Surface Inspection Integration

This software package acts as a client to an industrial robot controller, integrating a GigE camera and machine learning algorithm for automatic defect detection and reporting.

The software architecture is in its current form in this version of the code, but will still require testing and (likely) bugfixes.

Quickstart

Clone this repository to your machine
Download the trained network from the following link: https://filelocker.uconn.edu/public_download?shareId=0d857639f812a7baf4d3838015a8f4fa . When prompted for a password, enter resnet
Copy RES101TrainedNET.mat into the Data folder
Open MATLAB to the directory you cloned this repository to
Configure connections to the IRC and Camera, as described below.
Clear your MATLAB workspace: clear all;
Start the inspection by instantiating the AutoOpticalInspection class:

aoi = AutoOpticalInspection();

The IRC will be automatically connected to, and the system will respond to packets from the IRC. 8. Important: After testing, run aoi.dispose(); to close the TCP connection to the IRC, the connection to the camera, and shut down the state machine. If this is not done, connections may persist in the background and cause errors when trying to restart the inspection.

Configuring IRC Connection

At the top of the /Modules/Mod_ProcessControl.m file, adjust the IRC IP address and Port number to match the actual IRC.

Configuring the Camera Connection

In /Modules/Mod_ImageAcquisition.m modify the IP address for the camera. The default exposure (shutter speed) can be adjusted here as well.

dscl/Surface_Inspection_Integration

About

Resources

Stars

Watchers

Forks

Releases

Contributors 2

Languages