Function Module Details

AIVT-OV The kit is in LabVIEW The function panel path for is as follows: Block Diagram>>Functions>>Addons>>VIRobotics>>AI Vision Toolkit for OpenVINO

You will see the following three main modules:

opencv_yiku
OpenVINO
ModelZoo

Here is a brief introduction to the three modules:

🔷 Module 1:opencv_yiku

Provide OpenCV Image processing functions, camera capture, video input and output, traditional visual analysis, deep learning preprocessing and other functions.

Submodule	Function Description
`Mat`	Matrix Data Type Definition and Basic Operations
`Camera`	USB, network, industrial camera capture, video reading
`core`	Some basic processing of images
`imgproc`	Image color space, histogram, thresholding, contour, edge detection, filtering, mapping, Hough detection, corner detection, etc.
`imshow`	Image display control (OpenCV window)
`Draw`	Draw rectangles, lines, text, and other image annotation operations
`imgcodes`	Read, Write and Codec of Image File
`geometry`	Graphics geometry processing, point set, rectangle, contour processing
`cuda`	for management Information and switching of GPU(CUDA) environment
`features2d`	2D Feature description, extraction and matching of image feature points, for image registration, object recognition, image stitching and other tasks.
`dnn`	Call OpenCV-DNN Module to carry out AI Inference
`ml`	Support for basic machine learning models (SVMs, etc.)
`VideoWriter`	Image frame writing to video file
`calib3d`	Camera calibration, hand-eye calibration, 3D reconstruction
`face`	face detection and recognition
`RT`	RT System Path

✅ Applicable to all scenes of pre-processing, traditional vision, visual calibration and auxiliary AI input.

🔷 Module 2:OpenVINO

Provide ONNX, IR, Paddle and other model loading and acceleration inference functions, supporting multi-device inference (CPU/iGPU/dGPU)

Submodule	Function Description
`getAvailableDevices`	Gets the natively supported OpenVINO inference Device List
`getVersion`	Display OpenVINO runtime Current Version
`OV`	Call OpenVINO The deep learning model supports loading. IR, ONNX model, and execute inference, get results, set input and output, etc.

⚠️ Recommendations for use: - Model path is recommended to use English path, no spaces or Chinese;

Broad model support: Integrated OpenVINO™The deep learning inference engine function of supports IR, onnx, and paddle models generated inference mainstream frameworks such as PyTorch, TensorFlow, ONNX, and PaddlePaddle;
Multiple hardware accelerationsLeverage the AI high-performance inference of Intel CPUs, integrated graphics and discrete graphics to provide high-performance, low-latency visual feedback for industrial and commercial applications;

🔷 Module 3:ModelZoo

Built-in common task model inference encapsulation, including YOLO, DeeplabV3/V3, SAM, PaddleOCR It is suitable for quickly building a complete process.

Submodule	Function Description
`Object_Detection`	YOLOv5 ~ YOLOv13, RT-DETR Fast inference and automatic tracking of other models
`Segmentation`	Semantic segmentation model calls (e. g. DeepLabV3/3, Unet, SAM) fast implementation
`PaddleOCR`	text detection + Text recognition integration process

🧩 Advantages: - High encapsulation: each task only needs to call 1~2 One VI can be completed; - Strong scalability: support model replacement, custom category, running parameter setting; - Supports multiple input modes of video streams, image sets, and real-time cameras, and provides input of Mat and nivision data types.

🧠 Use recommendations

Task Type	recommend module
Image preprocessing, image acquisition	✅ opencv_yiku
Load your own ONNX/IR Model	✅ OpenVINO
Quickly build complete AI System	✅ ModelZoo
Teaching demonstration, beginner's introduction	✅ ModelZoo Sample Matching

🔍 Deep reading

If you need to view the input and output parameters, type descriptions, and usage of each module function in detail, please refer to the Help document of the built-in function of the toolkit: