Normalization method for video images
The present invention relates to a method and apparatus for automatically and adaptively normalizing analog signals representative of video images in object detection systems. Such normalization maximizes the average information content of the video images and, thereby, provides optimal digitized images for object detection and identification. The present invention manipulates two system control signals -- gain control signal and offset control signal -- to convert an analog image signal into a transformed analog image signal, such that the corresponding digitized image contains the maximum amount of information achievable with a conventional object detection system. In some embodiments of the present invention, information content is measured using parameters selected from image entropy, image mean, and image variance.
- Research Organization:
- Sandia National Labs., Albuquerque, NM (United States)
- Sponsoring Organization:
- USDOE, Washington, DC (United States)
- DOE Contract Number:
- AC04-94AL85000; AC04-76DP00789
- Assignee:
- USDOE, Washington, DC (United States)
- Patent Number(s):
- PATENTS-US-A7863928
- Application Number:
- ON: DE94010161; PAN: 7-863,928; BR: GB0103012
- OSTI ID:
- 10143348
- Resource Relation:
- Other Information: PBD: 1992
- Country of Publication:
- United States
- Language:
- English
Similar Records
Video image position determination
Method and apparatus for reading meters from a video image