-
激光成像是在激光探测技术基础发展而成的,是激光探测技术和成像技术相结合而产生的一种新型技术手段,前者奠定了激光成像发射和接收技术的基础,后者则奠定了激光成像整体技术的基础。由于激光具有亮度高、方向性强、单色性好、相干性强和波长短等特点,这就使得激光在成像领域具有与生俱来的优势[1]。
激光成像与雷达成像在技术原理上可以说是一脉相承,只不过激光成像将探测媒介由微波和毫米波变为激光。激光波长比微波和毫米波短得多,波束宽度比微波和毫米波要更窄。激光成像与雷达成像相比,图像分辨率更高,信息量更大。激光成像与可见光和红外成像相比,大多数体制的激光成像可以全天时工作,不受白天和黑夜光照条件限制,而且获得的信息更多。
目标的激光图像提供了物体大小、形状、表面材料等特征参数,大大丰富了对物体的描述信息,增强了人们对物体的了解和掌握程度。激光成像的过程依靠成像系统来实现。
图1中表述了从激光发射到信号接收处理以及后续信息处理的激光成像全过程。在这里,信号处理指的是传感器输出信号的处理,处理后的信号包含更加易于识别利用的信息,但并不能直接为用户所用。而信息处理就是指利用处理后的信号实现各类激光成像任务,从无法直接利用的信号中提取目标信息,从而实现目标识别、重建等复杂任务。在实际的成像场景中,通过激光发射系统和接收器的发和收,可以得到场景中的激光回波信号,再经过成像传感器的处理,就可以使得激光信号转化为更加易于处理的信号(比如数字信号等)。以上的成像过程依靠精密设计的光学硬件系统来实现,但通过硬件接收处理的场景信息可能会包含许多冗余的、存在误差的信息,比如在对海上重要目标进行激光成像的过程中,海面上还会存在许多非重点关注的目标、海面杂波等,这就会产生一定的多余信号。另外,在存在光学干扰(比如变化的温度、大气湍流等)的场景下,信号依然会存在一定的误差,成像信号呈现出来的场景信息是不易于识别,甚至是错误的。所以还需要对传感器的信号进行进一步的处理,比如信号去噪、辐射、位置等信息的校正等。处理后的信号最终是为了实现具体的成像任务的,所以对处理后的信号进一步分析,通过一定的算法得到所需的信息,比如目标识别、跟踪等。
-
为了说明多种激光成像体制下的成像处理技术的特点,找出不同体制的激光成像处理技术存在的异同,下面将对典型成像体制下的激光成像处理技术做一对比分析。
从表1中可以看出,典型成像体制的激光成像在信号处理和信息处理所包含的内容有所差异。究其原因,主要有三个方面:一是不同成像体制对应的激光成像原理不尽相同。例如,激光扫描成像是通过多个激光光束对目标进行快速扫描获得的回波点云图像;而合成孔径激光成像则是通过激光光源的移动获得时间分辨率上小孔径合成大孔径成像的效果,二者的载荷状态不同,接收到的信号和对应的信息处理也会产生较大的区别。二是各类体制的激光成像对应不同的应用场景。激光成像在不同的应用场景中需要完成的信息处理任务有所区别,例如,连续波激光成像一般应用场景中激光光源视场不存在强干扰和遮挡,激光能够直接探测到目标;而非视域激光成像的应用场景就是激光视场中存在无法透射的遮挡物,激光无法直线传播,需要根据激光的多次反射信号对遮挡目标进行探测和识别。所以二者在信息处理时由于激光回波信号的质量、稳定度的差异使得非视域成像中几乎无法实现高分辨成像。三是不同体制的激光成像对应的激光光源存在差异。例如,激光扫描成像和阵列成像均使用单激光发射系统作为光源,而相干激光成像需要获得相干的多个激光束对目标进行探测,通过光场的衍射或干涉效应来间接获得目标的信号,对应的相干成像会造成图像散斑噪声和伪影等现象,需要对信号进行处理。
Type of imaging Imaging process description Typical signal processing Typical information processing methods Laser scan imaging[1] Through the scanning of laser emitting system, we can obtain many kinds of information about the surface reflection signal of the object, so that we can image the object in two or three dimensions. ①Laser Point Cloud Denoising
②Laser Point Cloud Intensity Correction
③Point Cloud Location Correction
④Point Cloud Distribution Processing①(Single) Photon ranging
②Point cloud 3D reconstruction
③Target detection
④Target trackingLaser array imaging[2] Laser signals emitted by the array's array of laser emitters can be scanned or directly detected to obtain a variety of information about reflected signals on the surface of a detected object, thus completing 2D or 3D object imaging. ①Laser Point Cloud Denoising
②Laser Point Cloud Intensity Correction
③Point Cloud Location Correction
④Point Cloud Distribution Processing①(Single) Photon ranging
②Point cloud 3D reconstruction
③Target detection
④Target trackingLaser coherence imaging[3] Using coherent laser as the light source, two or three dimensional object imaging is achieved by obtaining wavefront phase information through matter diffraction or interferometric radiation distribution. ①Echo Signal Denoising
②Echo Radiation Correction
③Echo Geometry Calibration
④Laser image speckle noise suppression
⑤Laser Image False Removal①Laser Weak Signal Enhancement
②Laser Range
③Relevance Information Target Reconstruction
④Target detection
⑤High resolution imagingSynthetic aperture laser imaging[4] By using the small aperture laser imaging system, the image fields of each subsystem are synchronized and the same phase is superimposed into the large aperture system. ①Echo Signal Denoising
②Echo Radiation Correction
③Echo Position Correction
④Laser image speckle noise suppression
⑤Laser Signal Phase Compensation
⑥Laser Signal Motion Compensation①On-board, airborne laser remote sensing ranging
②Multi-Perspective Target Reconstruction
③Remote sensing target detection
④High resolution imagingContinuous wave laser imaging[5] Continuous wave laser imaging uses a continuous light signal as the detection signal, based on phase laser ranging technology, and uses single-frequency signal modulation laser. By phase detection of the reflected light signal, the target distance information is obtained, thus achieving high-efficiency imaging. ①Echo signal denoising
②Echo radiation correction
③Echo geometric correction
④Laser signal phase compensation
⑤Laser signal motion compensation①Laser dynamic ranging
②Target reconstruction (in motion)
③Target detection (in motion)
④Motion target velocity measurement
⑤High-resolution imagingNon-line-of-sight laser imaging[6] Laser signals may encounter relay obstructions and undergo diffuse reflection during transmission. Sparse information contained in the reflection can be used to create two-dimensional or three-dimensional images of objects that are out of sight. ①Light field noise reduction
②Reflection and echo radiation correction
③Reflection and echo geometric correction①Non-line-of-sight laser ranging
②Non-line-of-sight 2D/3D target reconstruction
③Non-line-of-sight 2D/3D target detection
④Non-line-of-sight target trackingCorrelated photon imaging[7] Using a single-pixel laser detector for detection, the total radiation value of the target object's information light field is recorded in chronological order. The system calculates the target image by using this value and correlating it with the speckle field matrix of the illuminated object. ①Echo signal denoising
