Outside recognition system, vehicle controller, travel path recognition method, and storage medium

1 . An outside recognition system comprising:

a front scene image acquirer that acquires an image of a scene on a front side seen from an own vehicle as a front scene image; and

a travel path recognizer that recognizes a travel path of the own vehicle based on the front scene image,

wherein the travel path recognizer includes:

a classifier that performs segmentation processing on the front scene image to classify photographic objects in the front scene image into a plurality of classes including an own vehicle division line dividing the travel path, and to generate an edge image in which edges of the classes are extracted;

a parallel edge extractor that extracts, from the edge image for the classes, an own vehicle division line edge that is an edge of the own vehicle division line and a plurality of parallel edges that are parallel to the own vehicle division line edge when a field of view of the front scene image acquirer is seen in a plan view, and selects one of the plurality of parallel edges as a reference edge;

a feature ratio parameter calculator that calculates a value of a feature ratio parameter correlated to a feature ratio between a first distance, when the edges other than the reference edge in the plurality of parallel edges are designated as comparison edges, between the reference edge and each of the comparison edges and a second distance between the reference edge and the own vehicle division line edge along a horizontal reference line extending in width directions in the edge image;

a memory that stores a value of the feature ratio parameter, the value being calculated by the feature ratio parameter calculator at a time earlier than a current time; and

an own vehicle division line estimator that estimates, based on positions of the reference edge and the comparison edges at the current time and the value of the feature ratio parameter, the value being stored in the memory, a position of the own vehicle division line at the current time.

2 . The outside recognition system according to claim 1 , wherein the parallel edge extractor extracts, as the parallel edges, edges extending toward a vanishing range centered around a vanishing point that is common in the edge image.

3 . The outside recognition system according to claim 2 , wherein

the parallel edge extractor

uses one or more section lines that is or are parallel to the horizontal reference line to partition, in a virtual manner, the edges existing in the edge image each into a plurality of edge line segments, and

extracts the edges each including the plurality of edge line segments each extending toward the vanishing range designated for each of partition divisions divided in a virtual manner by the one or more section lines in the plurality of edges included in the edge image as the parallel edges.

4 . The outside recognition system according to claim 3 , wherein

the feature ratio parameter calculator calculates a value of the feature ratio parameter for each of the partition divisions,

the memory stores the value of the feature ratio parameter, the value being calculated by the feature ratio parameter calculator, for each of the partition divisions, and

the own vehicle division line estimator estimates, based on the value of the feature ratio parameter, the value being stored in the memory for each of the partition divisions, the position of the own vehicle division line at the current time for each of the partition divisions.

5 . The outside recognition system according to claim 1 , wherein

the travel path recognizer further includes a confidence degree calculator that calculates a value of a confidence degree for each of the comparison edges based on a comparison between a position of the own vehicle division line edge extracted by the parallel edge extractor and the position of the own vehicle division line, the position being estimated by the own vehicle division line estimator,

the memory stores a value of the confidence degree, the value being calculated by the confidence degree calculator, at a time earlier than the current time, and

the own vehicle division line estimator estimates, based on the positions of the reference edge and the comparison edges at the current time and the values of the feature ratio parameter and the confidence degree, the values being stored in the memory, the position of the own vehicle division line at the current time.

6 . The outside recognition system according to claim 1 , wherein the own vehicle division line estimator estimates the position of the own vehicle division line at the current time to allow the value of the feature ratio at the current time and the value of the feature ratio, the value being calculated based on the value of the feature ratio parameter, the value being stored in the memory, to be equal to each other.

7 . The outside recognition system according to claim 2 , wherein the own vehicle division line estimator estimates the position of the own vehicle division line at the current time to allow the value of the feature ratio at the current time and the value of the feature ratio, the value being calculated based on the value of the feature ratio parameter, the value being stored in the memory, to be equal to each other.

8 . The outside recognition system according to claim 3 , wherein the own vehicle division line estimator estimates the position of the own vehicle division line at the current time to allow the value of the feature ratio at the current time and the value of the feature ratio, the value being calculated based on the value of the feature ratio parameter, the value being stored in the memory, to be equal to each other.

9 . The outside recognition system according to claim 4 , wherein the own vehicle division line estimator estimates the position of the own vehicle division line at the current time to allow the value of the feature ratio at the current time and the value of the feature ratio, the value being calculated based on the value of the feature ratio parameter, the value being stored in the memory, to be equal to each other.

10 . The outside recognition system according to claim 5 , wherein the own vehicle division line estimator estimates the position of the own vehicle division line at the current time to allow the value of the feature ratio at the current time and the value of the feature ratio, the value being calculated based on the value of the feature ratio parameter, the value being stored in the memory, to be equal to each other.

