矿井灾害事故发生后，快速掌握井下受灾情况是成功救援的关键。无人机具有机动灵活、环境适应性强等优势，通过搭载深度相机（RGB-D相机），可以为矿井灾变巷道态势侦测提供新途径，为专家掌握灾区受灾情况并制定针对性救援方案提供支持。本文采用理论分析和实验研究相结合的方法，从矿井低照度复杂环境下图像增强方法和矿井灾变巷道环境三维重构方案研究等方面展开研究，主要成果如下：

确定了灾害应急侦测无人机搭载D435i深度相机是适用于矿井灾变巷道环境三维重建系统解决方案，自主设计并搭建了模拟矿井灾变巷道试验平台并开展不同试验工况下图像采集试验，掌握了环境参数对图像采集质量影响规律，分别开展针对彩色图像和深度图像的图像增强方法研究：①针对彩色图像成像效果灰暗、纹理细节模糊、对比度低等问题，提出了一种抑制光晕效应的彩色图像去雾算法，基于区域适应性估计方法进行大气光参数精确估计，采用双边滤波引导空间同态滤波优化DCP算法进行图像去雾，有效去雾的同时保持良好的色彩真实性；②针对深度图像噪声大、存在深度值缺失空洞、边缘模糊等问题，提出了一种改进联合双边滤波的深度图像空洞修复算法，引入快速高斯变换优化联合双边滤波的高斯滤波部分，利用灰度图作为引导进行深度图像修复，效率比传统JBF算法提升了45.317%，能够有效修复深度空洞，并且具有较好的边缘特性。

开展了灾变巷道环境点云三维重建方法研究，通过离群点剔除和体素网格降采样方法实现点云简化，简化率达43.48%，较好得保持了原有点云结构并且极大得提升了点云处理效率；提出了两步式点云配准方案改进ICP点云配准算法，通过试验得出NDT-ICP算法相较PFH-ICP、FPFH-ICP算法配准精度和配准效率更高，有效解决了点云初始位置较差和随机采样点几何特征不稳定的问题。最后通过模拟矿井灾变巷道三维重建试验对提出的三维重构方案进行试验验证，分析了全局视角下灾变环境三维重建模型，对模型中主要障碍物进行重建精度分析，结果表明，重建模型能够完整还原巷道内障碍物布置情况，重建精度达到厘米级，证明了本文方案在矿井灾变巷道环境中真实可行。

Rapid understanding of the underground disaster situation is crucial for successful mine rescue operations following mining accidents. The unmanned aerial vehicle (UAV) possesses the advantages of mobility and flexibility, strong environmental adaptability, etc. By carrying Depth Cameras (RGB-D Camera), it can provide a new way to detect the situation of the mine disaster tunnel, providing support for experts to grasp the affected areas and formulate targeted rescue plans. This study combines theoretical analysis and experimental research to investigate methods for image enhancement in low-light and complex mining environments, as well as 3D reconstruction of disaster-stricken tunnel environments. The main achievements are as follows:

The suitability of equipping D435i depth camera on emergency detection UAV for 3D reconstruction of disaster-stricken tunnel environments in mines was determined. A simulated mine disaster tunnel test platform was autonomously designed and constructed, and image acquisition experiments were conducted under different test conditions. The impact of environmental parameters on image acquisition quality was analyzed. Image enhancement methods were separately developed for color images and depth images: 1. To address issues such as dim imaging effects, blurred texture details, and low contrast in color images, a color image dehazing algorithm was proposed. It accurately estimates atmospheric light parameters based on region-adaptive estimation methods, and applies a bilateral-filter-guided spatial homomorphic filtering optimization to the dark channel prior DCP algorithm for effective dehazing while preserving color fidelity. 2. To address problems such as high noise levels, depth value gaps, and edge blurring in depth images, an improved joint bilateral filtering-based depth image hole-filling algorithm was proposed. The fast Gaussian transform was introduced to optimize the Gaussian filtering component of joint bilateral filtering, and the grayscale image was used as a guide for depth image restoration. The efficiency of the proposed algorithm was found to be 45.317% higher than that of the traditional joint bilateral filtering JBF algorithm. It can effectively repair depth holes and maintain good edge characteristics.

