19 Jun 2012
CVB Video Stabilizer is a high performance tool for recording and compensating for image transformations caused by the movement of the camera. The algorithms used in the new CVB tool allow the developer to record image shifts rapidly and reliably up to half the image width or height respectively, with a resolution of up to 0.1 pixels, and to determine rotations between two images in a range between -90° and +90° with an angle resolution of 0.1°. Additionally the tool can detect changes in image size of up to 30% with a resolution of the scaling factor of 0.001.
This new tool can also be used to compensate for high frequency camera movements, providing simultaneous gentle ‘guiding’ of the camera image. Previously this could only be achieved by using mechanical SteadyCam systems.
The CVB Video Stabilizer uses a correlation-based analytical procedure. Therefore, no markers are needed, nor are predetermined known structures within the image required to perform the detection of image transformations. Instead, all of the image information is used to determine the transformation parameters and the dominant changes in the image are recorded automatically.
Detection is extremely robust in poor or rapidly changing light conditions. By using CVB?Video?Stabilizer, automated image processing can now be used in situations where it was previously impossible or severely limited due to intrinsic camera movement or oscillation.
The new CVB Optical Flow tool exploits the computer-supported determination of motion components in images, also known in image processing as "optical flow". The local determination of moving image contents is of major importance when viewing and recording surroundings. However, computer-supported determination of optical flow within images is a highly complex task and implementing this function robustly, rapidly and flexibly is extremely challenging. For this reason this functionality is only rarely available as part of image processing libraries.
With CVB Optical Flow, STEMMER IMAGING now provides the developer with an efficient tool for the high quality determination of optical flow. This allows the detection and display of local motion vectors with less than 2° mean deviation when determining the direction, and less than 1/10 pixel deviation when determining the amount of movement.
Compared with conventional methods for detecting movement, such as calculating the difference between images, CVB Optical Flow not only allows detection but also the quantitative determination of moving image contents, making the algorithm robust against poor or rapidly changing light conditions as well as vibrations.
Both tools are ideal for use in a range of industrial markets such as security and traffic control technology. For example, CVB Optical Flow allows the analysis of traffic flow or the movement of people or systems with mobile components.