Roke | Image Processing

Corner / Feature Detection & Edge Detection	Corner / Feature Detection reduces the image to a number of point features which can then be tracked or used to extract 3D information from a scene. The most commonly used method of feature extraction is our patented Harris Corner Detector. Edge Detection finds lines in an image which allows us to identify particular shapes. Alternatively, it can be used in combination with Corner Detection to define the properties of a scene.
Motion Analysis	Motion Analysis is used to predict motion in a scene. By tracking the movement of features in a scene the motion of the camera or an object can be extracted. Motion Analysis therefore gives us a powerful tool as it can be used to predict further motion, or used to determine the global motion of the camera and compensate for it in order to calculate the motion of the objects in the scene.
Deblurring	Deblurring is used to enhance an image. Where an image is unclear or blurred due to poor resolution, our algorithms will reduce the blurring and allow additional information to be obtained from it.
Image Stabilisation	Image Stabilisation removes the effects of unintentional camera movement. Since cameras are often used or installed in places which are prone to shaking or vibration, our Image Stabilisation techniques are used to remove these effects.
Camera Calibration	Camera Calibration is the determination of the cameras' pan, tilt and zoom settings. This is necessary if information about the size and position of an object is required from the scene.
3D Geometry	3D Geometry uses the camera calibration to convert objects seen in 2D images into real world coordinates. If a number of cameras are used we are able to combine these measurements in order to extract further information.

Our advanced and proven capabilities are put to use in a variety of ways. Here is a short list of technologies and tools developed by us to help our customers.

DROID
Structure from Motion	Reconstructs its surrounding 3D environment using only images from a vision system. This can be used to provide robots with sight, allowing them to navigate.

RAPiD
3D Object Tracking	Accurately locks onto and tracks the position and orientation of a known object. Has been used in computer assisted surgery and Unmanned Air Vehicle landing systems.

OSCAR
Measurements from Photographs	A photogrammetric tool which enables distances between objects in a set of photographs to be determined. Applications include crime scene data extraction, and reconstruction of 3D models. Vision Processing for Measurement, Inspection and Evaluation

VMAD

Video Motion Anomaly Detector

VMAD learns 'normal' movements in a scene and triggers an alert when unusual activity occurs. Can be used for perimeter monitoring, and wrong-way driving, for example.

VMAD Datasheet

Image Processing

Technologies

More Information