upalx-mz/grommets-vision-artificial

Purpose

grommets-Vision-Artificial is an industrial machine-vision repository for detecting/classifying grommets and related parts, with PLC/factory simulation assets. The minimal README provides no operating instructions; architecture is inferred from source, model, camera, and automation artifacts.

Primary Entry Point

• src/main.py is the likely application orchestrator.

• src/__init__.py defines the Python package boundary.

• pyproject.toml defines the Python project/runtime dependency boundary.

Vision and Inference Layer

• lib/analisys.py likely contains image-analysis and model-inference logic.

• Pretrained TensorFlow/Keras-style model assets are separated by product/domain:

- models/grommets/grommet_model.h5

- models/regco/regco_model.h5

• Class mappings are stored beside models:

- models/grommets/labels.txt

- models/regco/labels.txt

• models/modelos.txt likely provides a model registry or selection list.

Camera Acquisition Boundary

• lib/camaras.py is the camera/device integration layer.

• It should isolate live image capture from inference, allowing src/main.py to pass frames into lib/analisys.py without embedding device-specific behavior in model code.

Industrial Control and Simulation Integration

• Factory I/O simulation is represented by assets/Factory/OP180 -OP200 .factoryio.

• PLC logic/project configuration is represented by assets/Codesys/PLC Estadia.project.

• The likely control flow is: camera frame → visual classification/decision → PLC-facing signal or protocol update → Factory I/O simulation/actuation.

• sample/modbus.py indicates Modbus is the probable PLC communication protocol and an important external integration boundary.

Execution and Validation Examples

• sample/grommet.py, sample/regco.py, and sample/regco2.py appear to exercise model-specific inference paths.

• sample/sample1.py and sample/sample2.py likely provide general camera/vision experiments.

• These scripts are secondary entry points for manual validation rather than the main runtime.

Outputs and Operational Artifacts

• out/log.txt is the repository’s persisted runtime/log output location.

• img/ is the likely image capture, inspection, or visualization artifact directory.

• The architecture depends on local camera hardware, trained .h5 model files, and optionally a Modbus-connected PLC or Factory I/O/CODESYS simulation environment.