Tuesday, 20 October 2026Support us

Aube.

News of progress
Single source

Microintelligence Links Factory Inspection to Robot Grinding

Languages for this article
Original · ENESFRITPT

Originally written in English. 5 languages available; yours is one click away.

Generated illustration for “Microintelligence Links Factory Inspection to Robot Grinding”
AI-generated illustration — artist’s view, not documentary

On a die-cast metal part more than two metres long, the job was to find imperfections across close to 3,000 imaging positions and finish the work within ten minutes. A conventional robot following a fixed grinding path, says Zhang Zhiqi, chairman and chief executive of Microintelligence, would need about an hour. A skilled worker needed only seven or eight minutes, because the worker looked for flaws and concentrated on the places that needed attention, while the programmed machine followed its route.

Microintelligence was founded in 2018 in Changzhou and started with a narrower task: helping manufacturers check whether their products were good enough to ship. That is harder than it sounds. A defect can be small, its appearance can vary, and the surface carrying it can be difficult to photograph. Baidu, a technology partner, described an early system that used machine vision and deep learning to locate and classify defects, with cameras covering several sides of an object, and identified lighting and algorithms as its two central difficulties.

The casting project pushed the company one step further. Die casting forces molten metal into a mould, and the finished part often still needs trimming, such as removing a burr, a thin ridge of excess metal along an edge. Where that work sits differs from part to part. A 2024 report from the China Electronics Standardization Institute documents how Microintelligence's systems link the two steps: inspection software records and analyses defects, then generates positions and paths for a grinding robot. The sources note that not every defect can be repaired. What changes is that the machine's next movement depends on what the inspection found.

The company also changed shape. In December 2022 it acquired Jiangsu Zhiyun Tiangong, which its listing document describes as a turning point toward embodied-intelligence industrial robots. Zhang, who had founded Zhiyun Tiangong in 2021, became chairman in January 2023; his earlier career included roles at SAP and Baidu's intelligent cloud business.

For a factory, the change is practical. Inspection and rework have usually been separate jobs, and the second waits for the first. If a camera can decide where a robot works, the rework can follow the flaw rather than a fixed programme. The sources do not say how many customers use this, what it costs, or when it would spread beyond this kind of casting. The one figure on the table, the CEO's comparison, is his own and awaits independent testing. The company's proposed Hong Kong listing will bring more detail into the open, and that is where the next numbers should appear.

seven or eight minutesTime a skilled worker needed for one large die-cast part, per the CEO; a fixed-path robot took an hour

Sources — read the originals(Paris time)

TechNodeEN
0000

Read next

Comments

Loading the thread…

Sign in to leave a comment. Sign in