Statistical pressure snakes based on color images.
- ORION International Technologies, Albuquerque, NM
The traditional mono-color statistical pressure snake was modified to function on a color image with target errors defined in HSV color space. Large variations in target lighting and shading are permitted if the target color is only specified in terms of hue. This method works well with custom targets where the target is surrounded by a color of a very different hue. A significant robustness increase is achieved in the computer vision capability to track a specific target in an unstructured, outdoor environment. By specifying the target color to contain hue, saturation and intensity values, it is possible to establish a reasonably robust method to track general image features of a single color. This method is convenient to allow the operator to select arbitrary targets, or sections of a target, which have a common color. Further, a modification to the standard pixel averaging routine is introduced which allows the target to be specified not only in terms of a single color, but also using a list of colors. These algorithms were tested and verified by using a web camera attached to a personal computer.
- Research Organization:
- ORION International Technologies, Albuquerque, NM; Sandia National Lab. (SNL-NM), Albuquerque, NM (United States)
- Sponsoring Organization:
- USDOE
- DOE Contract Number:
- DE-AC04-94AL85000
- OSTI ID:
- 1145144
- Report Number(s):
- SAND2004-1867; 267607
- Country of Publication:
- United States
- Language:
- English
Similar Records
Extracting primary features of a statistical pressure snake.
Matching a statistical pressure snake to a four-sided polygon and estimating the polygon corners.