Intelligent image compression deals with ‘real-world’ wireless links
08 April 2010
Plextek has developed a suite of image processing functions that allows high quality video to be sent over data links with widely varying bit rates, such as wireless links deployed in demanding environments. By identifying objects and areas of interest in the image and intelligently assigning different amounts of image compression, video quality can be maintained for important image areas whilst reducing it in others. Together with real-time motion and target tracking to identify important elements of the scene, the system can adapt dynamically as the link quality varies, allowing users freedom to concentrate on the important aspects of a scene.
“Many image compression techniques are designed to compress the entire area of the scene at a fixed level of quality - but when the data link varies, the severe compression may not give enough quality to evaluate important objects in the scene,” says Graham Tootell, Project Engineer in Plextek’s Image Processing group. “Our processing provides much better resolution for specific areas whilst trading it in others, giving better overall system performance. This has significant advantages, for example for cameras deployed in the field for video and security surveillance.”
Plextek’s software processing can be applied to a wide range of cameras and sensors, including high-definition cameras incorporating mega-pixel sensors. In this situation the processing can maintain the higher definition in specific areas but allowing the use of standard video channels.
Plextek has traditionally had a strong video imaging capability but recently it has become a primary criterion in many of the new project briefs the company is seeing.
Colin Forster, Senior Project Engineer at Plextek comments: “Video imaging is a big focus area for Plextek during 2010. As well as working on bespoke customer projects, we will also soon be launching a series of customisable off-the-shelf platforms for applications such as portable high definition camera surveillance and precision medical instrumentation. We encourage any companies out there currently facing a challenging video imaging design dilemma to get in touch with our highly skilled and experienced team.”
Other current projects include work on vehicle tracking for motorways to highlight potential incident situations, 3D processing for biometric and medical measurements, and low-latency imaging for safety critical applications. Safety critical applications using existing video compression methods such as M-JPEG, MPEG and H264 are not always the best solution where short delays and rapid recovery from signal loss is required. More specialised methods such as wavelet compression matched to highly-robust transmission protocols may therefore be needed, to minimise the visual impact of errors and provide rapid system recovery.
Contact Details and Archive...