SMore ViMo – Super Platform for Industrial Manufacturing

Introduction

SMore ViMo provides a complete closed loop from image acquisition to model deployment and upgrade to the production line. By interfacing with imaging devices to achieve image acquisition, users can label the collected data and then train the model with one click, export and deploy the model to the production line to directly inspect materials in real time. As the production line changes, SMore ViMo will process self-training and upgrade the model without involving professional algorithm personnel.

Solution

SMore ViMo is a set of industrial vision end-to-end solutions, mainly consisting of two major system platforms: model training and model running, which can provide the whole process, one-stop and management capabilities for industrial manufacturing and create a standardized delivery system for intelligent manufacturing.

ViMo model training platform trains algorithms for complex scenarios such as material tracking, defect location, workpiece counting, multi-category appearance defect detection, etc. ViMo model running platform imports the models generated by the training platform and deploys them directly to the production line, seamlessly docking with the production line equipment to achieve real-time inspection.

The deployment of SMore ViMo does not require on-site programming by algorithm personnel, but only requires simple operations such as mouse clicks to complete the training and deployment of the model, which can be applied to industrial production lines to improve production efficiency and realize the numerical upgrade of traditional manufacturing industries.

The solution can be widely used in many fields such as consumer electronics, automotive, new energy, and pan-industrial, and has already provided services for many Fortune 500 manufacturers.

Core algorithm function

SMore ViMo unites the research results of 20 years of advanced technology by the core team of SmartMore and provides 4 core vision algorithm capabilities for the industrial field, such as product diversification and fast product update cycle.

OCR

Adopt end-to-end solution based on technology precipitation, support single character and multi-character standard and recognition, which can recognize different background characters, such as steel print, laser engraving, printing, textile, etc., break the technical limitations of traditional methods, solve the complex problems of curved character recognition, low contrast character recognition, larger character recognition, etc.

Detection 

Locate and classify the targets in the detected materials, applicable to multi-target detection, small target detection, counting, etc., such as solving the scenarios of drug pill counting, 3C device detection.

Classification

Classify and judge the detected objects. Such as OK/NG secondary classification judgment of materials, detection object color, food material type, 3C defect sub-classification and other issues.

Segmentation 

Fine to pixel level detection and edge recognition of detection objects. Such as identifying cracked areas of silicon wafers, bruised areas of bearings, etc.

Deployment

SMore ViMo offers 2 deployment methods, which users can choose flexibly according to their needs. For single production line, only need to solve the single link problem, then you can use the single machine deployment method. SMore ViMo can be adapted to multiple types of interfaces and communication protocols in industrial production scenarios, and supports docking multiple types of cameras, light sources, industrial control machines, common sensors, MES and databases, etc, opening up the data link of production line.