Wilt N. CUDA专家手册-GPU权威编程指南[M].北京:机械工业出版社,2014.(Wilt N. CUDA expert manual-GPU authoritative programming guide[M]. Beijing:Machinery Industry Press,2014.)