APPLICATIONS

AI INSPECTION

Play Video

Glass Bottle AI Inspection

 

Stains, scratches, and cracks are common cosmetic defects on glass surfaces and are difficult for rule-based vision systems like AOI to detect. Solvision requires only a few image samples to learn and identify each defect type. Users can also label multiple defects on the same image to speed up training and inspection time.

Play Video

Sewing Defect AI Inspection

 

In many inspection applications, an array of objects or defects need to be recognized simultaneously. Solvision can easily label and identify up to 30 classifications in the same image, significantly reducing the time required to train a model. In this example, different sewing irregularities such as loose thread and broken stitches are detected using AI.

Play Video

Chicken Nugget AI Inspection

 

Food grade inspection is crucial to assessing quality, freshness, and market value. With Solvision, unique features of produce or food items can be used to categorize them into different levels. All it requires is a few minutes and a small number of sample images to train the AI model.

Play Video

Automotive Part AI Inspection

 

For manufactured objects with highly consistent layouts or configurations and where defect samples are hard to obtain, Solvision offers an “unsupervised learning” mode that allows labeling and training using only good samples.

Play Video

Car Hood AI Inspection

 

Stains, scratches, and cracks are common cosmetic defects on metal surfaces and are difficult for rule-based vision systems like AOI to detect. Solvision requires only a few image samples to identify each defect type. Users will also find it handy to be able to label multiple defect types on the same image.

Play Video

Textile Bobbin AI Inspection

 

In many inspection applications, an array of objects or defects need to be recognized simultaneously. Solvision can easily label and identify up to 30 different classifications in the same image, significantly reducing the amount of time required to train a model. In this example, defects such as dirty or broken threads are detected on textile bobbins.