Sojitz Announces Surface Scanner for Pre-owned Vehicle Inspections

In Japan, the Automobile Fair Trade Council requires that repair history be disclosed in the sale of pre-owned vehicles in accordance with the Automobile Fair Competition Code. However, inspections for pre-owned vehicles generally involve visual inspections by an inspector, and there is urgent demand to address discrepancies and ensure precision, secure human resources, and increase transparency of car inspections.

The drive-through car surface scanner has a gate-like design with a width of 5m, height of 3.6m and depth of 1.2m. The scanner can be retrofitted to existing vehicle inspection lanes. The scanner was jointly developed together with PFN, which is engaged in the development and sale of AI-related solutions and products. Utilizing a camera with optical technology, the scanner can simultaneously scan the front, back, top, and sides of a vehicle. Image processing is then used for detection and monitor-based visualization of scratches, dents, rust, repaint marks, and other damage on the car's surface. The scanner is highly adept at detecting damage on white, silver, and grey vehicles, which are particularly difficult to discern for inspectors conducting visual inspections.*¹ A scan is estimated to be 30 seconds for a single car. Demonstration tests are scheduled to begin in fall 2025. By acquiring vast amounts of training data for AI learning, Sojitz aims to automate detection of damages and increase detection accuracy while also developing the technology for auto-generated inspection reports.

JU Gifu Hashima is a leading auto auction company that auctions 230,000 units annually in Japan. The scanner will be installed as part of JU Gifu Hashima's vehicle inspection lane for conducting car surface damage inspections. At the same time, JU Gifu Hashima will also introduce Sojitz's in-house developed tire tread depth gauge to measure tire wear down to 0.1 millimeter units as well as Sojitz's vehicle underbody imaging system to visualize corrosion and damage under the vehicle. Based on the results of the scanner demonstration tests to be completed in March 2026, improvements will be made to the system, AI will be used to enhance the system's functions, and adjustments will be made to auto auction workflows. Sales of the scanner service are expected to begin in spring 2026.