使用能量匹配的監(jiān)控視頻自適應(yīng)速率壓縮感知
doi: 10.11999/JEIT190750 cstr: 32379.14.JEIT190750
-
云南大學(xué)信息學(xué)院 昆明 650504
基金項目: 國家自然科學(xué)基金(61861045)
Adaptive-Rate Compressive Sensing Using Energy Matching for Monitoring Video
-
School of Information Science and Engineering, Yunnan University, Kunming 650504, China
Funds: The National Natural Science Foundation of China (61861045)
-
摘要: 獲取信號稀疏度對壓縮感知(CS)性能的提升有重大意義，但在采樣端不進(jìn)行完整信號數(shù)字化采集和存儲的情況下，對信號稀疏度進(jìn)行估計比較困難。現(xiàn)有方法在稀疏度估計性能和計算復(fù)雜度方面難以取得較好的平衡。針對采樣端對信號特性未知的監(jiān)控視頻應(yīng)用，該文提出一種新的使用能量匹配的自適應(yīng)速率壓縮感知方法(ARCS-EM)，通過觀測一個恒定低速率的壓縮感知觀測結(jié)果來對當(dāng)前幀實際稀疏度進(jìn)行估計，然后根據(jù)估計結(jié)果決定當(dāng)前幀應(yīng)執(zhí)行的壓縮感知測量數(shù)，再進(jìn)行補(bǔ)充測量得到當(dāng)前幀的優(yōu)化壓縮感知采樣結(jié)果。實驗結(jié)果表明，該方法可以較好地適應(yīng)視頻中前景稀疏度的變化，為每幀圖像分配適當(dāng)?shù)膲嚎s感知測量速率，在不顯著提高采樣端計算復(fù)雜度的前提下，有效提高重建視頻的質(zhì)量。
-
關(guān)鍵詞:
- 圖像信號處理 /
- 壓縮感知 /
- 自適應(yīng)速率采樣 /
- 能量匹配 /
- 監(jiān)控視頻
Abstract: Signal sparsity is of great significance for the improvement of Compressive Sensing (CS) performance. However, it is difficult to estimate the sparsity when the whole signal is not captured and stored at the sampling side. Few existing mothed can achieve good balance in terms of the sparsity estimation performance and the computational complexity. For the monitoring video applications where the signal characteristics is unknown for sampling devices, a new Adaptive-Rate CS using Energy Matching (ARCS-EM) method is proposed. By observing the measurement results of a low-rate compressive sensing, the actual sparsity of the current frame is estimated and then the rate of measurement for the current frame is determined. Finally, supplementary measurements are performed to obtain the optimized compressive sensing result for the current frame. Experiment results show that the proposed method could allocate suitable measurement rate for each frame to adapt to the variation of sparsity in different frames. The quality of reconstructed videos is effectively improved without noticeably increasing computational complexity in the sampling side. -
表 1 實驗參數(shù)
參數(shù) $\varSigma $ $a$ $b$ $\tau $ $r$ 視頻序列Hall 2.65 16 128 8 600 視頻序列PETS 2.45 16 128 8 600 
下載: 導(dǎo)出CSV
表 2 不同方法的自適壓縮感知平均性能對比
實驗結(jié)果 Hall視頻平均壓縮
感知采樣率Hall視頻平均峰值
信噪比(dB)PETS視頻平均壓縮
感知采樣率PETS視頻平均峰值
信噪比(dB)Oracle 0.2040 36.59 0.1317 40.02 CDSAM方法 0.2297 36.34 0.2001 39.53 ARCS-CV方法 0.2137 37.03 0.1191 39.07 ARCS-EM方法 0.2232 37.26 0.1350 40.26
下載: 導(dǎo)出CSV
表 3 采樣運行時間對照表(ms)
運行時間 Hall視頻T1 Hall視頻T2 Hall視頻T3 Hall視頻T PETS視頻T1 PETS視頻T2 PETS視頻T3 PETS視頻T CDSAM方法 7.48 104.18 0 111.66 6.76 94.93 0 101.69 ARCS-CV方法 957.26 0.14 2.99×105 3.00×105 526.56 0.11 1.44×105 1.45×105 ARCS-EM方法 800.87 0.42 0 801.29 498.75 0.51 0 499.26
下載: 導(dǎo)出CSV
-
CANDèS E J, ROMBERG J, and TAO T. Robust uncertainty principles: Exact signal reconstruction from highly incomplete frequency information[J]. IEEE Transactions on Information Theory, 2006, 52(2): 489–509. doi: 10.1109/TIT.2005.862083 DONOHO D L. Compressed sensing[J]. IEEE Transactions on Information Theory, 2006, 52(4): 1289–1306. doi: 10.1109/TIT.2006.871582 CANDèS E J. The restricted isometry property and its implications for compressed sensing[J]. Comptes Rendus Mathematique, 2008, 346(9/10): 589–592. doi: 10.1016/j.crma.2008.03.014 王鋼, 周若飛, 鄒昳琨. 基于壓縮感知理論的圖像優(yōu)化技術(shù)[J]. 電子與信息學(xué)報, 2020, 42(1): 222–233. doi: 10.11999/JEIT190669WANG Gang, ZHOU Ruofei, and ZOU Yikun. Research on image optimization technology based on Compressed Sensing[J]. Journal of Electronics &Information Technology, 2020, 42(1): 222–233. doi: 10.11999/JEIT190669 楊森林, 萬國賓. 聯(lián)合結(jié)構(gòu)預(yù)測和運動補(bǔ)償?shù)囊曨l自適應(yīng)壓縮感知[J]. 