Coastline monitoring using computer vision

The CSIR is developing software and algorithms to assist scientists and engineers tasked with monitoring South African coastlines. The oceans economy contributes significantly to GDP, and is a valuable resource that needs ongoing monitoring if it is to be managed effectively.

Researchers are developing computer vision algorithms to support coastal monitoring to protect coastal infrastructure. A particular threat to coastal infrastructure is soil erosion, which can exacerbate rising sea levels. The CSIR has developed an adaptive surf-zone tracker that tracks near-shore regions of breaking waves to monitor soil erosion and a sand segmentation algorithm to assist in sediment transfer studies, which also helps scientists study soil erosion.

The entrances to South African ports are protected by dolos armour breakwaters, and their condition needs to be monitored regularly.  However, these existing monitoring methods are time consuming and expensive. The CSIR has investigated options for breakwater inspection and recommended that 3D models of the breakwater be captured using a 3D camera mounted on a drone. Researchers have developed a software suite allowing drone control and 3D mapping, which has been tested in a simulator developed in-house and at the CSIR's coastal and hydraulics laboratory.



Key Concept

Computer vision
Humans use their eyes to collect information about their surroundings which is processed and interpreted by the brain forming the basis for certain tasks. In engineering, computer vision is an interdisciplinary field that seeks to automate and improve the tasks that the human visual system can do, mostly by developing mathematical algorithms to gain a high-level understanding of an environment based on digital images or videos.