Moving body speed derivation method and moving body speed derivation program

A non-transitory computer-readable medium storing a moving body speed derivation program that causes at least one computer to execute a process, the process comprising:

• acquiring moving image data including a plurality of time-series images captured from an imaging device installed in a first moving body;
• recognizing a second moving body region included in the moving image data using a model to recognize a second moving body;
• extracting a road region by applying semantic segmentation to the moving image data using a model to extract the road region;
• deriving a reference point in the second moving body region in each image of the plurality of time-series images based on the road region and the second moving body region;
• deriving a position of the second moving body in an actual space based on a distance between the imaging device and a feature point in the second moving body corresponding to the reference point and an absolute position of the first moving body acquired by using a global navigation satellite system;
• and
• deriving a speed of the second moving body based on a change amount of the position of the second moving body in the actual space corresponding to the second moving body region included in the plurality of time-series images.

A moving body speed derivation apparatus comprising:

• an imaging device installed in a first moving body;
• one or more memories;
• and
• one or more processors coupled to the one or more memories and the one or more processors configured to:
• acquire moving image data including a plurality of time-series images captured from an imaging device installed in a first moving body,
• recognize a second moving body region included in the moving image data using a model to recognize a second moving body,
• extract a road region by applying semantic segmentation to the moving image data using a model to extract the road region,
• derive a reference point in the second moving body region in each image of the plurality of time-series images based on the road region and the second moving body region,
• derive a position of the second moving body in an actual space based on a distance between the imaging device and a feature point in the second moving body corresponding to the reference point and an absolute position of the first moving body acquired by using a global navigation satellite system, and
• derive a speed of the second moving body based on a change amount of the position of the second moving body in the actual space corresponding to the second moving body region included in the plurality of time-series images.