西北大學(xué)學(xué)報: 自然科學(xué)版, 2019, 49(6): 909–917. doi: 10.16152/j.cnki.xdxbzr.2019-06-010YANG Senlin and WAN Guobin. Adaptive compressed sensing of video by combining structural predictions and motion compensation[J]. Journal of Northwest University:Natural Science Edition, 2019, 49(6): 909–917. doi: 10.16152/j.cnki.xdxbzr.2019-06-010 SHAHRASBI B and RAHNAVARD N. Model-based nonuniform Compressive Sampling and recovery of natural images utilizing a Wavelet-domain universal hidden Markov model[J]. IEEE Transactions on Signal Processing, 2017, 65(1): 95–104. doi: 10.1109/TSP.2016.2614654 LAKSHMI T C S, GNANADURAI D, and MUTHULAKSHMI I. Energy conserving texture-based adaptable Compressive Sensing scheme for WVSN[J]. Concurrency and Computation: Practice and Experience, 2019: e5178. doi: 10.1002/cpe.5178 ZHANG Xufan, WANG Yong, WANG Dianhong, et al. Adaptive image compression based on compressive sensing for video sensor nodes[J]. Multimedia Tools and Applications, 2018, 77(11): 13679–13699. doi: 10.1007/s11042-017-4981-6 DUARTE M F, DAVENPORT M A, TAKHAR D, et al. Single-pixel imaging via compressive sampling[J]. IEEE Signal Processing Magazine, 2008, 25(2): 83–91. doi: 10.1109/msp.2007.914730 范劍英, 馬明陽, 趙首博. 基于壓縮感知高反光成像技術(shù)研究[J]. 電子與信息學(xué)報, 2020, 42(4): 1013–1020. doi: 10.11999/JEIT190512FAN Jianying, MA Mingyang, and ZHAO Shoubo. Research on high reflective imaging technology based on compressed sensing[J]. Journal of Electronics &Information Technology, 2020, 42(4): 1013–1020. doi: 10.11999/JEIT190512 吳新杰, 閆詩雨, 徐攀峰, 等. 基于稀疏度自適應(yīng)壓縮感知的電容層析成像圖像重建算法[J]. 電子與信息學(xué)報, 2018, 40(5): 1250–1257. doi: 10.11999/JEIT170794WU Xinjie, YAN Shiyu, XU Panfeng, et al. Image reconstruction algorithm for electrical capacitance tomography based on sparsity adaptive Compressed Sensing[J]. Journal of Electronics &Information Technology, 2018, 40(5): 1250–1257. doi: 10.11999/JEIT170794 SHANGGUAN Wentao, YAN Qiurong, WANG Hui, et al. Adaptive single photon compressed imaging based on constructing a smart threshold matrix[J]. Sensors, 2018, 18(10): 3449. doi: 10.3390/s18103449 YUAN Xin, YANG Jianbo, LLULL P, et al. Adaptive temporal compressive sensing for video[C]. 2013 IEEE International Conference on Image Processing, Melbourne, Australia, 2013: 14–18. doi: 10.1109/ICIP.2013.6738004. WANG Yeru, TANG Chaoying, CHEN Yueting, et al. Adaptive temporal Compressive Sensing for video with motion estimation[J]. Optical Review, 2018, 25(2): 215–226. doi: 10.1007/s10043-018-0408-5 練秋生, 田天, 陳書貞, 等. 基于變采樣率的多假設(shè)預(yù)測分塊視頻壓縮感知[J]. 電子與信息學(xué)報, 2013, 35(1): 203–208. doi: 10.3724/SP.J.1146.2012.00590LIAN Qiusheng, TIAN Tian, CHEN Shuzhen, et al. Block compressed sensing of video based on variable sampling rates and multihypothesis predictions[J]. Journal of Electronics &Information Technology, 2013, 35(1): 203–208. doi: 10.3724/SP.J.1146.2012.00590 LI Honggui. Compressive domain spatial–temporal difference saliency-based realtime adaptive measurement method for video recovery[J]. IET Image Processing, 2019, 13(11): 2008–2017. doi: 10.1049/iet-ipr.2019.0116 WARNELL G, BHATTACHARYA S, CHELLAPPA R, et al. Adaptive-rate compressive sensing using side information[J]. IEEE Transactions on Image Processing, 2015, 24(11): 3846–3857. doi: 10.1109/TIP.2015.2456425 DUARTE-CARVAJALINO J M, YU Guoshen, CARIN L, et al. Task-driven adaptive statistical compressive sensing of Gaussian mixture models[J]. IEEE Transactions on Signal Processing, 2013, 61(3): 585–600. doi: 10.1109/TSP.2012.2225054 VAN DER BERG E and FRIEDLANDER M P. Probing the Pareto Frontier for basis pursuit solutions[J]. SIAM Journal on Scientific Computing, 2009, 31(2): 890–912. doi: 10.1137/080714488 DONOHO D L and TANNER J. Precise undersampling theorems[J]. Proceedings of the IEEE, 2010, 98(6): 913–924. doi: 10.1109/jproc.2010.2045630 CEVHER V, SANKARANARAYANAN A, DUARTE M F, et al. Compressive sensing for background subtraction[C]. The 10th European Conference on Computer Vision, France, Marseille, 2008: 155–168. doi: 10.1007/978-3-540-88688-4_12. WARD R. Compressed sensing with cross validation[J]. IEEE Transactions on Information Theory, 2009, 55(12): 5773–5782. doi: 10.1109/tit.2009.2032712 JOHNSON W B and LINDENSTRAUSS J. Extensions of Lipschitz mappings into a Hilbert space[J]. Contemporary Mathematics, 1984, 26(12): 189–206. doi: 10.1090/conm/026/737400 -
圖(4) / 表(3)
計量
- 文章訪問數(shù): 857
- HTML全文瀏覽量: 234
- PDF下載量: 99
- 被引次數(shù): 0
下載:
