Abstract: A method for detecting wheels of a vehicle on a road comprises directing a vehicle classification sensor onto a section of the road and recording a 3D representation of the vehicle, directing a camera onto the section and recording a source 2D image, determining a bounding box circumscribing the 3D representation, a side and four corner points of said bounding box side, identifying corner points in a source 2D image plane, defining four corner points of a rectangle in a destination 2D image plane, calculating a projective transformation between the source corner image points and the destination corner image points, transforming source 2D image pixels to destination 2D image pixels of a first destination 2D image, and detecting the wheels using the destination 2D image.

Abstract: A method for detecting wheels of a vehicle on a road comprises directing a vehicle classification sensor onto a section of the road and recording a 3D representation of the vehicle, directing a camera onto the section and recording a source 2D image, determining a bounding box circumscribing the 3D representation, a side and four corner points of said bounding box side, identifying corner points in a source 2D image plane, defining four corner points of a rectangle in a destination 2D image plane, calculating a projective transformation between the source corner image points and the destination corner image points, transforming source 2D image pixels to destination 2D image pixels of a first destination 2D image, and detecting the wheels using the destination 2D image.

Abstract: The present disclosure relates to a method of controlling a traffic surveillance system. The method comprises the steps of: capturing first images and second images over time by a plurality of the stereoscopic sensors; processing, by the processing unit, a first image and a second image from a first stereoscopic sensor of the plurality of the stereoscopic sensors to produce a first height image; analyzing said first height image to detect a moving object such as a vehicle located within the primary view; and analysing a part of a primary view of at least a first image, a second image or a combination of the first image and the second image captured by a second stereoscopic sensor based on the detected moving object in said first height image to determine a characteristic of the moving object.

Abstract: The present disclosure relates to a method of controlling a traffic surveillance system. The method comprises the steps of: capturing first images and second images over time by a plurality of the stereoscopic sensors; processing, by the processing unit, a first image and a second image from a first stereoscopic sensor of the plurality of the stereoscopic sensors to produce a first height image; analyzing said first height image to detect a moving object such as a vehicle located within the primary view; and analysing a part of a primary view of at least a first image, a second image or a combination of the first image and the second image captured by a second stereoscopic sensor based on the detected moving object in said first height image to determine a characteristic of the moving object.