条纹阵列探测激光雷达测距精度与三维测绘技术研究

发布时间：2018-09-01 06:15

【摘要】：条纹阵列探测激光雷达具有作用距离远、探测视场宽、距离景深大和数据率高等优点,在远距离目标三维重建和高空机载宽幅测绘中具有重要的应用前景。在现有的条纹阵列探测激光雷达研究工作中,始终未能建立一个完善的测距精度理论模型,同时还缺乏可有效提高测距精度的技术手段,也没有根据系统的成像特性对其三维测绘技术进行优化和改进。针对这些问题,本文对条纹阵列探测激光雷达的测距精度及其三维测绘技术进行了理论和实验研究。本文从条纹阵列探测激光雷达的距离分辨原理和级联成像过程出发,结合激光雷达方程和探测器的线扩展函数,推导出单个时间分辨通道内信号分布函数的理论表达式。发现信号强度的分布函数服从高斯分布,并且其条纹宽度由激光脉冲宽度和探测器的线扩展函数共同决定。根据噪声与信号强度的相关性对雷达系统的噪声特性进行了分类讨论,并分别给出了乘性噪声和加性噪声的均值和方差的理论表达式,进而建立了含噪条纹信号的理论模型,为系统的测距精度分析和探测过程研究提供了必要的理论依据。针对条纹阵列探测激光雷达自身的成像特性,在信号探测过程中建立了恒定发射功率和近饱和成像两种的工作模式,从而可在不同的探测距离和应用需要下获得更为理想的原始条纹图像。根据系统的信号和噪声特性,以质心权重算法作为信号鉴别方式,利用误差传递理论建立了测距精度的理论模型。分别在两种不同工作模式下推导出加性噪声引起的误差、乘性噪声引起的误差以及采样误差的理论表达式。利用平面目标距离成像结果的残差抖动分析方法,对测距精度的理论模型进行了仿真和实验验证。在仿真验证中,通过模拟激光的发射、接收过程和探测器的信号采集过程,对三类主要误差随系统各关键参量的变化关系进行了探讨。在实验验证中,建立了一套可对噪声源进行选择控制的条纹阵列探测激光雷达原理验证系统,并着重讨论了条纹宽度对测距精度的影响。实验结果表明,在恒定发射功率工作模式下,乘性噪声引起的误差随条纹宽度的增加而线性增加,加性噪声引起的误差不随条纹宽度的变化而发生改变;在近饱和成像工作模式下,乘性噪声引起的误差与条纹宽度的平方根成正比,加性噪声引起的误差则与条纹宽度成反比。基于测距精度的理论模型,本文提出了四种可提高系统距离分辨能力和目标识别能力的测距精度优化方法:(1)研究了基于最优条纹宽度的参量优化方法,给出了不同工作模式下最优条纹宽度的理论表达式和数值计算结果。实验上,利用聚焦极电压时空轴分离调节方法,在保证探测器具有最高空间分辨率的前提下实现了对时间轴线扩展函数的独立控制,进而获得了具有最优条纹宽度的回波信号。参量优化后,系统可在173m的距离门宽下对1.7km的远距离目标实现分米级的距离像重建,将测距均方根误差降低至0.19m;(2)在条纹图像的距离提取过程中提出了一种最优阈值方法。基于仿真模拟结果讨论了不同条纹宽度和噪声强度下阈值设定对测距精度的影响,建立了最优阈值的经验公式,并利用最优阈值方法抑制了外场测绘中强背景噪声对距离像的干扰;(3)讨论时隙宽度的最佳取值方案。仿真和实验结果表明,在保证距离门宽度足够覆盖被测目标景深的前提下,可通过减小时隙宽度来获得更高的测距精度。在室内小景深测距实验中,利用0.07ns的时隙宽度可使系统的测距均方根误差减小至3mm;(4)提出了一种基于迭代加权质心算法的超分辨率信号鉴别方法,并分析了该算法在距离提取过程中的作用机制。通过引入高斯型加权因子,该算法有效抑制了由于临近通道的条纹弥散所引起的边界模糊效应,实现了对1.4km被测目标特征细节的准确识别。实验结果表明,当条纹宽度为4.4ns时,经过15次迭代运算可将系统的测距均方根误差降低至0.15m,测绘结果的距离精度优于系统的最小距离分辨率,实现了条纹阵列探测激光雷达的超分辨率成像。最后,本文建立了一套搭载于飞机平台的对地三维测绘系统,并根据测距精度优化方法和回波强度计算结果讨论了不同飞行高度下系统工作模式的选择以及系统关键参量的取值问题。针对机载线阵扫描体制下测绘幅宽受到探测器视场角限制这一技术难题,提出了一种新型的扫帚式扫描体制,并结合数字高程地图仿真分析了该扫描体制下激光脚点对测绘区域的覆盖能力。对探测器扫描电场不均匀性引起的测距偏差和激光脚点扫描轨迹非线性引起的水平定位偏差进行了定标修正。利用优化和定标后的系统实现了距离门宽为173m、数据率为500k Hz的高精度机载宽幅测绘,在3000m和5800m的飞行高度下,测距均方根误差分别可以达到0.11m和0.16m,对面积为30km2的平原区域进行测绘所需的探测时间为2分13秒,相比于传统的推帚式扫描体制测绘效率提高了近10倍。
[Abstract]:The stripe array detection lidar has many advantages, such as long range, wide detection field, large range depth and high data rate. It has important application prospects in three-dimensional reconstruction of long-range targets and airborne wide-range mapping. In this paper, the ranging accuracy and three-dimensional mapping technology of the stripe array detection lidar are studied theoretically and experimentally. Based on the range resolution principle and the cascade imaging process of the column detection lidar, combining with the lidar equation and the linear spread function of the detector, the theoretical expression of the signal distribution function in a single time-resolved channel is derived. The noise characteristics of radar system are classified and discussed according to the correlation between noise and signal strength. The theoretical expressions of mean and variance of multiplicative noise and additive noise are given respectively. Then the theoretical model of noisy fringe signal is established to analyze and detect the ranging accuracy of the system. According to the imaging characteristics of the stripe array detection lidar, two working modes, constant transmitting power and near saturation imaging, are established in the signal detection process, so that more ideal original stripe images can be obtained at different detection distances and application requirements. The theoretical model of ranging accuracy is established by using the error transfer theory. The theoretical expressions of additive noise error, multiplicative noise error and sampling error are deduced under two different operating modes respectively. The theoretical model of ranging accuracy is simulated and verified