Automated Engine Block Inspection
Confirm the presence of all necessary parts on a wide range of engines
Graphical programming environment for deep learning-based industrial image analysis
Vehicle engine assembly lines often handle several ranges and models of engines that vary in size, component types, and component arrangements. At final engine assembly inspection, proper construction needs to be verified to ensure that all necessary components are in place and properly installed before proceeding to the next manufacturing step.
These components can be in various locations, depending on the model, and be hard to identify against the engine’s complex background. As customization has become more common, there are greater variations between one engine and another. This includes significant variations in size, which can require adjusting the field of view between inspections.
Conventional machine vision is adequate for more predictable variations between engines but is hard to program as variation increases. It also takes a significant amount of time and effort to reprogram when designs change.
Cognex Deep Learning is trained on the full range of possible engine types and component arrangements. A lens with a wide field of view can result in showing engine components at differing angles, depending on how far from the center they are. The classification tool learns to identify each component, no matter the angle at which it appears.
For accurate engine assembly verification, the classification tool learns the types and required positions for various types of oil filters, wiring, hoses, and other components, and instantly flags any engine with a missing or improperly installed part, so that the error can be identified before the engine is readied for installation in a vehicle.
If a new model appears, the classification tool can be trained on images of the new engine and its components in minutes, without requiring any programming.