@Brzhou wrote:
Hi Sir,
Just a open discussion, labview + opencv solution is much easier than that of ni vision ?
Because we have AI now, can use AI for opencv easily.
Thanks.
Bruce
In short words — no, a LabVIEW + OpenCV solution is generally more complicated than using NI Vision.
If you plan to develop everything from the scratch, you will need to create a wrapper, because OpenCV is a C++ library. (I would not recommend using Python as the wrapper layer.) Alternatively, you can use something “ready to use,” for example the OpenLvVision_OpenCv Toolkit, but then you are limited by the functionality that toolkit provides and you become dependent on it. A slightly easier option is the Intel IPP library, because all functions are exported in a C‑compatible form, you will avoid wrapper in many cases, because you can call them directly from LabVIEW. However, you still need to understand DLLs, memory layouts and how image data is stored. So, my recommendation is to stay with NI Vision and extend it with external libraries only when really necessary. Of course, check the toolkit mentioned above — it may already be sufficient for your needs.
It’s also unclear what you mean by “can use AI for OpenCV easily.” If by “AI use” you mean using AI tools (Copilot, Claude, Cursor, etc.) to generate code and speed up development, then you still need a C/C++ wrapper around large parts of the vision code to take full advantage of AI‑assisted development. But if you mean Deep Learning (OpenCV’s DNN module, available since OpenCV 3.1), then you again face the same requirement: you must create a wrapper to expose these functions to LabVIEW.
If you describe the problem you are trying to solve, we can probably give more specific help.
