Utilises the OpenMV smart camera to detect surface defects on aluminium plates in real time

Once aluminium sheet surface defects enter subsequent processes, it often implies higher rework costs. Manual visual inspection is unstable and difficult to adapt to high-speed production lines. OpenMV deploys defect identification capability directly on the device end.

During the production, processing, and subsequent surface treatment of aluminium plates, surface defects such as scratches, dents, and indentations not only affect the aesthetic quality of the product but may also adversely impact subsequent assembly, structural strength, and customer acceptance. If such defects are not detected promptly at an early stage of production, they often lead to increased rework costs and, in severe cases, result in the scrapping of entire batches, causing significant losses to the enterprise.

The traditional manual visual inspection method heavily relies on the operator’s experience and concentration, and is prone to missed detections or inconsistent judgment criteria under prolonged, high-intensity production conditions, making it difficult to meet modern manufacturing’s quality requirements for stability and consistency.

Based on SingTown Technology’s OpenMV intelligent camera, continuous and stable online visual inspection can be achieved on aluminium plate production lines. The camera is installed at critical inspection positions along the production line to capture real-time images of the aluminium plate surface during high-speed motion, providing a clear and stable data foundation for defect analysis.

OpenMV runs edge-enhancement, surface texture comparison, and defect feature matching algorithms on-device, enabling fine-grained analysis of aluminium plate surfaces. By comparing normal textures with anomalous features, the system effectively magnifies subtle defects—such as scratches, pits, or indentations—that are otherwise imperceptible to the naked eye, thereby enhancing detection sensitivity and accuracy.

When defective features that do not meet quality standards are detected, OpenMV can instantly output recognition results at the edge and coordinate with sorting, rejection, or alarm devices on the production line to promptly isolate defective products, enabling forward-shifted handling of quality issues and preventing defects from proceeding to the next process.

Since the detection logic runs entirely on the camera edge device, this solution does not rely on cloud computing, effectively reducing system latency and ensuring real-time responsiveness even in high-speed production line environments. OpenMV brings industrial vision inspection capabilities to the device edge, delivering a low-latency, high-consistency, and easy-to-deploy quality control solution for manufacturing enterprises, thereby supporting continuous optimisation and stable operation of production processes.