Our idea is to identify and geolocate manholes, inlets, traffic signals, streetlights, curbs, pavements, and fire hydrants using drones that can be utilized by civil engineering and construction firms undertaking large-scale land development projects for municipalities, townships, or counties.
While there are solutions for aerial surveys that work well with larger objects, identifying and precisely locating smaller objects in aerial images is a challenging task due to a high degree of variance in object colors, background, surroundings, shadows, and aspect ratios. Our solution is specifically focused on detecting smaller objects such as manholes, inlets, curbs, streetlights, and traffic lights. The specific nature of our solution would greatly help civil engineering and construction firms in eliminating errors that arise from outdated surveys and mapping. It can help eliminate rework costs resulting from damage caused to existing infrastructure, such as sewer lines connecting manholes, storm drain system connecting inlets, underground cable connecting streetlights and traffic lights, due to outdated or incorrect geographic location which can run above $10,000 for a single site, and reduce the time required for surveying by up to 75%.
This is an emerging area with several companies using drone technology in many different domains for aerial imaging, mapping and modeling. Our key differentiator is the use of drone technology in combination with machine learning and computer vision in surveying land development sites and mapping objects of interest that are relatively smaller in size and harder to identify. The advantage is the precision it can offer in locating existing infrastructure, such as manholes, inlets, streetlights, traffic lights, etc., that can help decision-makers giving them spatial data to oversee areas relevant to project planning. Currently the surveying and mapping is largely carried out manually and is not only time consuming but also subject to human error.
SkaiVue is a drone-based system that captures aerial shots of the land being surveyed and uses machine learning for object detection along with their highly accurate location.