Pedestrian crossing alert
A method for pedestrian crossing alert, the method may include receiving, by a vehicle computerized system, pedestrian crossing indicators; obtaining sensed information regarding an environment of the vehicle; processing the sensed information, wherein the processing comprises searching for at least one of the pedestrian crossing indicators; autonomously determining, when finding the at least one of the pedestrian crossing identifiers, that the vehicle is approaching a pedestrian crossing; and generating an alert when determining that the vehicle is approaching the pedestrian crossing.
1. A method for pedestrian crossing alert, the method comprises:
receiving, by a vehicle computerized system, pedestrian crossing indicators;
obtaining sensed information regarding an environment of the vehicle;
processing the sensed information, wherein the processing comprises searching for at least one of the pedestrian crossing indicators;
autonomously determining, when finding the at least one of the pedestrian crossing identifiers, that the vehicle is approaching a pedestrian crossing; and
generating an alert when determining that the vehicle is approaching the pedestrian crossing;
wherein the pedestrian crossing indicators are generated by:
obtaining a first group of tagged videos, wherein each tagged video comprises one or more pedestrian crossing tagged video frames;
obtaining a second group of untagged videos that were obtained at multiple conditions, wherein an untagged video is a video that is without a pedestrian crossing tag;
generating signatures of at least (a) pedestrian crossing tagged video frames and (b) at least some video frames of the second group of untagged videos:
clustering the signatures to provide signature clusters, wherein the signature clusters comprise different signature clusters for different pedestrian crossing types;
selecting video segments of the second group of untagged videos, wherein each selected video segment is acquired before reaching a pedestrian crossing and in timing proximity to the reaching of the pedestrian crossing;
determining pedestrian crossing identifiers based on the selected video segments and on unselected segments of the second group of untagged video; and
searching in the second group of untagged videos, identified pedestrian crossings, wherein each identified pedestrian crossing is identified at least one of the pedestrian crossing identifiers.
2. The method according to claim 1 , comprising generating the pedestrian crossing indicators.
3. The method according to claim 2 , comprising checking whether an identified pedestrian crossing is represented in one of the signature clusters.
4. The method according to claim 3 , comprising adding, to at least one of the signature clusters, a signature of an identified pedestrian crossing that was not represented in either one of the signature clusters.
5. The method according to claim 2 , comprising adding to a signature cluster of a pedestrian crossing of a given pedestrian crossing type, a signature of an identified pedestrian crossing that is of the given pedestrian crossing type and was not included in the signature cluster of the pedestrian crossing of the given pedestrian crossing type.
6. The method according to claim 2 , wherein the conditions are selected out of image acquisition conditions and environmental conditions.
7. The method according to claim 2 , comprising generating an ADAS response when determining that the vehicle is approaching the pedestrian crossing.
8. The method according to claim 7 , wherein the ADAS response comprises recommendation aimed to a human driver.
9. The method according to claim 7 , wherein the ADAS response comprises controlling an aspect of driving a vehicle.
10. The method according to claim 1 comprising checking whether an identified pedestrian crossing is represented in one of the signature clusters.
11. The method according to claim 10 comprising adding, to at least one of the signature clusters, a signature of an identified pedestrian crossing that was not represented in either one of the signature clusters.
12. The method according to claim 1 , comprising adding to a signature cluster of a pedestrian crossing of a given pedestrian crossing type, a signature of an identified pedestrian crossing that is of the given pedestrian crossing type and was not included in the signature cluster of the pedestrian crossing of the given pedestrian crossing type.
13. The method according to claim 1 , wherein the conditions are selected out of image acquisition conditions and environmental conditions.
14. The method according to claim 1 , comprising generating an ADAS response when determining that the vehicle is approaching the pedestrian crossing.
15. The method according to claim 14 , wherein the ADAS response comprises recommendation aimed to a human driver.
16. The method according to claim 14 , wherein the ADAS response comprises controlling an aspect of driving a vehicle.
17. The method according to claim 1 , wherein a pedestrian crossing indicator identifies a pedestrian crossing that (a) appears or occurs at a first probability in a pedestrian crossing and (b) appears or occurs outside a pedestrian crossing at a second probability, the second probability is significantly lower than the first probability.