②Echo radiation correction
③Echo geometric correction
④Speckle suppression in laser imaging
⑤Artifacts removal in laser imaging①Laser remote sensing ranging
②Laser image reconstruction
③Remote sensing target detection
④Anti-interference high-resolution imaging
⑤Laser image encryption与此同时,不同体制的激光成像在信号和信息处理方面存在诸多共性内容。在信号处理方面,首先,不论是何种体制的成像,均在激光成像过程中面临环境和光学系统内部干扰,进而出现噪声,所以在信号处理过程中,信号去噪是后续任务的前提,也是不同体制信号处理的共性内容之一。第二,激光成像过程中还面临着大气、温度、湿度等环境的变化,其对激光传输和光学硬件均会产生不同程度的影响,进而导致激光辐射强度发生衰减或其他变化,所以激光信号的辐射校正也是不同体制信号处理的共性内容之一。第三,由于激光载荷和探测目标的运动状态未知,且不同体制的激光成像均会面临对运动目标成像的需求,哪怕定位信息可以通过辅助定位传感器进行融合,也会由于信号时延、融合效率等问题导致激光信号出现几何误差,所以激光信号的几何校正也是不同体制信号处理的共性内容。第四,当前随着高维信息需求的不断增长,不同体制的激光成像也向着三维成像发展,而点云信息是目前表示目标三维信息非常成功的一种方式,激光成像通过获取目标点云能够获得不同运动状态、不同表面信息的三维目标成像,所以点云处理也是激光成像信号处理的一个重要的共性问题。从信息处理方面来看,激光成像的下游任务十分繁杂,根据用户的需求信息处理的内容也不一样。然而,不论是何种体制的激光成像,首先都要获得激光测距信息,这不仅是激光成像的基础,也是激光成像区别于其他模态成像的关键因素。其次,随着不同体制的激光成像运用场景的拓展,对成像质量和精度要求也越来越高,直接获取得到的激光图像已经无法满足任务需求,所以需要进行进一步的图像重建处理,这一信息处理任务在典型成像体制的激光成像中均有所应用。最后,目前利用激光成像的下游任务中最常见的是目标识别、跟踪、分割等,而这些任务的前提是对成像目标实现目标检测,只有完成了对成像目标的检测后才能够确定目标的一系列属性特征,进而完成各类下游任务。所以,目标检测也是各种体制的激光成像信息处理的共性内容之一。
综上对比可以发现,虽然对应不同体制的信号和信息处理内容名称上有所差异,但可以将激光成像信号处理的共性内容归纳为信号去噪、辐射校正、几何校正和点云处理4个方面,而成像信息处理的共性内容归纳为激光测距、图像重建、目标检测3种共性处理内容。所以在下文中,主要围绕激光成像信号和信息处理技术的共性内容进行总结和分析。
Recent progress and prospect of laser imaging processing technology (invited)
-
摘要: 激光成像在多个领域中应用广泛,其成像时包含一系列的信号和信息处理过程,会对成像质量造成关键性的影响,并在成像信息运用的过程中起着关键性作用。文中主要针对典型激光成像处理技术进行综述性研究。首先,讨论了典型成像体制的激光成像处理技术特征,按照信号处理和信息处理两个阶段归纳了主要的共性处理方法和内容。其次,分析了激光信号去噪、辐射与几何校正和激光点云处理技术,激光测距、图像重建、目标检测的信息处理技术的发展现状,研究了典型处理技术特别是基于深度学习的激光成像智能处理技术的实现过程。最后,分析了激光成像处理技术的发展需求和未来发展方向,希望能对激光成像的相关研究带来一定的参考作用。Abstract:
Significance Laser imaging refers to an imaging method that emits a specially designed laser signal, receives the laser echo, and processes it to obtain attribute information such as an image of the target. Laser imaging has wide applications in target detection, satellite surveying, smart agriculture, national defense and aerospace, and other fields. It contains a series of signals and information processing processes, including denoising, radiation, geometric correction, point cloud processing of laser echo signals, and subsequent data processing of various imaging tasks (such as laser ranging, laser image reconstruction, target detection, etc.), and have a critical impact on imaging quality and play a crucial role in the application of imaging information. Currently, with the continuous development of imaging systems and imaging hardware, laser imaging processing technology has increasingly high requirements for processing accuracy and speed, and involves a wider range of technical fields. Especially with the rapid development of machine learning technology represented by deep learning, it has achieved better results than traditional technologies in many classic problems, and has also been successfully applied in laser imaging processing technology, providing a new development direction for laser imaging processing. Progress This paper first introduces the characteristics of laser imaging processing technology of typical imaging system (Fig.1). We explained the characteristics of imaging processing technologies under various laser imaging systems, identified the similarities and differences between laser imaging processing technologies under different systems, and conducted a comparative analysis of laser imaging processing technologies under typical imaging systems (Tab.1). In summary, it can be found that although there are differences in the names of signal and information processing contents corresponding to different systems, the common contents of laser imaging signal processing can be summarized into four aspects of signal denoising, radiation correction, geometric correction, and point cloud processing. The common contents of imaging information processing can be summarized into three common processing contents of laser ranging, image reconstruction, and object detection. Based on the summarized common methods of laser imaging signals and information processing technology, we conducted separate studies. In the current research status of laser imaging signal processing technology, we focus on the laser signal denoising, correction and laser point cloud processing technology. In the research of signal denoising, we have conducted research based on wavelet transform, empirical mode decomposition, variational mode decomposition, and hybrid methods. We have also conducted specialized research on the application of deep learning algorithms in laser signal denoising. Representative algorithms are shown (Fig.5). The laser signal correction focuses on two aspects of laser signal radiation and geometric correction. And in point cloud signal processing, we mainly summarized the work on denoising and background removal, and focused on the work based on deep