TOBCAT - Industrial applications of object categorisation (R-4479)

Detecting objects in (camera) images is already widely used for some time in the industry, insofar as the outlook of the object to be detected is well know in advance. In this project we will go a step further. If the objects that are to be detected vary a lot in outlook, object categorization techniques are necessary, a group of object detection approaches who recently came up in the academic research domain of computer vision. We see a great potential for these object categorization techniques. There are many industrial sectors where objects must be detected which vary in their oulook: plants and animals in the agricultural sector, for product inspection of food products, industrial automation, personal detection for monitoring applications in health care and in security cameras, traffic analysis, processing of remote sensing and mobile mapping-images, nature-monitoring, microscope images of micro-organisms, luggage control at airports, ... In this project we will design and evaluate a number of such applications as test case, in cooperation with companies from different sectors. IMOB is responsible for: a) the detection of different classes of road users (passenger cars, lorries, cyclists, pedestrians, ...) based on video images. b) registering relevant traffic parameters such as speeds and distances between moving and fixed objects (vehicles and fixed obstacles).

Keywords:object categorisation

Disciplines:Economics and business

Project type:Collaboration project