Solafune, Inc. (headquartered in Okinawa City, Okinawa Prefecture, Japan; President & CEO, Ren Uechi; hereinafter referred to as "the Company"), is pleased to announce the release of new competition to develop AI technology to detect vehicles from satellite and aerial imagery data in multi-resolutions.

• Background

People on the ground use a wide variety of means of transportation. Among them, the largest proportion is the use of cars. By understanding the movement and parking of vehicles, it is possible to analyze people's behavior and its consequences. Traditional traffic volume surveys generally involve manually counting the number of vehicles or installing cameras on roads and parking lots.

In this project, we will develop a technology to detect vehicles using images taken from the sky to the ground. The detection technology to be developed is envisioned to be applied to various situations, such as understanding traffic congestion, detecting vehicles involved in accidents, and predicting the performance of commercial facilities based on the number of vehicles parked.

• Applicability of the Technology

The technology to detect objects from satellite and aerial imagery can be applied not only to cars, but also to all kinds of objects, such as aircraft, ships, containers, and properties. In this project, we envision developing a technology that can handle differences in data properties between satellite and aerial imagery, as well as between different resolutions. By establishing a method that can handle a wide range of resolutions, we aim to achieve more efficient vehicle detection.

In addition, By supporting a wide range of resolutions, we expect not only to reduce data acquisition costs but also to cover areas where it is difficult to acquire high-resolution data. By utilizing the "aerial perspective" and "computer eyes," we will contribute to promoting various industries with digital technology.

Image: Actual vehicle detection