learning. Besides, we have organized and summarized the research on laser information processing for laser ranging, image reconstruction and target detection information processing technology. In the section of laser image reconstruction, we conducted research on three aspects of stereo matching, point cloud data stitching, and laser reflection tomography reconstruction. In object detection, the traditional method and deep-learning based method were elaborated, and classic point cloud object detection algorithms based on deep learning algorithms were studied (Fig.9-10). Based on the classification of laser imaging processing technology in this paper, we finally analyzed the current challenges and future development directions of laser imaging processing technologies, and summarized the current development of laser imaging technology and future laser imaging processing technology examples. It is hoped that it can provide some reference for the research related to laser imaging. Conclusions and Prospects Laser imaging has always been a hot topic in the field of optical imaging and signal processing. In the past 20 years, laser imaging signal and information processing technology has made great progress. In the previous studies, deep learning has been deeply applied to laser imaging processing. Through the powerful representation learning ability of deep learning, great improvements have been made in laser imaging processing quality, precision, robustness and other aspects. In the future research on different signal and information processing tasks, the standardization of large-scale data sets for imaging tasks and more robust deep neural network processing paradigm will be the further development direction of the research. It should be noted that laser imaging processing technology is not limited to the contents in this paper. There are many other signal and information processing technologies not involved in this paper, which worth further study and exploration by researchers. -
Key words:
- laser imaging /
- signal processing /
- information processing /
- point cloud processing /
- deep learning
-
Type of imaging Imaging process description Typical signal processing Typical information processing methods Laser scan imaging[1] Through the scanning of laser emitting system, we can obtain many kinds of information about the surface reflection signal of the object, so that we can image the object in two or three dimensions. ①Laser Point Cloud Denoising
②Laser Point Cloud Intensity Correction
③Point Cloud Location Correction
④Point Cloud Distribution Processing①(Single) Photon ranging
②Point cloud 3D reconstruction
③Target detection
④Target trackingLaser array imaging[2] Laser signals emitted by the array's array of laser emitters can be scanned or directly detected to obtain a variety of information about reflected signals on the surface of a detected object, thus completing 2D or 3D object imaging. ①Laser Point Cloud Denoising
②Laser Point Cloud Intensity Correction
③Point Cloud Location Correction
④Point Cloud Distribution Processing①(Single) Photon ranging
②Point cloud 3D reconstruction
③Target detection
④Target trackingLaser coherence imaging[3] Using coherent laser as the light source, two or three dimensional object imaging is achieved by obtaining wavefront phase information through matter diffraction or interferometric radiation distribution. ①Echo Signal Denoising
②Echo Radiation Correction
③Echo Geometry Calibration
④Laser image speckle noise suppression
⑤Laser Image False Removal①Laser Weak Signal Enhancement
②Laser Range
③Relevance Information Target Reconstruction
④Target detection
⑤High resolution imagingSynthetic aperture laser imaging[4] By using the small aperture laser imaging system, the image fields of each subsystem are synchronized and the same phase is superimposed into the large aperture system. ①Echo Signal Denoising
②Echo Radiation Correction
③Echo Position Correction
④Laser image speckle noise suppression
⑤Laser Signal Phase Compensation
⑥Laser Signal Motion Compensation①On-board, airborne laser remote sensing ranging
②Multi-Perspective Target Reconstruction
③Remote sensing target detection
④High resolution imagingContinuous wave laser imaging[5] Continuous wave laser imaging uses a continuous light signal as the detection signal, based on phase laser ranging technology, and uses single-frequency signal modulation laser. By phase detection of the reflected light signal, the target distance information is obtained, thus achieving high-efficiency imaging. ①Echo signal denoising
②Echo radiation correction
③Echo geometric correction
④Laser signal phase compensation
⑤Laser signal motion compensation①Laser dynamic ranging
②Target reconstruction (in motion)
③Target detection (in motion)
④Motion target velocity measurement
⑤High-resolution imagingNon-line-of-sight laser imaging[6] Laser signals may encounter relay obstructions and undergo diffuse reflection during transmission. Sparse information contained in the reflection can be used to create two-dimensional or three-dimensional images of objects that are out of sight. ①Light field noise reduction
