Thanks to the development of different internal R&D projects in recent years, DIELMO 3D has a solid base on which to develop applications based on artificial intelligence from alphanumeric data, images, videos and/or geospatial data.
It allows to detect a rectangle that encompasses the object to be detected.
It is suitable for multiple applications such as cropping an image to the surroundings of the detected object, blurring faces and license plates in images or videos, etc.
Based on these detections, we can implement quality controls to confirm that an object is present in thousands of images, count objects, select the best image among several images of the same object, etc.
It allows to identify the pixels within the image or video that correspond to the detected object in addition to the rectangle that surrounds it.
It is suitable for multiple applications such as vectorizing buildings or other types of elements in an orthophoto, obtaining a geo-referenced vector layer.
Based on these detections we can automate processes for the detection of illegal constructions, quality controls in the production of vectorial cartography or automatically avoid that the stitching lines in an orthophoto pass inside buildings.
We also have experience in training models for LiDAR point cloud classification to distinguish different types of features such as soil, buildings, vegetation, towers and power lines.
In this way we can improve conventional automatic classifications and reduce manual classification time.
From this automatic classification we can also obtain vector layers with the position of power lines or use input vector layers with the precise position of towers and cables to improve automatic classifications.
Getting started with AI.dielmo
Anonymization of images or videos
Draw detections in image
Blur detections in image
Draw detections in video
Blur detections in video