Research on point cloud-based 3D reconstruction methods in disaster-stricken tunnel environments was conducted. Outlier removal and voxel grid downsampling methods were employed to simplify the point cloud, achieving a simplification rate of 43.48%. The original structure of the point cloud was well preserved, and the efficiency of point cloud processing was significantly improved. An improved two-step point cloud registration scheme was proposed to enhance the Iterative Closest Point ICP algorithm. Through experiments, it was determined that the Normal Distributions Transform NDT-ICP algorithm outperforms the Point Feature Histogram PFH-ICP and Fast Point Feature Histogram FPFH-ICP algorithms in terms of registration accuracy and efficiency. This effectively addresses the issues of poor initial point cloud alignment and unstable geometric features of randomly sampled points. The proposed 3D reconstruction scheme was experimentally validated through simulated mine disaster tunnel reconstruction tests. The 3D reconstruction model of the disaster environment from a global perspective was analyzed, and the reconstruction accuracy of the main obstacles in the model was evaluated. The study demonstrate that the reconstruction model can faithfully restore the obstacle arrangement in the tunnel, with the reconstruction accuracy reaching the centimeter level, which proves that the scheme of this paper is real and feasible in the disaster tunnel environment of the mine.

[1]Khan I, Zakari A, Dagar V, et al. World energy trilemma and transformative energy developments as determinants of economic growth amid environmental sustainability[J]. Energy Economics, 2022, 108: 105884.

[2]王双明,刘浪,朱梦博,等.“双碳”目标下煤炭绿色低碳发展新思路[J].煤炭学报,2024,49(1):152-171.

[3]谢和平,任世华,谢亚辰,等.碳中和目标下煤炭行业发展机遇[J].煤炭学报,2021,46(7):2197-2211.

[4]王国法,任世华,庞义辉,等.煤炭工业“十三五”发展成效与“双碳”目标实施路径[J].煤炭科学技术, 2021,49(9):1-8.

[5]刘峰,曹文君,张建明,等.我国煤炭工业科技创新进展及“十四五”发展方向[J]. 煤炭学报, 2021,46(1):1-15.

[6]袁亮.煤炭工业碳中和发展战略构想[J].中国工程科学,2023,25(5):103-110.

[7]梁运培,郑梦浩,李全贵,等.我国煤与瓦斯突出预测与预警研究现状[J].煤炭学报,2023,48(8):2976-2994.

[8]李泉新,刘飞,方俊,等.我国煤矿井下智能化钻探技术装备发展与展望[J].煤田地质与勘探,2021,49(6):265-272.

[9]袁亮.深部采动响应与灾害防控研究进展[J].煤炭学报,2021,46(3):716-725.

[10]袁亮,张平松.煤炭精准开采地质保障技术的发展现状及展望[J].煤炭学报,2019,44(8):2277-2284.

[11]袁亮,杨科.再论废弃矿井利用面临的科学问题与对策[J].煤炭学报,2021,46(1):16-24.

[12]聂高众,高建国,苏桂武,等.地震应急救助需求的模型化处理——来自地震震例的经验分析[J].资源科学,2001(1):69-76.

[13]Xue XS, Cao XG Zhang ZH, et al. Electromagnetic Wave Attenuation Mechanism and Distribution Strategy for Coal Mine Rescue Robot Under the Typical Obstacle Environment[J]. Radio Science, 2020, 55(3).

[14]温学雷.煤矿井下应急救援超前侦测发射装置设计研究[J].煤炭科学技术,2019,47(8):164-168.

[15]覃海明,吴燕清,肖勇.煤矿井下远程灾区侦测系统研制[J].煤矿安全,2015,46(11):111-113.

[16]Perkins G, du Toit E, Cochrane G, et al. Overview of underground coal gasification operations at Chinchilla, Australia[J]. Energy Sources, Part A: Recovery, Utilization, and Environmental Effects, 2016, 38(24): 3639-3646.

[17]Huang B, Wang Y. Roof weakening of hydraulic fracturing for control of hanging roof in the face end of high gassy coal longwall mining: a case study[J]. Archives of Mining Sciences, 2016, 61(3).

[18]Guo W, Wang H, Chen S. Coal pillar safety and surface deformation characteristics of wide strip pillar mining in deep mine[J]. Arabian Journal of Geosciences, 2016, 9(2): 1-9.

