Overcoming machine vision challenges

MV systems can often have issues with the precision that characterises picking applications. Higher precision provides more accurate object detection of faults when installing an MV system, so engineers will want to look for vision systems with high dynamic range, high resolution and precision. This can be an issue when using prebuilt rigs because the specifications of the camera are predetermined and accompanied by other hardware that may not be the best fit for a particular application. 

Using a vision software that can be integrated with a wide range of cameras and imaging technologies will give engineers more flexibility when incorporating MV. 

MV software is critical to find the best solution for an application. Using familiar programming languages and configuration tools facilitates the configuration of vision systems and programming of image processing. This makes MV systems easier to manage while reducing latency and improving synchronisation with other systems. 

Another common challenge with MV is scalability. In our experience, uptake of MV systems would be greater if they could be scalable. If engineers are looking to use MV on a larger scale, they should consider systems that can be used with their existing skills and software tools. Removing this barrier would allow engineers to scale up their MV capacities – you would only need to have MV systems that work in harmony with other processes. 

Beckhoff’s TwinCAT vision systems, for example, can help achieve this scalability and reduce latency and synchronisation using the programming languages and configuration tools of the TwinCAT 3 open automation platform. TwinCAT vision systems allow image processing to be programmed and executed which helps reduce latency and improves synchronisation with other systems by bringing MV into the same control environment as motion applications, robotics and programmable logic controllers (PLCs).

For obvious reasons, if a product with nuts is labelled incorrectly, problems might arise that endanger consumers and affect the company's reputation. However, normal variations in products in the food industry are unpredictable, making it hard to create criteria for a vision system to detect defects. 

However, there are areas of the food industry for which MV is usable – including packaging and safety applications that assess cross-contamination. Because inspection challenges in the food production industry are different to those of other manufacturers, the best way to address these challenges is to find a solution that is specifically designed for the industry.

A general visual inspection solution will not address the needs of the food industry, which is why Beckhoff is focused on developing vision software that can be applied to most hardware, making it easier for manufacturers to tailor the hardware to their needs.