Method and system for adaptive ray based scene analysis of semantic traffic spaces and vehicle equipped with such system
A method and a system computationally performs scene analysis of semantic traffic spaces based on an adaptive spatio-temporal ray-based approach. The method includes acquiring a spatial semantic environment map including semantic context data, calculating at least one feature for at least one location on the spatial environment map taking into account the semantic context of the location and determining a category for the at least one location based on the at least one calculated feature. A feature is a ray based feature calculated by integrating input values along at least one ray extending in least one of a space dimension and a time dimension. The ray may have a ray shape corresponding to a general shape of a road and/or the at least one spatially extending ray follows a course of the road and/or is perpendicular to a surrounding road segment orientation.
1. A method for spatial analysis of a traffic scene, comprising:
acquiring a spatial semantic environment map including semantic context data, wherein the semantic context data are confidence values or binary values for a semantic label for grid cells of the spatial semantic environment map,
calculating at least one feature for at least one cell on the spatial environment map taking into account the semantic context of the cell,
wherein the at least one feature is a ray based feature calculated by integrating confidence values or binary values to provide an integrated semantic label along at least one ray extending from the at least one cell in at least one of a space dimension and a time dimension,
determining a category for the at least one cell based of the at least one calculated feature, wherein the at least one ray has a ray shape which is determined based on the semantic context of the at least one cell where it starts, and
generating and outputting an analysis signal comprising data about the determined category of the at least one cell for predicting a traffic scene.
2. The method according to claim 1 , wherein
the at least one ray is a spatially extending ray,
wherein the spatially extending ray has a ray shape corresponding to a general shape of a road or the at least one spatially extending ray follows a course of the road and/or is perpendicular to a surrounding road segment orientation or bypasses at least one area on the spatial semantic environment map.
3. The method according to claim 1 , wherein
the at least one feature corresponds to a distance at which integrated input values integrated along the at least one ray exceed at least one predetermined threshold.
4. The method according to claim 1 , further comprising:
performing the method for spatial evaluation of a traffic scene separately for at least one further cell of the spatial semantic environment map.
5. The method according to claim 1 , further comprising:
applying a label each to the at least one cell of the spatial semantic environment map,
wherein the label is determined based on the determined category for the at least one cell.
6. The method according to claim 1 , further comprising:
determining the category for each ray based feature of a set of ray based features or on a selected number of ray based features of a set of ray based features.
7. The method according to claim 1 , further comprising:
determining a general scene category for the traffic scene based on the spatial analysis of the traffic scene.
8. The method according to claim 1 , wherein
the spatial semantic environment map comprises areas labelled with an area type, and/or
the spatial semantic environment map comprises multiple layers, wherein the layers refer to different points in time in a time dimension.
9. A computer program embodied on a non-transitory computer-readable medium, said program comprising program-code for executing the steps according to claim 1 , when the program is executed on a computer or digital signal processor.
10. A system for spatial analysis of a traffic scene, the system comprising:
an acquisition means configured to acquire a spatial semantic environment map including semantic context data, wherein the semantic context data are confidence values or binary values for a semantic label for grid cells of the spatial semantic environment map,
a spatial ray feature extraction means configured to calculate at least one feature for at least one cell of the spatial environment map taking into account the semantic context of the cell,
wherein the spatial ray feature extraction means is configured to calculate the at least one feature as a ray based feature by integrating confidence values or binary values along at least one ray extending in at least one of a space dimension and a time dimension,
a ray shape determination means configured to determine the at least one ray having a ray shape based on the semantic context of the at least one cell,
a classification means configured to determine a category for the at least one cell based on the at least one calculated feature, and
an output means configured to generate and output an analysis signal comprising data about the determined category of the at least one cell for predicting a traffic scene.
11. A vehicle including the system for the spatial analysis of a traffic scene according to claim 10 , wherein
the driver assistance system is configured to influence control of the ego-vehicle based on output signals of the driver assistance system.