Images and processing routines
This section will provide a summary of the important steps in processing UAV images prior to information retrieval. The primary output of the processing steps is normally an orthophoto (i.e. a seamless image that has been corrected for terrain distortions), from which several other products/information can be derived. Best practices, mostly learned from the STARS project, will be shared.
UAVs come with associated software packages for processing the images. Within the STARS project, Pix4D and AGISoft were the two main applications used. Other software packages used include: CAN-EYE and Microsoft ICE. The purpose of this portal is not to reproduce the respective manuals of these applications, but to share with the reader experiences gained in dealing with aspects of the processing steps that could be tricky.
Irrespective of the software used, there are generally four steps to be followed. These are:
- loading and stitching images,
- generating and editing point cloud,
- digital surface model generation and
- making orthophotos.