11 . The outside recognition system according to claim 1 , wherein

the own vehicle division line estimator includes:

a model estimator that uses an own vehicle division line position estimation model causing the positions of the reference edge and the comparison edges to be associated with the position of the own vehicle division line to estimate the position of the own vehicle division line at the current time; and

a model updater that updates the own vehicle division line position estimation model based on the value of the feature ratio parameter, the value being stored in the memory.

12 . The outside recognition system according to claim 2 , wherein

the own vehicle division line estimator includes:

a model estimator that uses an own vehicle division line position estimation model causing the positions of the reference edge and the comparison edges to be associated with the position of the own vehicle division line to estimate the position of the own vehicle division line at the current time; and

a model updater that updates the own vehicle division line position estimation model based on the value of the feature ratio parameter, the value being stored in the memory.

13 . The outside recognition system according to claim 3 , wherein

the own vehicle division line estimator includes:

a model estimator that uses an own vehicle division line position estimation model causing the positions of the reference edge and the comparison edges to be associated with the position of the own vehicle division line to estimate the position of the own vehicle division line at the current time; and

a model updater that updates the own vehicle division line position estimation model based on the value of the feature ratio parameter, the value being stored in the memory.

14 . The outside recognition system according to claim 4 , wherein

the own vehicle division line estimator includes:

a model estimator that uses an own vehicle division line position estimation model causing the positions of the reference edge and the comparison edges to be associated with the position of the own vehicle division line to estimate the position of the own vehicle division line at the current time; and

a model updater that updates the own vehicle division line position estimation model based on the value of the feature ratio parameter, the value being stored in the memory.

15 . The outside recognition system according to claim 5 , wherein

the own vehicle division line estimator includes:

a model estimator that uses an own vehicle division line position estimation model causing the positions of the reference edge and the comparison edges to be associated with the position of the own vehicle division line to estimate the position of the own vehicle division line at the current time; and

a model updater that updates the own vehicle division line position estimation model based on the value of the feature ratio parameter, the value being stored in the memory.

16 . A vehicle controller comprising:

the outside recognition system according to claim 1 ; and

a controller that controls at least one of a steering mechanism, a travel driving device, or a braking device in the own vehicle based on a result of the recognition by the outside recognition system,

wherein the controller controls, when there is no photographic object classified into the own vehicle division line seen in the front scene image, at least one of the steering mechanism, the travel driving device, or the braking device based on the position of the own vehicle division line, the position being estimated by the own vehicle division line estimator.

17 . A travel path recognition method of causing a computer to recognize, based on a front scene image acquired by a front scene image acquirer that images a scene on a front side seen from an own vehicle, a travel path of the own vehicle, the travel path recognition method comprising:

performing segmentation processing on the front scene image to classify photographic objects in the front scene image into a plurality of classes including an own vehicle division line dividing the travel path, and to generate an edge image in which edges of the classes are extracted;

extracting, from the edge image for the classes, an own vehicle division line edge that is an edge of the own vehicle division line and a plurality of parallel edges that are parallel to the own vehicle division line edge when a field of view of the front scene image acquirer is seen in a plan view, and selecting one of the plurality of parallel edges as a reference edge;

calculating a value of a feature ratio parameter correlated to a feature ratio between a first distance, when the edges other than the reference edge in the plurality of parallel edges are designated as comparison edges, between the reference edge and each of the comparison edges and a second distance between the reference edge and the own vehicle division line edge along a horizontal reference line extending in width directions in the edge image;

allowing a memory to store a value of the feature ratio parameter at a time earlier than a current time; and

estimating, based on positions of the reference edge and the comparison edges at the current time and the value of the feature ratio parameter, the value being stored in the memory, a position of the own vehicle division line at the current time.

18 . A non-transitory medium storing a computer program causing a computer to recognize, based on a front scene image acquired by a front scene image acquirer that images a scene on a front side seen from an own vehicle, a travel path of the own vehicle,

the computer program causing the computer to execute a process comprising:

performing segmentation processing on the front scene image to classify photographic objects in the front scene image into a plurality of classes including an own vehicle division line dividing the travel path, and to generate an edge image in which edges of the classes are extracted;

extracting, from the edge image for the classes, an own vehicle division line edge that is an edge of the own vehicle division line and a plurality of parallel edges that are parallel to the own vehicle division line edge when a field of view of the front scene image acquirer is seen in a plan view, and selecting one of the plurality of parallel edges as a reference edge;

selecting one of the plurality of parallel edges as a reference edge;

calculating a value of a feature ratio parameter correlated to a feature ratio between a first distance, when the edges other than the reference edge in the plurality of parallel edges are designated as comparison edges, between the reference edge and each of the comparison edges and a second distance between the reference edge and the own vehicle division line edge along a horizontal reference line extending in width directions in the edge image;

allowing a memory to store a value of the feature ratio parameter at a time earlier than a current time; and

estimating, based on positions of the reference edge and the comparison edges at the current time and the value of the feature ratio parameter, the value being stored in the memory, a position of the own vehicle division line at the current time.