②Reflection and echo radiation correction
③Reflection and echo geometric correction①Non-line-of-sight laser ranging
②Non-line-of-sight 2D/3D target reconstruction
③Non-line-of-sight 2D/3D target detection
④Non-line-of-sight target trackingCorrelated photon imaging[7] Using a single-pixel laser detector for detection, the total radiation value of the target object's information light field is recorded in chronological order. The system calculates the target image by using this value and correlating it with the speckle field matrix of the illuminated object. ①Echo signal denoising
②Echo radiation correction
③Echo geometric correction
④Speckle suppression in laser imaging
⑤Artifacts removal in laser imaging①Laser remote sensing ranging
②Laser image reconstruction
③Remote sensing target detection
④Anti-interference high-resolution imaging
⑤Laser image encryption -
[1] 胡以华. 激光成像目标侦察[M]. 北京: 国防工业出版社. 2014: 1 Hu Y H. Laser Imaging Target Reconnaissance[M]. Beijing: National Defense Industry Press, 2014: 1.(in Chinese) [2] 侯阿慧, 胡以华, 赵楠翔, 等. 扩展目标光子测距回波特性及误差研究[J]. 中国激光, 2021, 48(04): 261-268. doi: 10.3788/CJL202148.0401016 Hou A H, Hu Y H, Zhao N X, et al. Echo characteristics and error of extended target for photon ranging [J]. Chinese Journal of Lasers, 2021, 48(4): 0401016. (in Chinese) doi: 10.3788/CJL202148.0401016 [3] 胡以华. 激光相干探测应用理论方法[M]. 北京: 科学出版社. 2022 Hu Y H. Applied Theory and Method of Laser Coherent Detection [M]. Beijing: Science Press, 2022. (in Chinese) [4] 胡以华. 空间激光成像目标精确侦察技术[J]. 国防科技, 2016, 37(01): 30-36. doi: 10.13943/j.issn1671-4547.2016.01.08 Hu Y H. Space laser imaging target precise reconnaissance technology [J]. National Defense Technology, 2016, 37(1): 30-36. (in Chinese) doi: 10.13943/j.issn1671-4547.2016.01.08 [5] 陈鹏, 赵继广, 宋一铄, 等. 气溶胶环境下FMCW与脉冲激光探测性能对比[J]. 红外与激光工程, 2020, 49(06): 187-196. doi: 10.3788/IRLA20190399 Chen P, Zhao J G, Song Y S, et al. Comparison on detection performance of FMCW and pulsed lidar in aerosol environment [J]. Infrared and Laser Engineering, 2020, 49(6): 20190399. (in Chinese) doi: 10.3788/IRLA20190399 [6] 赵禄达, 董骁, 徐世龙, 等. 典型成像模式下非视域成像重建算法研究现状[J/OL]. 中国光学(中英文). doi: 10.37188/CO.2022-0186 Zhao L D, Dong X, Xu S L, et al. Recent progress of non-line-of-sight imaging reconstruction algorithms in typical imaging modalities [J]. Chinese Optics, 2023, 16(3): 479-499. (in Chinese) doi: 10.37188/CO.2022-0186 [7] 龚文林, 孙建锋, 邓陈进, 等. 基于相干探测的单像素激光成像雷达研究进展[J]. 激光与光电子学进展, 2021, 58(10): 36-48. doi: 10.3788/LOP202158.1011003 Gong W L, Sun J F, Deng C J, et al. Research progress of single-pixel laser imaging radar based on coherent detection [J]. Progress in Laser and Optoelectronics, 2021, 58(10): 1011003. (in Chinese) doi: 10.3788/LOP202158.1011003 [8] Pathak R S. The Wavelet Transform[M]. Paris: Atlantis Press, 2009. [9] Jansen M, Malfait M, Bultheel A. Generalized cross validation for wavelet thresholding [J]. Signal Processing, 1997, 56(1): 33-44. doi: 10.1016/S0165-1684(97)83621-3 [10] Fang H T, Huang D S. Noise reduction in lidar signal based on discrete wavelet transform [J]. Optics Communications, 2004, 233(1-3): 67-76. doi: 10.1016/j.optcom.2004.01.017 [11] Blu T, Luisier F. The SURE-LET approach to image denoising [J]. IEEE Transactions on Image Processing, 2007, 16(11): 2778-2786. doi: 10.1109/tip.2007.906002 [12] Luisier F, Blu T, Unser M. A new SURE approach to image denoising: Interscale orthonormal wavelet thresholding [J]. IEEE Transactions on Image Processing, 2007, 16(3): 593-606. doi: 10.1109/tip.2007.891064 [13] Wu S, Liu Z, Liu B. Enhancement of lidar backscatters signal-to-noise ratio using empirical mode decomposition method [J]. Optics Communications, 2006, 267(1): 137-144. doi: 10.1016/j.optcom.2006.05.069 [14] Tian P, Cao X, Liang J, et al. Improved empirical mode decomposition based denoising method for lidar signals [J]. Optics Communications, 2014, 325: 54-59. doi: 10.1016/j.optcom.2014.03.083 [15] Chang J, Zhu L, Li H, et al. Noise reduction in Lidar signal using correlation-based EMD combined with soft thresholding and roughness penalty [J]. Optics Communications, 2018, 407: 290-295. doi: 10.1016/j.optcom.2017.09.063 [16] Dragomiretskiy K, Zosso D. Variational mode decomposition [J]. IEEE Transactions on Signal Processing, 2013, 62(3): 531-544. doi: 10.1109/TSP.2013.2288675 [17] Wang Y, Markert R, Xiang J, et al. Research on variational mode decomposition and its application in detecting rub-impact fault of the rotor system [J]. Mechanical Systems and Signal Processing, 2015, 60: 243-251. doi: 10.1016/j.ymssp.2015.02.020 [18] Yang Y, Jiang D. Casing vibration fault diagnosis based on variational mode decomposition, local linear embedding, and support vector machine [J]. Shock and Vibration, 2017, 2017: 5963239. doi: 10.1155/2017/5963239 [19] Yang W, Peng Z, Wei K, et al. Superiorities of variational