Real time traffic sign recognition
Advanced driver assistance systems can be designed to recognize and to classify traffic signs under real time constraints, and under a wide variety of visual conditions. This disclosure provides techniques that employ binary masks extracted by color space segmentation, with a different binary mask generated for each sign shape. Temporal tracking is employed to add robustness to the detection system. The system is generic, and is trainable to the traffic signs used in various countries.
1. A method comprising:
receiving an image;
extracting binary masks from the image, wherein each of the binary masks is associated with a respective color;
performing shape detection operations on the image using the binary masks such that each of the shape detection operations is performed on a respective subset of the binary masks, wherein the performing of the shape detection operations identifies a shape in the image;
determining a location of interest based on the shape;
determining a localized window within the image around the location of interest; and
performing classification on the localized window within the image to determine whether an object is present within the localized window.
2. The method of claim 1 further comprising determining whether the object is a traffic sign.
3. The method of claim 1 further comprising performing morphological opening on the binary masks prior to the performing of the shape detection operations using the binary masks.
4. The method of claim 1 , wherein the performing of the shape detection operations using the binary masks includes:
performing edge detection on the image to determine a first set of edges; and
for each of the shape detection operations:
determining a second set of edges by applying the respective subset of the binary masks to the first set of edges to remove an edge that does not correspond to a color of the respective subset of the binary masks; and
determining whether a respective shape is present based on the second set of edges.
5. The method of claim 4 , wherein the performing of the shape detection operations includes, for a first operation of the shape detection operations, determining a center point and a radius of the respective shape.
6. The method of claim 5 , wherein the location of interest is based on at least one of the center point and the radius of the respective shape.
7. The method of claim 1 further comprising performing histogram equalization on the image prior to the extracting of the binary masks from the image.
8. The method of claim 1 further comprising converting a color space of the image prior to the extracting of the binary masks from the image.
9. The method of claim 1 further comprising subsampling the image prior to the extracting of the binary masks from the image.
10. The method of claim 1 , wherein the performing of the classification on the localized window includes applying a feature extraction operation to the image to determine a set of feature planes for the image.
11. The method of claim 10 , wherein the feature extraction operation includes performing adaptive bosting classification.
12. An advanced driver assistance system comprising:
a processor;
a non-transitory memory configured to store instructions that, when executed by the processor, cause the processor to:
receive an image;
extract binary masks from the image;
perform shape detection operations on the image using the binary masks such that each of the shape detection operations is performed on a respective subset of the binary masks, wherein the performing of the shape detection operations identifies a shape in the image;
determine a location of interest based on the shape;
determine a localized window within the image and around the location of interest; and
perform classification on the localized window within the image to determine whether an object is present within the localized window.
13. The advanced driver assistance system of claim 12 , wherein the instructions cause the processor to determine whether the object is a traffic sign.
14. The advanced driver assistance system of claim 12 , wherein the instructions cause the processor to perform morphological opening on the binary masks prior to the performing of the shape detection operations using the binary masks.
15. The advanced driver assistance system of claim 12 , wherein the instructions to perform the shape detection operations include instructions that cause the processor to:
perform edge detection on the image to determine a first set of edges; and
for each of the shape detection operations:
determine a second set of edges by applying the respective subset of the binary masks to the first set of edges to remove an edge that does not correspond to a color of the respective subset of the binary masks; and
determine whether a respective shape is present based on the second set of edges.
16. The advanced driver assistance system of claim 15 , wherein the instructions to perform the shape detection operations cause the processor to, for a first operation of the shape detection operations, determine a center point and a radius of the respective shape.
17. The advanced driver assistance system of claim 16 , wherein the location of interest is based on at least one of the center point and the radius of the respective shape.
18. The advanced driver assistance system of claim 12 , wherein the instructions cause the processor to perform histogram equalization on the image prior to the extracting of the binary masks from the image.
19. The advanced driver assistance system of claim 12 , wherein the instructions cause the processor to convert a color space of the image prior to the extracting of the binary masks from the image.
20. The advanced driver assistance system of claim 12 , wherein the instructions cause the processor to subsample the image prior to the extracting of the binary masks from the image.