Project Area

Introduction

This article offers a general description of how SURE handles the Project Area, what ways to specify it does the user have, as well as some benefits gained by using this functionality. The Area is defined in object space (world coordinates) and it indicates that the products selected by the user (e.g. DSM, True Ortho, Point Clouds, Meshes) will be generated within its boundaries.

Important

The products generated by SURE can be georeferenced if a proper Project Coordinate System is provided in WKT format. Please note, this file must refer to the input coordinate system (associated to the Exterior Orientation of the input images) , as SURE does not convert to new coordinate systems. See more information in the dedicated page.

User interaction

If there is no indication from the user, SURE will automatically compute a 3D bounding box, based on the point cloud generated in the initializer. This is available after Analysis stage is done, during which the input images are preliminary processed at low resolution. For all practical purposes, this volume is the most extended domain in which the user can expect 3D data, based on the amount of input images.

Currently the following ways to manually define the Project Area exist:

  • Limited range for X, Y, Z coordinates. The user has the possibility to specify minimum and maximum values for these.
  • Shape file (*.shp) containing a 2D domain which can potentially be defined by multiple 2D polygons. See Shape File Specification
  • Combinations of the above, e.g. Shape file + limiting X/Y/Z range

The actual parameters and interfaces are stated in General Parameters.

Important

Within the control files interface, all the parameters that have an effect on the project processing area are condensed in controlArea.txt. This control file is necessary for all the steps in the standard SURE workflow, the 2.5D Tool, as well as for the Standalone Modules.


Functionality

  • Determining the Project Area

The user should expect results only within the boundaries of what is referred to as Project Area, considering, of course, the captured regions by the input images. The Project Area is determined based on the restrictions that the user sets, with the help of the above mentioned interfaces. Any manually specified Area takes priority over the automatically computed one. 

Examples (not limited to these):

  • The limits set manually to the X/Y/Z coordinates will take priority over the Automatically determined ones;
  • Same as above, a Shape File set as Area will override the Automatic bounding box";
  • A combination of Manually set bounding box with a Shape file will define the Project Area as the intersection between the two of them. If this intersection is the empty set, the user will be warned in this regard and no processing will start.

Important

An initial terrain height can be considered in combination with a 2D area set already known to SURE. This can be either rectangular (manually specified area / bounding box), or more complex (described by a shape file).

  • Selection of Input Images

It is possible to automatically select the input images, which "see" the Project Area, for processing. For that, it is required that either a bounded volume (2D limitation + Z range), or a 2D area, coupled with the input of a terrain height, is specified. Furthermore, SURE performs an image masking, such that only pixels that cover the Project Area will be matched. The image selection and the image masking are implemented both for rectangular 2D area, as well as for polygons contained within a Shape file.

  • Results

Once the Project's processing area is set, its rectangular bounding box will determine the tiling scheme for the results. The tiles for the individual products will be aligned to the upper left corner of this bounding box (North-West). Note that not all tiles may necessarily exist, only those that overlap with the processing area and contain actual data. The user may expect to find results in this region, as long as the input images cover it sufficiently in terms of photogrammetric requirements.

  • Border quality and consistency

In order to ensure quality preservation at the borders of the Project's processing area, a buffer region is considered (padded). As a consequence, the raw Dense Clouds (3D_Points folder) will contain this additional data, as they represent the input to the products further down the processing pipeline. However, in the subsequent processing steps, the padded area will be segregated from the actual data that are situated inside the specified area. Therefore, the deliverable results are to be found by the user in the form, that they contain exclusively data which are inside the Project's processing area, as it was initially defined (i.e. without the padded region). We refer to DSM, True Ortho, DSM Mesh (textured, LOD), Filtered Clouds and Mesh (textured, LOD) as "deliverable results".

The second benefit from the padding occurs in the case of Distributed Processing or automatic project splitting. Because of the buffer area considered around each subproject's processing area, consistency and smooth transition between them is guaranteed. More details about the functionality of the distributed processing are available in the dedicated page.

Best practice

It is recommended to define the Project Area whenever possible. Because of the image selection and image masking, the processing can potentially be sped up significantly (the effect increases with the ratio between the size of the entire data set capturing region and the Project's processing area, defined by the user). Examples include (but are not limited just to) the situations bellow:

  • A manual specification of Project Area is almost always applicable, as photogrammetric flights usually capture a somewhat extended area around the actual region of interest. This means there will be an overhead of data to be processed that can easily be discarded.
  • With the use of Shape files, it is possible to follow the complex borders of Administrative Territorial Units.
  • Because the input images will be automatically selected from the full data set, tests can be performed rapidly on small areas, on full resolution, to find the best parameters and for prompt Quality Inspection.
  • Area based sequential processing can be automatically applied to large projects by making use of the Project Splitting feature. This means the user can have sooner full results for a subsection of the project, before processing the other area based partitions (e.g. for Quality Inspection and/or delivery to customer).