by experiment. In the simulation, the relationship between the three main errors and the key parameters of the system is discussed by simulating the process of laser emission, receiving and signal acquisition of the detector. The principle validation system of fringe array detection lidar with selective control of noise sources is presented, and the effect of fringe width on ranging accuracy is discussed emphatically. The experimental results show that the error caused by multiplicative noise increases linearly with the increase of fringe width and the error caused by additive noise does not with the fringe width under constant transmitting power mode. The error caused by multiplicative noise is directly proportional to the square root of the fringe width and the error caused by additive noise is inversely proportional to the fringe width in near saturation imaging mode. Distance precision optimization methods: (1) The parameter optimization method based on the optimal fringe width is studied, and the theoretical expression and numerical results of the optimal fringe width under different operating modes are given. After parameter optimization, the system can reconstruct the range profile of 1.7 km long-range target at a distance gate width of 173 m, and reduce the ranging root mean square error to 0.19 M. (2) A new method is proposed in the range extraction process of fringe image. Based on the simulation results, the influence of threshold setting on ranging accuracy under different fringe widths and noise intensities is discussed, and the empirical formula of the optimal threshold is established, and the interference of strong background noise on range profile in outfield mapping is suppressed by the optimal threshold method. (3) The optimal scheme of time slot width is discussed. The experimental results show that higher ranging accuracy can be achieved by reducing the time slot width while the width of the distance gate is enough to cover the depth of field of the target to be measured. By introducing Gaussian weighting factor, this algorithm can effectively suppress the boundary blur effect caused by fringe dispersion in the adjacent channel, and realize the accurate recognition of 1.4 km target feature details. When the fringe width is 4.4ns, the RMS error of the system can be reduced to 0.15m after 15 iterations. The range precision of the mapping result is better than the minimum range resolution of the system, and the super-resolution imaging of the fringe array detection lidar is realized. Finally, a set of three-dimensional mapping system for the ground based on the aircraft platform is established. According to the optimization method of ranging accuracy and the results of echo intensity calculation, the selection of operating mode and the selection of key parameters of the system at different flight altitudes are discussed. The coverage of the laser footprint to the surveying and mapping area under the scanning system is analyzed by combining with the digital elevation map simulation. The ranging deviation caused by the non-uniformity of the scanning electric field of the detector and the horizontal positioning deviation caused by the non-linearity of the scanning trajectory of the laser footprint are calibrated and corrected. The gate width is 173m and the data rate is 500kHz. At the flight altitudes of 3000m and 5800m, the RMS error of ranging can reach 0.11M and 0.16m, respectively. The detection time for surveying and mapping the plain area of 30km2 is 2 minutes and 13 seconds, which is nearly 10 times more efficient than the traditional broom scanning system.
【学位授予单位】：哈尔滨工业大学
【学位级别】：博士
【学位授予年份】：2016
【分类号】：TN958.98

【相似文献】