Design of Airborne Video Dehazing System for UCAV Based on HSV Transmission Weighted Correction
-
摘要: 针对无人机机载雾天实时获取图像降质的问题,设计一种基于HSV (hue,saturation,value)透射率加权修正的无人机视频去雾处理系统. 首先,根据机载低功耗实时去雾系统要求,完成去雾系统总体设计;其次,根据去雾系统视频采集需要,设计数字视频BT.656/BT.1120隔行和逐行处理、视频控制、指令接收处理、TS1601视频去雾算法处理、H.264视频压缩处理和组帧等功能;最后,重点介绍系统去雾处理算法和平台设计、去雾参数化处理等功能模块的设计实现,分别使用本系统方法和相关文献方法对典型含雾图像处理,采用方差函数、平均梯度函数、TenenGrad函数等3种清晰度评价函数并做归一化处理,以进行客观评价. 研究结果表明:根据HSV分量透射率加权修正的无人机机载图像去雾处理系统设计,具有功耗小、易实现和适应性强等特点;对典型含雾图像3处理后方差函数归一化分别提高46.87%、1.44%、12.83%,平均梯度函数归一化分别提高12.54%、9.26%、11.15%,TenenGrad函数归一化分别提高53.19%、3.60%、8.82%,测试算法整体运算时间分别提高4.74、5.41倍和5.46倍.Abstract: As there is image degradation when the UCAV (unmanned combat air vehicle) captures the real-time image in the haze day, a low power-consumption system is designed on the basis of the HSV (hue, saturation, value) transmission weighted correction. First, the overall design of the dehazing system is completed, which can meet the requests of a low power-consumption and real-time image dehazing. Then, according to the needs of the video acquisition dehazing system, functions are being designed, including digital video BT.656/BT.1120 interlaced and progressive processing, video control, instruction receiving processing, TS1601 video dehazing algorithm processing, H.264 video compression processing and framing, etc. Lastly, the design and implementation are focused in terms of the system dehazing algorithm, platform design, dehazing parameter processing and other functional modules. Also this system and several other methods suggested in references are used to process typical hazy images respectively, and then evaluated after employing three definition evaluation functions (variance function, average gradient function and TenenGrad function) and normalization process. The results indicate that the design of this dehazing system has such merits as low power consumption, easy implementation, and high adaptability. After processing the typical hazy images, the variance function is increased by 46.87%, 1.44% and 12.83%, the average gradient function is increased by 12.54%, 9.26% and 11.15%, that of normalization of TenenGrad function is increased by 53.19%, 3.60% and 8.82%, respectively. The overall operation time of the test algorithm is respectively increased by 4.74 times, 5.41 times and 5.46 times.
-
表 1 图6去雾性能客观评价结果
Table 1. Objective evaluation results of dehazing performance for fig.6
含雾图像 性能指标 原图 文献[2]方法 文献[3]方法 文献[4]方法 本文方法 1 方差函数 0.33 0.547 0.645 0.715 0.745 平均梯度函数 0.32 0.698 0.723 0.754 0.798 TenenGrad函数 0.30 0.505 0.695 0.743 0.705 运算时间/ms 42.476 56.611 87.121 14.631 2 方差函数 0.23 0.532 0.702 0.642 0.742 平均梯度函数 0.25 0.674 0.765 0.712 0.792 TenenGrad函数 0.27 0.606 0.866 0.786 0.806 运算时间/ms 115.492 115.481 127.654 15.654 3 方差函数 0.12 0.431 0.624 0.561 0.633 平均梯度函数 0.13 0.797 0.821 0.807 0.897 TenenGrad函数 0.17 0.564 0.834 0.794 0.864 运算时间/ms 121.834 136.261 137.258 21.251 注:表中黑体数据表示各项评价指标中的最优结果. -
HE Kaiming, SUN Jian, TANG Xiaoou, et al. Single image haze removal using dark channel prior[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2011, 33(12): 2341-2353. doi: 10.1109/TPAMI.2010.168 HE Kaiming, SUN Jian, TANG Xiaoou, et al. Guided image filtering[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2013, 35(6): 1397-1409. doi: 10.1109/TPAMI.2012.213 BERMAN D, TREIBITZ T, AVIDAN S. Non-local image dehazing[C]//Proceedings of the IEEE International Conference on Computer Vision and Pattern Recognition. Los Alamitos: IEEE Computer Society Press, 2016: 1674-1682 MENG G F, WANG Y, DUAN J Y, et al. Efficient image dehazing with boundary constraint and contextual regularization[C]//Proceedings of the IEEE International Conference on Computer Vision. Los Alamitos: IEEE Computer Society Press, 2013: 617-624. 范新南,冶舒悦,史朋飞,等. 改进大气散射模型实现的图像去雾算法[J]. 