18. A non-transitory computer readable medium for pedestrian crossing alert, the non-transitory computer readable medium stores instructions for:
receiving, by a vehicle computerized system, pedestrian crossing indicators;
obtaining sensed information regarding an environment of the vehicle;
processing the sensed information, wherein the processing comprises searching for at least one of the pedestrian crossing indicators;
autonomously determining, when finding the at least one of the pedestrian crossing identifiers, that the vehicle is approaching a pedestrian crossing; and
generating an alert when determining that the vehicle is approaching the pedestrian crossing;
wherein the pedestrian crossing indicators are generated by:
obtaining a first group of tagged videos, wherein each tagged video comprises one or more pedestrian crossing tagged video frames;
obtaining a second group of untagged videos that were obtained at multiple conditions, wherein an untagged video is a video that is without a pedestrian crossing tag;
generating signatures of at least (a) pedestrian crossing tagged video frames and (b) at least some video frames of the second group of untagged videos;
clustering the signatures to provide signature clusters, wherein the signature clusters comprise different signature clusters for different pedestrian crossing types;
selecting video segments of the second group of untagged videos, wherein each selected video segment is acquired before reaching a pedestrian crossing and in timing proximity to the reaching of the pedestrian crossing;
determining pedestrian crossing identifiers based on the selected video segments and on unselected segments of the second group of untagged video; and
searching in the second group of untagged videos, for identified pedestrian crossings, wherein each identified pedestrian crossing is identified at least one of the pedestrian crossing identifiers.
19. The non-transitory computer readable medium according to claim 18 that stores instructions for generating the pedestrian crossing indicators.
20. The non-transitory computer readable medium according to claim 19 , that stores instructions for checking whether an identified pedestrian crossing is represented in one of the signature clusters.
21. The non-transitory computer readable medium according to claim 20 , that stores instructions for adding, to at least one of the signature clusters, a signature of an identified pedestrian crossing that was not represented in either one of the signature clusters.
22. The non-transitory computer readable medium according to claim 19 , that stores instructions for adding to a signature cluster of a pedestrian crossing of a given pedestrian crossing type, a signature of an identified pedestrian crossing that is of the given pedestrian crossing type and was not included in the signature cluster of the pedestrian crossing of the given pedestrian crossing type.
23. The non-transitory computer readable medium according to claim 19 , wherein the conditions are selected out of image acquisition conditions and environmental conditions.
24. The non-transitory computer readable medium according to claim 19 , that stores instructions for generating an ADAS response when determining that the vehicle is approaching the pedestrian crossing.
25. The non-transitory computer readable medium according to claim 18 , wherein a pedestrian crossing indicator identifies a pedestrian crossing that (a) appears or occurs at a first probability in a pedestrian crossing and (b) appears or occurs outside a pedestrian crossing at a second probability, the second probability is significantly lower than the first probability.
26. The non-transitory computer readable medium according to claim 18 , that stores instructions for checking whether an identified pedestrian crossing is represented in one of the signature clusters.
27. The non-transitory computer readable medium according to claim 26 , that stores instructions for adding, to at least one of the signature clusters, a signature of an identified pedestrian crossing that was not represented in either one of the signature clusters.
28. The non-transitory computer readable medium according to claim 18 , that stores instructions for adding to a signature cluster of a pedestrian crossing of a given pedestrian crossing type, a signature of an identified pedestrian crossing that is of the given pedestrian crossing type and was not included in the signature cluster of the pedestrian crossing of the given pedestrian crossing type.
29. The non-transitory computer readable medium according to claim 18 , wherein the conditions are selected out of image acquisition conditions and environmental conditions.
30. The non-transitory computer readable medium according to claim 18 , that stores instructions for generating an ADAS response when determining that the vehicle is approaching the pedestrian crossing.