Automatic buildings detection using Sobel, Roberts, Canny and Prewwitt detector

Libor Ladányi – Martin Králik

   This work deals with the possibilities of contemporary automatic identification of objects. Automatic object identification can be done by two computational procedures, namely object detection and object recognition. This work deals with the automatic buildings detection, specifically. Presented detection is performed using the edge detectors, namely Prewwitt, Roberts, Canny and Sobel. The main goal of our work was to automate the device for the detection of hazardous substances in the air, as the detection of hazardous substances is realized by laser-based CBRN (Chemical, Biological, Radiological and Nuclear) stand-off detectors, which evaluate the measured data from the reflected laser beam. In this case, buildings are the most reflective surfaces. In order to detect a building, it is necessary to find a suitable edge detector to be used in further research and serve as a basis for software solution of automatic identification.

Keywords: automatic object recognition, object recognition, buildings recognition, methods based on feature points

[full-paper]