mode decomposition over empirical mode decomposition particularly in time–frequency feature extraction and wind turbine condition monitoring [J]. IET Renewable Power Generation, 2017, 11(4): 443-452. doi: 10.1049/iet-rpg.2016.0088 [20] Lian J, Liu Z, Wang H, et al. Adaptive variational mode decomposition method for signal processing based on mode characteristic [J]. Mechanical Systems and Signal Processing, 2018, 107: 53-77. doi: 10.1016/j.ymssp.2018.01.019 [21] Mao X, Yang Q, Wang X, et al. Application of variational mode decomposition and whale optimization algorithm to laser ultrasonic signal denoising [J]. Sensors, 2022, 23(1): 354. doi: 10.3390/s23010354 [22] Komaty A, Boudraa A O, Augier B, et al. EMD-based filtering using similarity measure between probability density functions of IMFs [J]. IEEE Transactions on Instrumentation and Measurement, 2013, 63(1): 27-34. doi: 10.1109/tim.2013.2275243 [23] Li Z, Chen J, Zi Y, et al. Independence-oriented VMD to identify fault feature for wheel set bearing fault diagnosis of high speed locomotive [J]. Mechanical Systems and Signal Processing, 2017, 85: 512-529. doi: 10.1016/j.ymssp.2016.08.042 [24] Liu Y, Yang G, Li M, et al. Variational mode decomposition denoising combined the detrended fluctuation analysis [J]. Signal Processing, 2016, 125: 349-364. doi: 10.1016/j.sigpro.2016.02.011 [25] 丁红波, 王珍珠, 刘东. 激光雷达信号去噪方法的对比研究[J]. 光学学报, 2021, 41(24): 9-18. doi: 10.3788/AOS202141.2401001 Ding H B, Wang Z Z, Liu D. Comparison of de-noising methods of LiDAR signal [J]. Acta Optica Sinica, 2021, 41(24): 2401001. (in Chinese) doi: 10.3788/AOS202141.2401001 [26] Hua T, Dai K, Zhang X, et al. Optimal VMD-based signal denoising for laser radar via Hausdorff distance and wavelet transform [J]. IEEE Access, 2019, 7: 167997-168010. doi: 10.1109/access.2019.2949063 [27] 胡以华, 张鑫源, 徐世龙, 等. 激光反射层析成像技术的研究进展[J]. 中国激光, 2021, 48(4): 8-25. doi: 10.3788/CJL202148.0401002 Hu Y H, Zhang X, Xu S, et al. Research progress of laser reflective tomography techniques [J]. Chinese Journal of Lasers, 2021, 48(4): 0401002. (in Chinese) doi: 10.3788/CJL202148.0401002 [28] Zhang X, Hu Y, Xu S, et al. Application of image fusion algorithm combined with visual saliency in target extraction of reflective tomography lidar image [J]. Computational Intelligence and Neuroscience, 2022, 2022: 8247344. doi: 10.1155/2022/8247344 [29] Rakotosaona M J, La Barbera V, Guerrero P, et al. Pointcleannet: Learning to denoise and remove outliers from dense point clouds[C]//Computer Graphics Forum. 2020, 39(1): 185-203. [30] Guerrero P, Kleiman Y, Ovsjanikov M, et al. Pcpnet learning local shape properties from raw point clouds[C]//Computer Graphics Forum, 2018, 37(2): 75-85. [31] Zhang D, Lu X, Qin H, et al. Pointfilter: Point cloud filtering via encoder-decoder modeling [J]. IEEE Transactions on Visualization and Computer Graphics, 2020, 27(3): 2015-2027. doi: 10.1109/tvcg.2020.3027069 [32] Pistilli F, Fracastoro G, Valsesia D, et al. Learning robust graph-convolutional representations for point cloud denoising [J]. IEEE Journal of Selected Topics in Signal Processing, 2020, 15(2): 402-414. doi: 10.1109/jstsp.2020.3047471 [33] Höfle B, Pfeifer N. Correction of laser scanning intensity data: Data and model-driven approaches [J]. ISPRS Journal of Photogrammetry and Remote Sensing, 2007, 62(6): 415-433. doi: 10.1016/j.isprsjprs.2007.05.008 [34] Eitel J U H, Höfle B, Vierling L A, et al. Beyond 3-D: The new spectrum of lidar applications for earth and ecological sciences [J]. Remote Sensing of Environment, 2016, 186: 372-392. doi: 10.1016/j.rse.2016.08.018 [35] Yan W Y, Shaker A, Habib A, et al. Improving classification accuracy of airborne LiDAR intensity data by geometric calibration and radiometric correction [J]. ISPRS Journal of Photogrammetry and Remote Sensing, 2012, 67: 35-44. doi: 10.1016/j.isprsjprs.2011.10.005 [36] Carrea D, Abellan A, Humair F, et al. Correction of terrestrial LiDAR intensity channel using Oren–Nayar reflectance model: An application to lithological differentiation [J]. ISPRS Journal of Photogrammetry and Remote Sensing, 2016, 113: 17-29. doi: 10.1016/j.isprsjprs.2015.12.004 [37] Ding Q, Chen W, King B, et al. Combination of overlap-driven adjustment and Phong model for LiDAR intensity correction [J]. ISPRS Journal of Photogrammetry and Remote Sensing, 2013, 75: 40-47. doi: 10.1016/j.isprsjprs.2012.09.015 [38] 谭凯, 程效军, 张吉星. TLS辐射数据的入射角及距离效应改正方法[J]. 武汉大学学报(信息科学版), 2017, 42(02): 223-228. doi: 10.13203/j.whugis.20150502 Tan K, Cheng X, Zhang J. Correction for incidence angle and distance effects on TLS intensity data [J]. Journal of Wuhan University (Information Science Edition), 2017, 42(2): 223-228. (in Chinese) doi: 10.13203/j.whugis.20150502 [39] Yan W Y, Shaker A. Radiometric correction and normalization of airborne LiDAR intensity data for improving land-cover classification [J]. IEEE Transactions on Geoscience and Remote Sensing, 2014, 52(12): 7658-7673. doi: 10.32920/ryerson.14649288.v1 [40] Zhang C, Gao S, Li W, et al. Radiometric calibration for incidence angle, range and sub-footprint effects on hyperspectral LiDAR backscatter intensity [J]. Remote Sensing, 2020, 12(17): 2855. doi: 10.3390/rs12172855 [41] 纪红柱, 陈思颖, 张寅超, 等. 