The challenge
During cable production, the cable moves through the facility at a variable line speed. An eddy-current unit detects irregularities in the weld seam, but such a detection signal provides no usable evidence: there is no image of what exactly is wrong, and manual inspection is unsustainable on a continuous line. Without recording, the manufacturer could not reconstruct what had happened during the welding process in the event of a later complaint or malfunction.
The challenge lies in the timing. The defect moves along with the line and must appear in the image at the exact right moment. With a line speed that varies, the moment the camera needs to trigger shifts constantly. In addition, two types of defects can be distinguished: less critical deviations that only need to be logged, and critical defects where a sticker is physically applied to the cable and which the operator must be able to assess immediately.
The solution
Vision Partners built a dual-camera vision system that moves completely with the production line. One camera focuses on the weld seam and photographs the defect itself; a second camera is positioned slightly further along and captures the marker sticker placed by production. An I/O module processes up to four digital trigger signals, each linked to its own capture behavior—ranging from a single loose photo to an adjustable photo series with its own interval and quantity per camera.
The system consists of:
- Two industrial machine vision cameras with lenses: one for the weld seam, one for the sticker detection.
- Industrial Machine Vision processor (Windows 11 LTSC / Intel) that runs the application and can buffer images locally.
- I/O module that reads the line's trigger and speed signals and converts them into usable control data.
- Dynamic speed compensation: the line speed is provided via an analog signal and via TCP. Based on this, the application calculates the correct time interval and delay between the photos in real time, ensuring that the defect or sticker is always clearly captured in the image.
- Snapshot application with an operator interface where critical errors appear directly on the main screen, while less critical errors are logged in the background. Images are saved with length and batch number in the filename, searchable via a log screen with up to over 10,000 inspections.
To prevent production downtime, the system automatically falls back to local storage in the event of a lost network connection, with continuous status notifications on the screen. An audit log also records who changed which setting and when—useful for quality assurance and compliance.
We build these solutions on a strong technical foundation. Vision Partners is an MVTec HALCON integrator — the leading software for machine vision — and a proud distributor of leading machine vision brands such as LED2WORK, TPL Vision, and Hikrobot. This is how we combine proven software, top-quality cameras, and lighting with our own engineering to create reliable, custom inspection and quality control systems.
The result
The manufacturer now has real-time insight into the welding process. Every welding defect is automatically and accurately recorded, regardless of line speed, and directly linked to the corresponding cable length and batch.
(Main screen example)
Critical errors are immediately visible to the operator so that intervention can take place quickly; all other errors are logged in the background. This creates a complete photographic report of the production with which quality can be assessed afterwards and complaints or malfunctions can be accurately traced — down to the meter and the batch.
Vision Partners delivered the solution turn-key: from developing the vision solution and on-site installation, to setting up and testing communication with the higher-level system, and a test phase with fine-tuning and site acceptance testing (SAT).
“Since running Snapshot, we have a clear image of every welding defect — linked to the exact length and batch. When dealing with a complaint, we no longer have to guess; we simply pull up the image. That brings peace of mind on the shop floor and confidence to our customers.”
— Production manager at the cable manufacturer
Want to know more about automatic quality inspection in continuous production processes? Contact contact OP.
