DeGould’s two main products are the Auto-scan and Auto-compact. Both systems use ultra-high-resolution images captured in a controlled lighting environment. AI algorithms then process the images in the cloud and identify defects like scratches, dents, paint chips and wheel scuffs. Finally, the dashboard supplies a complete and accurate vehicle condition record in less than a minute. The Auto-compact system uses a combination of structured, dark and bright field lighting, as well as a range of cameras, to capture hundreds of ultra-high-definition images in seconds. As well as damage, the AI also checks the vehicle specification. Images are held in the cloud for easy access and storage, customers can review the data on the dashboard. The Auto-compact system can be used as a drive-through system or placed over a conveyor.
Full-frame sensor-based cameras
The image quality is only as good as the amount of light that gets through to the image sensors in the cameras. If DeGould used only machine vision-based technology with smaller sensors and lenses, they wouldn’t be able to identify the tiniest defects. By using Sony’s Alpha 7RIV and 9II cameras with an exclusive 61MP full-frame image sensor housed inside the Alpha 7RIV, the system can detect and analyse more surface defects. To maximise the amount of light going to the image sensor the Sony G Master lens is used; this increases the pixel count, which is vital for detecting minute defects on a vehicle’s surface. The full-frame sensor-based cameras provide a different image compared to a traditional industrial-based RAW image. The larger surface area of the sensor and back-illuminated structure of the Exmor R and Exmor RS image sensors allows for the maximum amount of light to reach the pixels.
In the OEM environment, it may be necessary to scan a black car, a red car and a white car; in quick succession. Sony cameras are advanced enough to ensure that the ISO will float and change, retaining the high image quality. Consider a tiny paint chip on the surface of a vehicle that an OEM customer wants to find. A low-resolution camera will result in fewer pixels in the image, and therefore the chances of the AI identifying the damage decrease. With the ultra-high-resolution images provided by the Sony cameras, the system achieves more pixels and more damage detection.
The partnership with Sony allows DeGould to make full use of the latest full-frame Alpha mirrorless cameras. Furthermore, in February 2020, Sony launched an SDK that allowed for the integration of their full-frame Alpha cameras into highly industrialised systems like the DeGould Auto-compact. Previously, users had to manually take the picture, whereas now it can all be done electronically, remotely, and controlled from a software GUI. The SDK helps DeGould software grab the images from the camera and transfer them to a separate folder. The proprietary AI algorithms then process these images in the cloud to identify defects on the surface of vehicles. Although the two processes of image acquisition and applying AI are working independently, the AI can only be so precise because of image acquisition. Therefore the AI performs at an optimum level, due to the powerful high-resolution images from the Alpha 7RIV.
*The article is based on an article by Sony