Supercam-Neuro Artificial neural network is an integral fully autonomous system for automatic analysis of the photo and video data using a deconvolutional artificial neural network.
Supercam-Neuro allows for automatic interpretation of photo images captured with the UAV – analyzing all images obtained during the flight, counting machinery, humans, animals, generating reports including the coordinates of every detected object in automatic mode, as well as generating primary orthomosaics.
Automatic interpretation reduces the number of human errors, as well as significantly decreases the time of receiving information after the flight.
The interpretation is performed using the pre-trained profiles. This system is constantly being trained and allows to integrate new data analysis profiles. Upon a customer’s request, it is possible to add additional objects of search. To train a profile, you will need 500-700 portraits of objects with at least 28 pixels in size as per the resolution requirements.
Based on the survey data (photo images and telemetry), it is possible to carry out express stitching of photo data into a primary orthomosaic. It will take about 15-20 minutes to generate an orthomosaic from 1000-1500 images (depending on the equipment).