Road marking extraction from in-vehicle video
AssigneeIntel Corporation
InventorsAlexander Bovyrin, Nikita Manovich
Life of this patent
Estimating the term…
Abstract
A two-stage pipeline includes a first stage that is a static filter applied to an inverse perspective mapping (IPM) image. A SLAT (Symmetrical Local Adaptive Threshold) method, according to some embodiments, is more adaptive to environment changes. Additional geometry restrictions in the static filter remove clutter. The second stage of the pipeline is a dynamic filter. The motion of blobs is analyzed to find regions that look and behave like road marking.
Claims24
A method comprising:
- applying a static filter to a current input image by:
- extracting connected components of bright pixels in the current input image, wherein extracting the connected components of bright pixels includes dynamically determining a threshold to distinguish bright pixels;
- applying geometric restrictions to remove connected components that do not look like road marking;
- and
- after applying the geometric restrictions, outputting static filtered connected components in the current input image;
- and
- applying a dynamic filter by:
- matching the static filtered connected components in the current input image to connected components of bright pixels from a previous input image;
- building a trajectory for each of the static filtered connected components in the current input image based on the matching;
- and
- removing any static filtered connected components with trajectories that are not substantially along a direction in which a vehicle is moving,
- wherein applying the static filter comprises:
- applying a symmetrical local filter with zero threshold;
- analyzing all found bright pixels to find an improved threshold;
- and
- removing all bright pixels that are below the improved threshold.
One or more non-transitory computer readable media storing instructions to perform a sequence comprising:
- applying a static filter to a current input image by:
- extracting connected components of bright pixels in the current input image wherein extracting the connected components of bright pixels includes dynamically determining a threshold to distinguish bright pixels;
- applying geometric restrictions to remove connected components that do not look like road marking;
- and
- after applying the geometric restrictions, outputting static filtered connected components in the current input image;
- and
- applying a dynamic filter by:
- matching the static filtered connected components in the current input image to connected components of pixels from a previous input image;
- building a trajectory for each of the static filtered connected components in the current input image based on the matching;
- and
- removing any static filtered connected components with trajectories that are not substantially along a direction in which a vehicle is moving,
- wherein applying the static filter comprises:
- applying a symmetrical local filter with zero threshold;
- analyzing all found bright pixels to find an improved threshold;
- and
- removing all bright pixels that are below the improved threshold.
An apparatus comprising:
- a processor to:
- apply a static filter to a current input image by:
- extracting connected components of bright pixels, wherein extracting the connected components of bright pixels includes dynamically determining a threshold to distinguish bright pixels,
- applying geometric restrictions to remove connected components that do not look like road marking, and
- after applying the geometric restrictions, outputting static filtered connected components in the current input image;
- and
- apply a dynamic filter by:
- matching the static filtered connected components in the current input image to connected components of bright pixels from a previous input image,
- building a trajectory for each of the static filtered connected components in the current input image based on the matching, and
- removing any static filtered connected components with trajectories that are not substantially along a direction in which a vehicle is moving,
- wherein applying the static filter comprises applying a symmetrical local filter with zero threshold, analyzing all found bright pixels to find an improved threshold, and removing all bright pixels that are below the improved threshold;
- and
- a memory coupled to said processor.
Dependencies are inferred from the claim text.