Machine Vision for Wood Inspection

98% accuracy in wood inspection

JLI vision specializes in developing advanced machine vision systems for automating challenging inspection tasks, the type of inspection that would normally require a highly trained human eye.

By combining 2D and 3D machine vision with machine learning, we can automate the inspection of items that would otherwise require the assessment of experienced employees.

For example, when using wooden boards for quality furniture, it's very important to detect the knots that could cause holes or dropouts, before the board is processed in expensive operations. Some resin pockets can cause problems, as the resin will get more fluent, if the furniture is placed in warm surroundings. It is very difficult to distinguish between acceptable resin pockets and rejectable resin pockets.

To solve this task, JLI vision has combined machine vision and machine learning. We have trained a neural network with 30,000 images to determine which knots and resin pockets can pass a quality control

Another example is inspection of lacquered surfaces of wooden elements. With a combination of 2D machine vision, multiple cameras, and machine learning, we detect a wide range of different types of defects, e.g. holes in the surface, edge and corner damage, dust in the lacquer, glossy stripes, orange peel, etc.

This built-up knowledge of inspection in the wood industry is the foundation of the new solutions we make. We have documented success with 98% accuracy in detecting defects in challenging inspection tasks.