[19]Ross M R V, McGlynn B L, Bernhardt E S. Deep impact: Effects of mountaintop mining on surface topography, bedrock structure, and downstream waters[J]. Environmental science & technology, 2016, 50(4): 2064-2074.

[20]Wang D, Li X, Wang X, et al. Effects of hoisting parameters on dynamic contact characteristics between the rope and friction lining in a deep coal mine[J]. Tribology International, 2016, 96: 31-42.

[21]Jimoh A Y, Ojo O J. Rock-Eval pyrolysis and organic petrographic analysis of the Maastrichtian coals and shales at Gombe, Gongola Basin, Northeastern Nigeria[J]. Arabian Journal of Geosciences, 2016, 9(6): 1-13.

[22]Wen H, Wu W, Zheng X, et al. Application and Development Trend of Radar Detection Technology in Mine Rescue[C]//Proceedings of the 11th International Mine Ventilation Congress. Springer, Singapore, 2019: 942-950.

[23]Nithya T, Ezak M M, Kumar K R, et al. Rescue and protection system for underground mine workers based on Zigbee[J]. Int. J. Recent Res. Asp, 2018, 4(4): 194-197.

[24]郑太雄,黄帅,李永福,等.基于视觉的三维重建关键技术研究综述[J].自动化学报,2020,46(4):631-652.

[25]Isgro F, Odone F, Verri A. An open system for 3D data acquisition from multiple sensor[C]//Seventh International Workshop on Computer Architecture for Machine Perception (CAMP'05). IEEE, 2005: 52-57.

[26]Ma Z, Liu S. A review of 3D reconstruction techniques in civil engineering and their applications[J]. Advanced Engineering Informatics, 2018, 37: 163-174.

[27]Xu Y, Stilla U. Toward building and civil infrastructure reconstruction from point clouds: A review on data and key techniques[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2021, 14: 2857-2885.

[28]QIU Z. Key techniques on cultural relic 3D reconstruction[J]. ACTA ELECTONICA SINICA, 2008, 36(12): 2423.

[29]Kiyasu S, Hoshino H, Yano K, et al. Measurement of the 3-D shape of specular polyhedrons using an M-array coded light source[J]. IEEE transactions on instrumentation and measurement, 1995, 44(3): 775-778.

[30]Snavely N, Seitz S M, Szeliski R. Photo tourism: exploring photo collections in 3D[M]//ACM siggraph 2006 papers. 2006: 835-846.

[31]Zhong Y, Zhang HF. Control points based semi-dense matching[C]//Proceedings of the 5th Asian Conference on Computer Vision. Melbourne, Australia, 2002, 144.

[32]Pollefeys M, Nistér D, Frahm J M, et al. Detailed real-time urban 3d reconstruction from video[J]. International Journal of Computer Vision, 2008, 78(2): 143-167.

[33]Zhang Z. Microsoft kinect sensor and its effect[J]. IEEE multimedia, 2012, 19(2): 4-10.

[34]陈湘源.煤矿无线通信系统的现状与发展[J].煤炭工程,2009(F01):62-65.

[35]Raval S, Banerjee B P, Singh S K, et al. A preliminary investigation of mobile mapping technology for underground mining[C]//IGARSS 2019-2019 IEEE International Geoscience and Remote Sensing Symposium. IEEE, 2019: 6071-6074.

[36]白琪.基于机器视觉的矿井救灾机器人路径规划的研究[D].北京:中国矿业大学,2018.

[37]刘洋,高磊,吴学群,等.点云特征保留的精简方法研究[J].应用激光,2024,44(1):144-154.

[38]房屹松.矿井钻孔救援监控系统的研究与开发[D].西安:西安科技大学,2015.

[39]顾海荣,罗佳,高子渝,等.基于深度相机的大直径救援井三维模型重建研究[J].煤田地质与勘探,2023,51(5):188-197.

[40]孙海波.基于机器视觉的矿井环境三维地图构建研究[D].北京:中国矿业大学,2020.

[41]陈清华,张旭.基于双目立体匹配的巷道三维重建研究[J].激光杂志,2022,43(10):208-212.

[42]周楠.基于机器视觉的矿井环境三维重建研究[D].北京:中国矿业大学,2022.

[43]徐靖雯.基于机器视觉的井壁扫描成像系统研究[D].徐州:中国矿业大学,2021.