地基离轴拉曼-米激光雷达几何因子校正[J]. 北京理工大学学报, 2019, 39(06): 644-649. doi: 10.15918/j.tbit1001-0645.2019.06.015 Ji H Z, Chen S Y, Zhang Y C, et al. Determination of geometric factor for ground-based Raman-Mie lidar with bi-static configuration [J]. Journal of Beijing University of Technology (Natural Edition), 2019, 39(6): 644-649. (in Chinese) doi: 10.15918/j.tbit1001-0645.2019.06.015 [42] Su J, Mccormick M P, Liu Z, et al. Obtaining a ground-based lidar geometric form factor using coincident spacebornelidar measurements [J]. Applied Optics, 2010, 49(1): 108-113. [43] Zhang X, Xue Z. Geometrical error correction research in high precision 2D laser measuring instrument[C]//Sixth International Symposium on Precision Engineering Measurements and Instrumentation, SPIE, 2010, 7544: 1710-1717. [44] Liu Y, Chen S, Zhang Y, et al. Algorithm of geometry correction for airborne 3D scanning laser radar[C]//2009 International Conference on Optical Instruments and Technology: Advanced Sensor Technologies and Applications. SPIE, 2009, 7508: 452-462. [45] Skaloud J, Lichti D. Rigorous approach to bore-sight self-calibration in airborne laser scanning [J]. ISPRS Journal of Photogrammetry and Remote Sensing, 2006, 61(1): 47-59. doi: 10.1016/j.isprsjprs.2006.07.003 [46] Rewehel E M, Li J, Keshk H M, et al. Geometric correction of aerial camera and LiDAR hybrid system data using GNSS/IMU[C]//2022 IEEE 13th International Conference on Software Engineering and Service Science (ICSESS), IEEE, 2022: 54-58. [47] Chen B, Pang Y. A denoising approach for detection of canopy and ground from ICESat-2's airborne simulator data in Maryland, USA[C]//AOPC 2015: Advances in Laser Technology and Applications, SPIE, 2015, 9671: 383-387. [48] Gao R, Park J, Hu X, et al. Reflective noise filtering of large-scale point cloud using multi-position LiDAR sensing data [J]. Remote Sensing, 2021, 13(16): 3058. doi: 10.3390/rs13163058 [49] Li Y, Fu H, Zhu J, et al. A filtering method for ICESat-2 photon point cloud data based on relative neighboring relationship and local weighted distance statistics [J]. IEEE Geoscience and Remote Sensing Letters, 2020, 18(11): 1891-1895. [50] Wang X, Pan Z, Glennie C. A novel noise filtering model for photon-counting laser altimeter data [J]. IEEE Geoscience and Remote Sensing Letters, 2016, 13(7): 947-951. doi: 10.1109/LGRS.2016.2555308 [51] Zhu X, Nie S, Wang C, et al. A noise removal algorithm based on OPTICS for photon-counting LiDAR data [J]. IEEE Geoscience and Remote Sensing Letters, 2020, 18(8): 1471-1475. [52] Zaman F, Wong Y P, Ng B Y. Density-based denoising of point cloud[C]//9th International Conference on Robotic, Vision, Signal Processing and Power Applications: Empowering Research and Innovation. Springer Singapore, 2017: 287-295. [53] Liu Z, Xiao X, Zhong S, et al. A feature-preserving framework for point cloud denoising [J]. Computer-Aided Design, 2020, 127: 102857. doi: 10.1016/j.cad.2020.102857 [54] Qi C R, Su H, Mo K, et al. Pointnet: Deep learning on point sets for 3D classification and segmentation[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2017: 652-660. [55] Roveri R, Öztireli A C, Pandele I, et al. Pointpronets: Consolidation of point clouds with convolutional neural networks[C]//Computer Graphics Forum, 2018, 37(2): 87-99. [56] Himmelsbach M, Hundelshausen F V, Wuensche H J. Fast segmentation of 3D point clouds for ground vehicles[C]//2010 IEEE Intelligent Vehicles Symposium. IEEE, 2010: 560-565. [57] Zermas D, Izzat I, Papanikolopoulos N. Fast segmentation of 3D point clouds: A paradigm on lidar data for autonomous vehicle applications[C]//2017 IEEE International Conference on Robotics and Automation (ICRA). IEEE, 2017: 5067-5073. [58] Moosmann F, Pink O, Stiller C. Segmentation of 3D lidar data in non-flat urban environments using a local convexity criterion[C]//2009 IEEE Intelligent Vehicles Symposium, 2009: 215-220. [59] Fischler M A, Bolles R C. Random sample consensus: a paradigm for model fitting with applications to image analysis and automated cartography [J]. Communications of the ACM, 1981, 24(6): 381-395. doi: 10.1145/358669.358692 [60] Narksri P, Takeuchi E, Ninomiya Y, et al. A slope-robust cascaded ground segmentation in 3D point cloud for autonomous vehicles[C]//2018 21st International Conference on Intelligent Transportation Systems (ITSC), IEEE, 2018: 497-504. [61] Paigwar A, Erkent Ö, Sierra-Gonzalez D, et al. Gndnet: Fast ground plane estimation and point cloud segmentation for autonomous vehicles[C]//2020 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), 2020: 2150-2156. [62] Xu J, Zhang R, Dou J, et al. Rpvnet: A deep and efficient range-point-voxel fusion network for lidar point cloud segmentation[C]//Proceedings of the IEEE/CVF International Conference on Computer Vision, 2021: 16024-16033. [63] Shen Z, Liang H, Lin L, et al. Fast ground segmentation for 3d lidar point cloud based on jump-convolution-process [J]. Remote Sensing, 2021, 13(16): 3239. doi: 10.3390/rs13163239 [64] 郭冠军, 李树楷, 胡以华. 