Advanced Sensors and Computer Vision for Automotive Exterior Inspection and detection of Paint Defects

In the automotive industry, where precision and perfection are not just goals but necessities, ensuring the highest quality in exterior parts is paramount. This is where the innovative collaboration between LandingAI and Micro-Epsilon comes into play with computer vision, revolutionizing the way automotive exterior parts with reflective surfaces are inspected.

The Challenge of Inspecting Automotive Exterior Parts

Traditional methods for inspecting automotive exterior parts are not only inefficient but also prone to errors. The intricate and reflective surfaces of these parts pose a challenge in identifying qualitative defects such as scratches, dents, and paint imperfections. Micro-Epsilon, a trusted partner of LandingAI, who has introduced reflectCONTROL sensors specifically designed to address this issue. By capturing high-resolution 2D images and 3D point clouds of these difficult surfaces, these sensors provide the finest quality images that make defects easily visible during surface inspection. As a result, users can now effortlessly develop automotive computer vision solutions for inspection and defect detection with the help of LandingAI.

The Foundation for Advanced Inspection: Advanced Sensor Technology and Auto Defect Recognition

The sensor captures directly reflective surfaces with exceptional clarity. These sensors generate detailed 2D images, forming the basis for creating 3D point clouds. Multiple layers of 2D images from two perspectives are generated: The base intensity picture to identify the brightness, the amplitude picture to detect differences in reflectivity, and a curvature picture for recognizing deviations in the surface structure.

The object can be analyzed using Micro-Epsilon’s 3DInspect software when processed into 3D point clouds, However, recognizing small defects within these large 3D point clouds is a complex task, often hindered by noise and inaccuracies, especially in peripheral areas. This is where the expertise of LandingAI comes into play with 2D images for a detailed automated analysis.

Figure a: Micro-Epsilon reflectCONTROL sensor

Enhance Auto Inspection Accuracy with AI & Computer Vision

Once the highest quality images are captured, LandingLens, LandingAI’s computer vision platform, identifies and qualifies defects such as small craters, dents, and paint inclusions with precision. LandingLens pinpoint the location and size of these defects, providing a high level of detail and accuracy. LandingLens could detect even the tiniest defects of 10 µm on the 2D images in greyscale.

Base intensity picture vs. curvature picture

Transformative Results and Future Outlook

The collaboration between LandingAI and Micro-Epsilon marks a significant leap forward in quality control within the automotive industry. By combining the advanced sensor technology of Micro-Epsilon with LandingAI’s computer vision solution, manufacturers now have access to a tool that ensures unparalleled accuracy in detecting and evaluating defects in automotive exterior parts. This not only saves time but also significantly reduces the margin for error in quality control processes.

Learn more about computer vision for defect detection and AI in the automotive industry, along with sensor technologies at Micro Epsilon, and start building your models for free today.