计算机辅助设计与图形学学报,2019,31(7): 1148-1155.FAN Xinnan, YE Shuyue, SHI Pengfei, et al. An image dehazing algorithm based on improved atmospheric scattering model[J]. Journal of Computer-Aided Design & Computer Graphics, 2019, 31(7): 1148-1155. CHEN Xuesong, LU Haihua, CHENG Kaili, et al. Sequentially refined spatial and channel-wise feature aggregation in encoder-decoder network for single image dehazing[C]//IEEE International Conference on Image Processing. [S.l.]: IEEE, 2019: 2776-2780. HU Haimiao, ZHANG Hongda, ZHAO Zichen, et al. Adaptive single image dehazing using joint local-global illumination adjustment[J]. IEEE Transactions on Multimedia, 2020, 22(6): 1485-1495. doi: 10.1109/TMM.2019.2944260 ZHAO D, XU L, YAN Y, et al. Multi-scale optimal fusion model for single image dehazing[J]. Signal Processing-Image Communication, 2019, 74: 253-265. doi: 10.1016/j.image.2019.02.004 LIU P J, HORNG S J, LIN J S, et al. Contrast in haze removal:configurable contrast enhancement model based on dark channel prior[J]. IEEE Transactions on Image Processing, 2019, 5(28): 2212-2227. QIN M J, XIE F Y, W Li, et al. Dehazing for multi-spectral remote sensing images based on a convolutional neural network with the residual architecture[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2018, 11(5): 1645-1655. SINGH P, KOMODAKIS N. Cloud-gan: cloud removal for sentinel-2 imagery using a cyclic consistent generative adversarial networks[C]//IEEE International Geoscience and Remote Sensing Symposium. [S.l.]: IEEE, 2018: 1772-1775. DING J, YAN Z, WEI X, et al. Light-weight residual learning for single image dehazing[J]. Journal of Elec- tronic Imaging, 2019, 28(3): 033013.1-033013.13. 周晓波,何魁华,周聪. 基于FPGA的图像高速去雾实时系统设计实现[J]. 电视技术,2018,42(4): 67-72.ZHOU Xiaobo, HE Kuijua, ZHOU Cong, et al. Design and implementation of high-speed real-time image dehazing system based on FPGA[J]. Video Engineering, 2018, 42(4): 67-72. 张海斌. 基于Zynq实时视频图像去雾系统的设计[D]. 上海: 上海师范大学, 2018. 匡娇娇,张瑞珏,王慧芳. 改进的暗通道去雾算法在多核DSP上的并行实现[J]. 武汉大学学报(理学版),2016,62(6): 519-524.KUANG Jiaojiao, ZHANG Ruijue, WANG Huifang, et al. Parallel implementaiton of improved dark channel dehazing algorithm on multicores DSP[J]. Journal of Wuhan University of Technology (Natural Science Edition), 2016, 62(6): 519-524. ZHANG B, WEI J. Hardware implementation for haze removal with adaptive filtering[J]. IEEE Access, 2019, 7(11): 142498-142506. Techwell Corporation. Data sheet of TW9912[M/OL]. [S.l.]: Techwell Corporation, 2011[2019-10-08]. http://www.techwelllinc.com. Semtech Corporation. Gennum corporation GS2971 data sheet[M/OL]. [S.l.]: Semtech Corporation, 2009[2019-10-08]. http://www.semtech.com. Surveillance Semiconductor. TS1601 3D HD video enhance processor product datasheet[M/OL]. [S.l.]: Surveillance Semiconductor, 2012[2019-10-08]. http://www.tssic.com. NXP Semiconductors. i.MX6 dual/6 quad multimedia applications processor reference manual[M/OL]. Freescale: NXP Semiconductors, 2012[2019-10-08]. http://www.nxp.com.cn. 王健,辛向龙,张修飞. 基于i.MX6无人机视频码率受控的压缩系统设计[J]. 火力与指挥控制,2016,41(11): 148-152. doi: 10.3969/j.issn.1002-0640.2016.11.034WANG Jian, XIN Xianglong, ZHANG Xiufei. Design of the embedded video process system based on i.MX6[J]. Fire Control & Command Control, 2016, 41(11): 148-152. doi: 10.3969/j.issn.1002-0640.2016.11.034 王健,秦春霞,杨珂,等. 无人机机载视频去雾系统设计[J]. 计算机测量与控制,2020,28(6): 135-139.WANG Jian, QIN Chunxia, YANG Ke, et al. Design of the airborne video demisting system for UCAV[J]. Computer Measurement & Control, 2020, 28(6): 135-139. 耿威. 基于视频图像的雾天能见度检测方法研究与实现[D]. 南京: 东南大学, 2013. 李雪,江旻珊. 光学显微成像系统图像清晰度评价函数的对比[J]. 光学仪器,2018,40(1): 28-38.LI Xue, Jiang Minshan. Comparison of sharpness function based on microscope[J]. Optical Instruments, 2018, 40(1): 28-38. WANG Zhou, BOVIK A C, SHEIKH H R, et al. Image quality assessment:from error visibility to structural similarity[J]. IEEE Transactions on Image Processing, 2004, 13(4): 600-612. doi: 10.1109/TIP.2003.819861