本文为澳大利亚塔斯马尼亚大学（作者：Stephen John Harwin）的博士论文，共209页。

无人机已成为一种具有成本效益的测绘工具。利用多视立体视觉（MVS）进行无人机摄影测量，简称UAV-MVS，是摄影测量与计算机视觉相结合的一种技术，在三维重建测量中越来越流行。使用无人机MVS进行三维重建需要严格的精度评估。

本文评估了无人机测量设计中需要考虑的因素（摄像机网络、摄像机标定和地面控制网络），并评估了它们对MVS点云精度的影响。本论文的目的是评估UAV-MVS测量技术的准确性，以便更好地了解可以检测到的变化尺度。选定的应用区域是自然地形变化，这种情况是一段受保护的海岸线正在以卫星或航空摄影难以监测的尺度被侵蚀。量化这些脆弱的防护海岸沿线发生侵蚀的空间和时间尺度，可以深入了解自然景观对许多不同变量（包括海平面上升）的响应。

侵蚀发生在分米和厘米级别，用低空UAV-MVS进行测量（精确控制的高度误差小于5-6毫米（σ=1-2毫米），使用差分GPS等效控制时小于10-11毫米（σ=22毫米）。这些测试中的飞行高度为地面以上20-25米。网络仿真和经验数据的比较表明，在使用差分GPS（σ=22mm）建立地面控制的典型无人机应用中，仿真可以可靠地预测目标的空间精度。然而，当使用精确的实地测量方法（σ=1–2 mm）建立地面控制时，模拟预测的可靠性较低，所获得的精度低于预测精度。这是由于摄像机标定中残余系统误差的影响。

研究结果进一步表明，70-80%的重叠最低点摄影加上倾斜摄影聚焦在地形的复杂部分，提供了这个海岸线最准确和完整的三维重建。点云差分是检测和量化变化的有效手段。当以UAV-MVS数据提供的精细比例尺提取海岸线时，陡坎边缘比植被边缘更容易描绘，并提供更准确的海岸线位置和相关变化指示。如果遵循本文提供的设计指南，可以放心地使用无人机进行测量和监测。利用UAV-MVS监测受保护海岸线的厘米级变化，可以提供区分事件驱动的变化和长期趋势所需的空间/时间分辨率数据集。

Unmanned aerial vehicles (UAVs) or droneshave become a cost-effective tool for surveying and mapping. UAV photogrammetryusing Multi-View Stereopsis (MVS), known as UAV-MVS, is a technique thatcombines photogrammetry and computer vision and is becoming increasinglypopular for 3D reconstruction surveys. There is a need for rigorous accuracyassessment of 3D reconstructions using UAV-MVS. This thesis evaluates UAVsurvey design considerations (camera network, camera calibration, and groundcontrol network) and assesses their impact on the accuracy of MVS point clouds.The aim of this thesis is to assess the accuracy of the UAV-MVS surveytechnique to better understand the scale of change that can be detected. Thechosen application area is natural landform change, in this case of a sectionof sheltered coastline that is eroding at scales that are difficult to monitorfrom satellite or aerial photography. Quantifying the spatial and temporalscales of the erosion occurring along these often fragile sheltered coastsprovides an insight into the response of the landscape to many differentvariables including sea level rise. The erosion is occurring at the decimetreand centimetre scale and UAV-MVS from low altitude (< 5–6 mm when usingprecise control (σ = 1–2 mm) and 10–11 mm when using differential GPSequivalent control (σ = 22 mm). The flying height in these tests was ～20–25 m above terrain. Comparisons of network simulations and empiricaldata demonstrate that a simulation can be used to reliably predict object spaceaccuracy for typical UAV operations, where ground control is being establishedusing differential GPS (σ = 22 mm). However, simulation predictions are lessreliable when the ground control is established using precise field surveymethods (σ = 1–2 mm), with achieved accuracy lower than predicted precision.This is attributed to the influence of residual systematic errors in cameracalibration. The findings further demonstrate that 70–80% overlap nadirphotography supplemented with oblique photography focussed on complex portionsof the terrain provides the most accurate and complete 3D reconstructions ofthis coastal shoreline. Point cloud differencing is an effective means ofdetecting and quantifying change. When extracting a proxy for shoreline at thefine-scale provided by UAV-MVS data, the scarp edge is more easily delineatedthan vegetation edge and provides a more accurate indication of shorelineposition and associated change. The use of UAVs for surveying and monitoringcan now be undertaken with confidence provided the design guidelines offered inthis thesis are adhered to. The use of UAV-MVS to monitor centimetrescalechange along sheltered coastlines can provide the spatial and temporalresolution datasets needed to distinguish event-driven change from longer termtrends.

1. 引言

2. 基于无人机图像的多角度立体视觉（MVS）生成的地理参考点云精度评估

3. 无人机多视立体图（UAV-MVS）和地面多视立体图（T-MVS）点云的地表表示评估

4. 校准方法对无人机多视立体视觉（UAV-MVS）点云精度的影响

5. 利用无人机图像进行自然景观三维重建的飞行轨迹形态摄影测量模拟

6. 无人机多视立体视觉（UAV-MVS）点云中的海岸变化探测

7. 结论

更多精彩文章请关注公众号：