[44]杨勇,岳建华,李玉良,等.一种矿井动态图像增强方法[J].工矿自动化,2015,41(11):48-52.

[45]赵谦,王奕婷,曾召华,等.一种矿井巷道图像边缘检测算法[J].工矿自动化,2013(2):71-75.

[46]王焱,关南楠,刘海涛.改进的多尺度Retinex井下图像增强算法[J].辽宁工程技术大学学报:自然科学版,2016(4):440-443.

[47]王文庆,晏婷.基于空间域的图像去噪方法比较研究[J].西安邮电学院学报,2012,17(2):75-79.

[48]张国伟.基于小波变换的图像去噪方法研究[D].昆明:昆明理工大学,2014.

[49]易翔,王蔚然.基于小波域统计混合模型的图像降噪方法[J].电子与信息学报,2005,27(11):1722-1725.

[50]张俞晴,何宁,甄雪艳,等.基于机器学习的图像去噪研究进展[J].中国计算机用户协会网络应用分会2017年第二十一届网络新技术与应用年会论文集,2017.

[51]肖俊,宋寿鹏,丁丽娟.空域同态滤波算法研究[J].中国图象图形学报,2008,13(12):2302-2306.

[52]姜柏军,钟明霞.改进的直方图均衡化算法在图像增强中的应用[J].激光与红外,2014,44(6):702-706.

[53]Zhang X, Huang W, Wang Q, et al. SSR-NET: Spatial–spectral reconstruction network for hyperspectral and multispectral image fusion[J]. IEEE Transactions on Geoscience and Remote Sensing, 2020, 59(7): 5953-5965.

[54]Vajha M, Balaji S B, Kumar P V. Small-d MSR codes with optimal access, optimal sub-packetization and linear field size[J]. IEEE Transactions on Information Theory, 2023.

[55]Song S, Zhu J, Li X, et al. Integrate MSRCR and mask R-CNN to recognize underwater creatures on small sample datasets[J]. IEEE Access, 2020, 8: 172848-172858.

[56]付燕,李瑶,严斌斌.一种煤矿井下视频图像增强算法[J].工矿自动化,2018, 44(7): 80-83.

[57]Lu Z, Hu X,Qian J,et al.A novel image denoising algorithm used for coal mine[J].Journal of Information and Computational Science, 2012, 9(5): 1321-1327.

[58]Zhao Yanqi. Multi-level denoising and enhancement method based on wavelet transform for mine monitoring[J]. International Journal of Mining Science and Technology, 2013, 23(1): 163-166.

[59]唐守锋,史可,仝光明,等.一种矿井低照度图像增强算法[J].工矿自动化,2021,47(10):32-36.

[60]李晓宇,吕进来,郝晓丽.一种改进的Retinex矿井图像增强算法[J].科学技术与工程,2020,20(29):12028-12034.

[61]田昊,王小玉.融合暗通道先验与粒子群算法的去雾改进算法[J].北京邮电大学学报,2024, 47(2):118.

[62]黄向东,邓莉洁.一种基于Retinex的矿井非均匀照度图像增强算法[J].科学技术与工程,2014,14(5):141-144+170.

[63]Le AV, Jung SW, Won CS. Directional joint bilateral filter for depth images[J]. Sensors, 2014, 14(7): 11362-11378.

[64]Shen YH, Li JH, Lü CH. Depth map enhancement method based on joint bilateral filter. 2014 7th International Congress on Image and Signal Processing[J]. IEEE, 2014, 153-158.

[65]吴倩,史晋芳,王德娇,等.结合彩色图像局部分割的Kinect深度图修复算法[J].计算机应用研究,2017,34(12):3852-3854+3884.

[66]万琴,朱晓林,陈国泉,等.分层联合双边滤波的深度图修复算法研究[J].计算机工程与应用,2021,57(6):184-190.

[67]王广建,刘德儿.顾及边缘交融特征的自适应深度图空洞修复算法[J]. Laser & Optoelectronics Progress, 2022, 59(18): 1810016-1810016-7.

[68]Ahn I, Kim C. A Novel Depth-Based Virtual View Synthesis Method for Free Viewpoint Video[J]. IEEE Transactions on Broadcasting, 2013, 59(4): 614-626.

