Development of a comprehensive model for concurrent water surface and bottom reconstruction in photo bathymetry (PhotoBathyWave)

在摄影测深学中，摄影测量技术用于从通过水面拍摄的陆地和航空图像中对水底地形进行三维测量。根据斯涅耳定律，在不同光学介质环境中进行几何建模必须考虑像线折射。这需要存在分离光学介质的界面的精确几何模型。如果图像是通过自由流动的水面拍摄的，波浪的存在严重阻碍了几何建模。到目前为止，只有近似的解决方案存在这个任务。拟议项目的目标是开发一个通用的和多功能的方法，通过开放的动态波浪影响的水面多媒体摄影测量。我们的目标是同时确定水面的形状和淹没的水底地形从多视角图像中的综合摄影测量参数估计过程。这就需要利用其折射或反射特性作为严格考虑多射线空间相交中的光路上的折射的基础，对不稳定的水面进行时空建模，以精确地确定无偏的水底坐标。将特别强调参数的可确定性，以避免过度参数化，其中包括适当的随机建模和规划的测量配置。在一般情况下，这将被解决在一个综合的方法，同时确定图像的方向和水的表面和底部的点坐标在一个扩展的光束区域网平差或在一个解耦的处理framework.While广泛的机构存在的文献建模无论是水面或底部的几何形状与图像为基础的技术，我们的目标是估计从相同的立体图像。考虑复杂的，动态波影响的水面在光测深处理管道构成一个创新的方法，它有可能（一）大幅提高精度水平的一个数量级和（二）大大扩展的应用领域的光测深。从研究成果中受益的应用领域包括水文生态学和水文形态学，沿海地区的航空测绘，基于无人机的浅水内陆水体测绘，珊瑚礁变化检测或水力机械实验室通道设施中的多种测量任务。

In photo bathymetry, photogrammetric techniques are used for 3D measurements of water bottom topography from terrestrial and aerial images taken through water surfaces. Geometric modelling in environments with different optical media must consider image ray refraction according to Snell’s Law. This requires the existence of a precise geometric model of the interface separating the optical media. If images are taken through a free-flowing water surface, the presence of waves severely impedes geometric modelling. So far, only approximate solutions exist for this task. The objective of the proposed project is to develop a generic and versatile approach for multimedia photogrammetry through open dynamic wave-affected water surfaces. Our goal is to simultaneously determine both the shape of the water surface and the submerged water bottom topography from multi-view images in an integrated photogrammetric parameter estimation procedure. This necessitates spatio-temporal modelling of an instationary water surface exploiting its refractive or reflective property as a basis for strict consideration of refraction on the optical paths in multi-ray spatial intersection for precise unbiased water bottom coordinate determination. Special emphasis will be laid on parameter determinability in order to avoid over-parametrization, which includes appropriate stochastic modelling and planning of the measurement configuration. In general, this will be addressed in an integrated approach for simultaneously determining image orientations and water surface and bottom point coordinates within an extended bundle block adjustment or in a decoupled processing framework.While a broad body of literature exists for modelling either water surface or bottom geometry with image based techniques, our aim is to estimate both from the same stereo images. The consideration of complex, dynamic wave effected water surfaces in photo bathymetry processing pipelines constitutes an innovative approach, which has the potential (i) to substantially raise the accuracy level by one order of magnitude and to (ii) considerably extend the application fields of photo bathymetry. Application fields benefiting from the research outcomes include hydroecology and hydromorphology, airborne mapping of coastal zones, UAV-based mapping of shallow inland water bodies, coral reef change detection, or manifold measurement tasks in hydromechanical laboratory channel facilities.