Thanks to feature tracking, bundle adjustement, correlation techniques as well as point clouds and mesh fusion, imajing enables its clients to evolve in a reconstructed 3D environment of their network.
imajing has developed a complete machine learning pipeline for processing different kind of surveys published on imajnet® web platform.
Imajing is constantly enlarging its machine learning models and recognition algorithms to automatically identify, characterize and geo-reference all visible equipment along surveyed zones.
This process is convenient for large data volumes, where many assets have to be geo-referenced. We currently propose a railway model, a road sign model, a road utilites model and a streetlight model. In a recurrent survey approach, machine learning also brings automatic change detection.
To preserve privacy, for images destined to public diffusion, imajing has developed an algorithm to blur faces and car plates that are visible in surveyed images.