SMART AVI_AVI-600L_AVI-1000P_AVI-YT600-BSVISION optical technology

SMART AVI（FPC light board inspection）

BSVISION WISE AVI is an industrial image acquisition, labeling, training and analysis system based on artificial intelligence deep learning. The product landing form includes AI inspection robot (BSW-AVI) and AI vision center (BSW-AIpro). It can solve the problem that it is difficult for traditional machine vision systems to identify small and complex defects.

✔Recognition rate can reach more than 99%
✔Small sample training quickly achieves high accuracy
✔Recognition speed can be as fast as milliseconds
✔80% reduction in labor cost
✔Efficiency increased by 6-8 times

Accurate collection-industrial cameras/sensors and other testing equipment

Select the most suitable video, image, and sound data capture and acquisition equipment for a specific scene, which can include edge computing functions

Intelligent detection-AI vision central system

Integration of software and hardware, equipped with Bestway BSW-AIpro, which can locally complete deep learning, model training, identification and verification. Optional private cloud or public cloud big data platform connection

Automatic control-automation equipment and robotic arms

Cooperate with industrial cameras to recognize products from multiple angles. Cooperate with existing product lines to automatically complete loading and unloading and defect classification.

Compared with human vision

More consistent-It runs around the clock and maintains the quality level in every production line, every shift and every factory.
More reliable-Identify every defect outside the set tolerance.
Faster-Identify defects in milliseconds, support high-speed applications and improve throughput.

Compared with traditional machine vision

High precision-The latest deep learning framework carried by BSW-AIpro can ensure the high precision and accuracy of defect detection, and the software has excellent performance in processing speed.
Neural Networks-It is no longer necessary to design specific algorithms for specific image features, but to extract and analyze features through convolutional neural networks.

Compared with deep learning open source

High compatibility-The software can flexibly run on a variety of computing chips without being restricted by hardware devices. Able to realize parallel processing of heterogeneous hardware.
Customization-In terms of training models, the software can provide more options, including algorithms, model layers, etc. Support for customizing specific algorithms for projects, and more flexibly and vertically to solve the fundamental needs of users.