[69]Choi J, Min D, Sohn K. Reliability-based multiview depth enhancement considering interview coherence[J]. IEEE transactions on circuits and systems for video technology, 2013, 24(4): 603-616.

[70]Xue H, Zhang S, Cai D. Depth image inpainting: Improving low rank matrix completion with low gradient regularization[J]. IEEE Transactions on Image Processing, 2017, 26(9): 4311-4320.

[71]任超锋,张楠.数码相机畸变模型的相互转换方法[J].激光与光电子学进展,2018,55(7):406-413.

[72]Zhang Z. Flexible camera calibration by viewing a plane from unknown orientations [C]//Proceedings of the seventh ieee international conference on computer vision. IEEE, 1999, 1: 666-673.

[73]Furgale P, Rehder J, Siegwart R. Unified temporal and spatial calibration for multi-sensor systems[C]//2013 IEEE/RSJ International Conference on Intelligent Robots and Systems. IEEE, 2013: 1280-1286.

[74]Qin T, Shen S. Online temporal calibration for monocular visual-inertial systems[C]//2018 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). IEEE, 2018: 3662-3669.

[75]McCartney E J. Optics of the atmosphere: scattering by molecules and particles[J]. New York, 1976.

[76]Nayar S K, Narasimhan S G. Vision in bad weather[C]//Proceedings of the Seventh IEEE International Conference on Computer Vision. IEEE, 2002.

[77]谢伟,余瑾,涂志刚,等.消除光晕效应和保持细节信息的图像快速去雾算法[J].ApplicationResearchofComputers/JisuanjiYingyongYanjiu,2019,36(4):1228-1231.

[78]杨燕,王志伟.基于均值不等关系优化的自适应图像去雾算法[J].电子与信息学报,2020,42(3):9.

[79]张江鑫,周佳波,孟利民.改进的暗通道先验快速图像去雾算法[J].计算机应用,2017,037(8):2324-2328.

[80]Kim J H, Jang W D, Sim J Y, et al. Optimized contrast enhancement for real-time image and video dehazing[J]. Journal of Visual Com-munication Image Representation, 2013, 24(3): 410-425.

[81]He Kaiming, Sun Jian, Tang Xiaoou. Single image haze removal using dark channel prior[J]. IEEE Trans on Pattern Analysis Machine Intelligence, 2011, 33(12): 2341-2353.

[82]刘金荣,李淳芃,欧阳建权,等.基于联合双边滤波的深度图像增强算法[J].计算机工程,2014,40(3):249-252.

[83]张竞阳,严利民,陈志恒.采用暗态点光源模型的夜间去雾算法[J].红外技术,2021,43(8):798-803.

[84]刘继忠,吴文虎,程承,等.基于像素滤波和中值滤波的深度图像修复方法[J].光电子．激光,2018,29(5):6.

[85]卫星,刘清,郭依正,等.联合双边滤波器耦合多分辨率的图像降噪算法[J].计算机工程与设计,2016,37(12):7.

[86]孟恬,全红艳.基于快速行进算法的深度图像增强算法[J].计算机应用与软件,2017,34(8):6.

[87]Balta H, Velagic J, Bosschaerts W, et al. Fast statistical outlier removal based method for large 3D point clouds of outdoor environments[J]. IFAC-PapersOnLine, 2018, 51(22): 348-353.

[88]李龙龙,宫成,李雨成,等.矿山巷道三维激光扫描点云复合滤波算法研究[J].金属矿山,2023,(11):205-212.

[89]Besl P J, McKay N D. Method for registration of 3-D shapes[C]//Sensor fusion IV: control paradigms and data structures. Spie, 1992, 1611: 586-606.

[90]Rusu R B, Blodow N, Marton Z C, et al. Aligning point cloud views using persistent feature histograms[C]//2008 IEEE/RSJ international conference on intelligent robots and systems. IEEE, 2008: 3384-3391.

[91]Rusu R B, Blodow N, Beetz M. Fast point feature histograms (FPFH) for 3D registration[C]//2009 IEEE international conference on robotics and automation. IEEE, 2009: 3212-3217.

[92]Biber P, Straßer W. The normal distributions transform: A new approach to laser scan matching[C]//Proceedings 2003 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS 2003)(Cat. No. 03CH37453). IEEE, 2003, 3: 2743-2748.