SMore ViMo covers more than 1,000 application scenarios in consumer electronics, automotive, new energy and pan-industrial industries, enabling more than 300 tasks such as defect detection, visual localization, OCR recognition, target classification, etc. It has provided services to the world’s top 500 companies, including the world’s largest consumer electronics manufacturer, the world’s top auto parts manufacturer, and Asia’s largest LCD panel manufacturer, bringing technological capabilities to the traditional manufacturing industry.

Wireless headphone charging coil detection

More than 20 types of difficult defect detection classes, passing the ARR test of the world’s largest smart terminal brand manufacturer in terms of market capitalization.

Project Overview

As the world’s largest smart terminal brand manufacturer in terms of market capitalization, the purpose is to introduce high technology in the production process for appearance defect detection, and improve product quality and reduce the rework rate. The project has more than 20 types of categories to be tested, plenty testing areas and types with high accuracy requirements.

Solution and application effect

According to the defect categories, the inspection area of the material is divided into seven parts, and the detection and segmentation module of SMore ViMo is used to build a multi-model solution, and the model is accelerated and optimized to reduce the processing speed to 0.2 sec/chip.

The processing speed is less than 0.2 sec/slice, 25,000+ inspections per production line per day, and labor cost is reduced by 95%.

Smartwatch Character Recognition

The first intelligent OCR project of the smart terminal brand manufacturer that has the highest global market capitalization in the world.

Project Overview

Before a smartwatch enters the sales process, the manufacturer descends to assign a unique material number to each product for product tracking and after-sales service. Since the material length of engraved numbers is usually less than 5mm, the engraving form varies, and the number recognition accuracy rate is as high as 99%, the manufacturer was slow to introduce traditional OCR technology.

Solution and Application Effect

SMore ViMo’s intelligent OCR recognition module, which can recognize characters on materials in particular environments and engraving processes.

The daily average processing amount exceed 170,000 materials to achieve fully automated identification of the production line, with the processing speed of less than 0.1 second/each. The overall recognition rate of the technology is greater than or equal to 99%, the character recognition rate is higher than 99.9%.

Wireless charger core device appearance inspection

The first project that had pass the ARR test of the client for multiple defects in appearance

Project Overview

The client of this project is a well-known wireless communication equipment provider in China. There is an urgent need for automated equipment to replace manpower in the production process of the device, due to the high requirements of quality and efficiency of appearance inspection, and the large manpower input. The inspection sub-categories required by the project include moon glue offset, crush, broken, foreign body, wrinkle, 2D code offset, 2D code missing, 2D code printing defect, etc.

Solution and Application Effect

SmartMore created intelligent appearance defect detection automation equipment based on SMore ViMo classification, segmentation, 2D code recognition and other modules. The equipment provides OK/NG material differentiation, 14 types of fine defect detection for NG materials and 6 major categories of sub-assembly tray, etc., which comprehensively improves the inspection efficiency of the production line and provides data for manufacturing process analysis.

The efficiency of achieving full inspection has been increased by 2 times, the detection rate is higher than 99% and the over-inspection rate is less than 1%. The daily average processing amount exceed 100,000 in the processing speed of less than 0.3 seconds/chip.