机载激光雷达测距性能研究[J]. 光电子·激光, 2001(06): 592-595. doi: 10.16136/j.joel.2001.06.013 Guo G J, Li S K, Hu Y H. Study on the ranging performance of the airborne lidar [J]. Journal of Optoelectronics ·Laser, 2001(6): 592-595. (in Chinese) doi: 10.16136/j.joel.2001.06.013 [65] 刘鸿彬, 李铭, 王凤香, 等. 一种少光子高精度多波束激光雷达系统及验证[J]. 红外与毫米波学报, 2019, 38(04): 535-541. doi: 10.11972/j.issn.1001-9014.2019.04.021 Liu H B, Li M, Wang F X, et al. A high accuracy multi-beam lidar system and its verification on several photons [J]. Journal of Infrared and Millimeter Waves, 2019, 38(4): 535-541. (in Chinese) doi: 10.11972/j.issn.1001-9014.2019.04.021 [66] 刘鸿彬, 李铭, 舒嵘, 等. 少光子灵敏度精密激光测距方法及验证[J]. 红外与激光工程, 2019, 48(1): 93-99. doi: 10.3788/IRLA201948.0106001 Liu H B, Li M, Shu R, et al. Estimation and verification of high-accuracy laser ranging on several photons [J]. Infrared and Laser Engineering, 2019, 48(1): 0106001. (in Chinese) doi: 10.3788/IRLA201948.0106001 [67] Hou A H, Hu Y H, Zhao N X, et al. Full-waveform fast correction method for photon counting lidar [J]. Chinese Optics Letters, 2021, 19(5): 87-92. doi: 10.3788/col202119.052701 [68] Zhou H, Chen Y, Hyyppä J, et al. An overview of the laser ranging method of space laser altimeter [J]. Infrared Physics & Technology, 2017, 86: 147-158. doi: 10.1016/j.infrared.2017.09.011 [69] Hao Q, Cao J, Hu Y, et al. Differential optical-path approach to improve signal-to-noise ratio of pulsed-laser range finding [J]. Optics Express, 2014, 22(1): 563-575. doi: 10.1364/oe.22.000563 [70] 胡善江, 贺岩, 俞家勇, 等. 基于深度学习的脉冲激光测距回波时刻解算方法[J]. 中国激光, 2019, 46(10): 302-311. doi: 10.3788/CJL201946.1010001 Hu S J, He Y, Yu J Y. Method for solving echo time of pulse laser ranging based on deep learning [J]. Chinese Journal of Laser, 2019, 46(10): 1010001. (in Chinese) doi: 10.3788/CJL201946.1010001 [71] Xu X, Chen Y, Zhu K, et al. Research on FPGA pulse laser ranging method based on deep learning [J]. IEEE Transactions on Instrumentation and Measurement, 2021, 70: 1-11. doi: 10.1109/tim.2021.3096281 [72] Li P, Wang R, Wang Y, et al. Evaluation of the ICP algorithm in 3D point cloud registration [J]. IEEE Access, 2020, 8: 68030-68048. doi: 10.1109/ACCESS.2020.2986470 [73] Yang J, Li H, Jia Y. Go-icp: Solving 3D registration efficiently and globally optimally[C]//Proceedings of the IEEE International Conference on Computer Vision, 2013: 1457-1464. [74] Rosen D M, Carlone L, Bandeira A S, et al. A certifiably correct algorithm for synchronization over the special Euclidean group[C]//Algorithmic Foundations of Robotics XII: Proceedings of the Twelfth Workshop on the Algorithmic Foundations of Robotics. Cham: Springer International Publishing, 2020: 64-79. [75] Izatt G, Dai H, Tedrake R. Globally optimal object pose estimation in point clouds with mixed-integer programming[C]//Robotics Research: The 18th International Symposium ISRR. Springer International Publishing, 2020: 695-710. [76] Maturana D, Scherer S. Voxnet: A 3D convolutional neural network for real-time object recognition[C]//2015 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), IEEE, 2015: 922-928. [77] Wu Z, Song S, Khosla A, et al. 3D ShapeNets: A deep representation for volumetric shapes[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2015: 1912-1920. [78] Aoki Y, Goforth H, Srivatsan R A, et al. Pointnetlk: Robust & efficient point cloud registration using pointnet[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2019: 7163-7172. [79] Chu J, Nie C. Multi-view point clouds registration and stitching based on SIFT feature[C]//2011 3rd International Conference on Computer Research and Development, IEEE, 2011, 1: 274-278. [80] 范强, 刘鹏, 杨俊, 等. 基于3 D-Harris与FPFH改进的3 D-NDT配准算法[J]. 图学学报, 2020, 41(04): 567-575. doi: 10.11996/JG.j.2095-302X.2020040567 Fan Q, Liu P, Yang J, et al. Improved 3D-NDT point cloud registration algorithm based on 3D-Harris and FPFH [J]. Journal of Graphics, 2020, 41(4): 567-575. (in Chinese) doi: 10.11996/JG.j.2095-302X.2020040567 [81] Zhong Y, Bai F, Liu Y, et al. Point cloud splicing based on 3D-Harris operator[C]//2021 3rd International Symposium on Smart and Healthy Cities (ISHC), IEEE, 2021: 61-66. [82] 杨彪, 胡以华. 代数迭代法在激光反射断层成像目标重构中的应用[J]. 红外与激光工程, 2019, 48(07): 287-293. doi: 10.3788/IRLA201948.0726002 Yang Biao, Hu Yihua. Laser reflection tomography target reconstruction algorithm based on algebraic iteration [J]. Infrared and Laser Engineering, 2019, 48(7): 0726002. (in Chinese) doi: 10.3788/IRLA201948.0726002 [83] 刘一凡, 胡以华, 徐世龙, 等. 基于波形分解的激光反射层析成像优化方法[J]. 