Map data generation device

A map data generation device comprising:

• one or more processors configured to:
• acquire input information in which an image in which an outside is captured from a mobile object and position information of the mobile object are associated,
• acquire visual saliency distribution information obtained by estimating a level of visual saliency in the image based on the image,
• set a reference sight line position in the image according to a predetermined rule, the reference sight line position corresponding to a sight line along a traveling direction of the mobile object in a traffic environment in which there is no obstacle or other mobile object,
• calculate a visual attention concentration degree with respect to the reference sight line position based on a visual saliency of each position constituting the visual saliency distribution information,
• determine, based on the visual attention concentration degree, whether or not a point or a section has a high risk, and
• add, upon a result of the determination that the point or the section has the high risk, attention information with the point or the section as a point or a section which requires visual attention, to map data.

A map data generation method executed by a map data generation device, the map data generation device adding predetermined information to map data based on an image in which an outside is captured from a mobile object, the map data generation method comprising:

• acquiring input information in which the image and the position information of the mobile object are associated;
• acquiring visual saliency distribution information obtained by estimating a level of visual saliency in the image based on the image;
• calculating a movement amount of an estimated gazing point based on the visual saliency distribution information by estimating that the estimated gazing point is a position on the image in which the visual saliency becomes the maximum in the visual saliency distribution information, the movement amount of the calculated estimated gazing point being based on a time sequence;
• determining, based on the movement amount of the calculated estimated gazing point, whether or not a visual recognition load tendency is high at a point or a section;
• and
• adding, upon a result of the determination that the visual recognition load tendency is high at the point or the section, attention information with the point or the section as a point or a section which requires visual attention, to map data.

A map data generation method executed by a map data generation device, the map data generation device adding predetermined information to map data based on an image in which an outside is captured from a mobile object, the map data generation method comprising:

• acquiring input information in which the image and the position information of the mobile object are associated;
• acquiring visual saliency distribution information obtained by estimating a level of visual saliency in the image based on the image;
• setting a reference sight line position in the image according to a predetermined rule, the reference sight line position corresponding to a sight line along a traveling direction of the mobile object in a traffic environment in which there is no obstacle or other mobile object;
• calculating a visual attention concentration degree with respect to the reference sight line position based on a visual saliency of each position constituting the visual saliency distribution information;
• determining, based on the visual attention concentration degree, whether or not a point or a section has a high risk;
• and
• adding, upon a result of the determining that the point or the section has a high risk, attention information with the point or the section as a point or a section which requires visual attention, to map data.