光学学报, 2023, 43(18): 1828002. doi: 10.3788/AOS222044 Liu Y F, Hu Y H, Xu S L, et al. Optimization method for laser reflective tomography imaging based on waveform decomposition method [J]. Acta Optica Sinica, 2023, 43(18): 1828002. (in Chinese) doi: 10.3788/AOS222044 [84] Hu Y. Theory and Technology of Laser Imaging Based Target Detection[M]. Singapore: Springer Press, 2018. [85] Moosmann F, Fraichard T. Motion estimation from range images in dynamic outdoor scenes[C]//2010 IEEE International Conference on Robotics and Automation, IEEE, 2010: 142-147. [86] Dewan A, Caselitz T, Tipaldi G D, et al. Motion-based detection and tracking in 3d lidar scans[C]//2016 IEEE International Conference on Robotics and Automation (ICRA), IEEE, 2016: 4508-4513. [87] Dewan A, Caselitz T, Tipaldi G D, et al. Rigid scene flow for 3D lidar scans[C]//2016 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), IEEE, 2016: 1765-1770. [88] Qi C R, Yi L, Su H, et al. Pointnet++: Deep hierarchical feature learning on point sets in a metric space[C]//Advances in Neural Information Processing Systems, 2017: 30. [89] Zeng Y, Hu Y, Liu S, et al. Rt3d: Real-time 3D vehicle detection in lidar point cloud for autonomous driving [J]. IEEE Robotics and Automation Letters, 2018, 3(4): 3434-3440. doi: 10.1109/lra.2018.2852843 [90] Zhou Y, Tuzel O. Voxelnet: End-to-end learning for point cloud based 3D object detection[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. 2018: 4490-4499. [91] Liu X, Qi C R, Guibas L J. Flownet3d: Learning scene flow in 3d point clouds[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2019: 529-537. [92] Wang Z, Li S, Howard-Jenkins H, et al. Flownet3d++: Geometric losses for deep scene flow estimation[C]//Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision, 2020: 91-98. [93] Mayer N, Ilg E, Hausser P, et al. A large dataset to train convolutional networks for disparity, optical flow, and scene flow estimation[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2016: 4040-4048. [94] Gojcic Z, Litany O, Wieser A, et al. Weakly supervised learning of rigid 3D scene flow[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2021: 5692-5703. [95] Wang G, Wu X, Liu Z, et al. Hierarchical attention learning of scene flow in 3D point clouds [J]. IEEE Transactions on Image Processing, 2021, 30: 5168-5181. doi: 10.1109/tip.2021.3079796 [96] Tishchenko I, Lombardi S, Oswald M R, et al. Self-supervised learning of non-rigid residual flow and ego-motion[C]//2020 International Conference on 3D Vision (3DV), IEEE, 2020: 150-159. [97] Baur S A, Emmerichs D J, Moosmann F, et al. SLIM: Self-supervised LiDAR scene flow and motion segmentation[C]//Proceedings of the IEEE/CVF International Conference on Computer Vision, 2021: 13126-13136. [98] Behl A, Paschalidou D, Donné S, et al. Pointflownet: Learning representations for rigid motion estimation from point clouds[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2019: 7962-7971. [99] Milioto A, Vizzo I, Behley J, et al. Rangenet++: Fast and accurate lidar semantic segmentation[C]//2019 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), IEEE, 2019: 4213-4220. [100] Cortinhal T, Tzelepis G, Erdal Aksoy E. SalsaNext: Fast, uncertainty-aware semantic segmentation of LiDAR point clouds[C]//International Symposium on Visual Computing. Cham: Springer, 2020: 207-222. [101] Li S, Chen X, Liu Y, et al. Multi-scale interaction for real-time lidar data segmentation on an embedded platform [J]. IEEE Robotics and Automation Letters, 2021, 7(2): 738-745. doi: 10.1109/lra.2021.3132059 [102] Kim J, Woo J, Im S. RVMOS: Range-view moving object segmentation leveraged by semantic and motion features [J]. IEEE Robotics and Automation Letters, 2022, 7(3): 8044-8051. doi: 10.1109/lra.2022.3186080 [103] Chen X, Li S, Mersch B, et al. Moving object segmentation in 3D LiDAR data: A learning-based approach exploiting sequential data [J]. IEEE Robotics and Automation Letters, 2021, 6(4): 6529-6536. doi: 10.1109/lra.2021.3093567 [104] Mersch B, Chen X, Vizzo I, et al. Receding moving object segmentation in 3D lidar data using sparse 4D convolutions [J]. IEEE Robotics and Automation Letters, 2022, 7(3): 7503-7510. doi: 10.1109/lra.2022.3183245 [105] Sun J, Dai Y, Zhang X, et al. Efficient spatial-temporal information fusion for Lidar-based 3d moving object segmentation[C]//2022 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), IEEE, 2022: 11456-11463. [106] Hu H K, Sun S, Lin H Z, et al. Denoising ghost imaging under a small sampling rate via deep learning for tracking and imaging moving objects [J]. Optics Express, 2020, 28(25): 37284-37293. doi: 10.1364/oe.412597 [107] Wang S, Li L, Yu Z, et al. Image-free target classification with semi-active laser detection system [J]. IEEE Sensors Journal, 2022, 22(23): 22331642. doi: 10.1109/jsen.2022.3217281 [108] Chen Y, Tang J, Jiang C, et al. The accuracy comparison of three simultaneous localization and mapping (SLAM)-based indoor mapping technologies [J]. Sensors, 2018, 10(18): 3225. [109] Yin T, Zhou X, Krähenbühl P. Multimodal virtual point 3D detection[C]//Advances in Neural Information Processing Systems, 2021, 34: 16494-16507.