Visual Inspection Manual

The Visual Quality Control software library has been developed at the Georg-August-University Goettingen. The software provides a Graphical User Interface (GUI) for creating, editing and executing Computer Vision pipelines. The pipelines are called flows and should enable easier development and reading of Computer Vision algorithms. Flows are made out of blocks that represent atomic Computer Vision operations. These output of blocks have to be connected to subsequent blocks, thus creating flows.

Edit Flow ModeEdit Flow Mode

The Template mode enables on-the-fly cropping and selection of image regions, often needed for the aforementioned algorithms.

Template ModeTemplate Mode

Apart from creating and modifying flows, the GUI offers the opportunity to test flows, i.e. to apply the pipeline to a given image.

Testing ModeTesting Mode

Lastly, the monitoring mode allows for inspection of a series of images or even entire camera input streams.

Monitoring ModeMonitoring Mode

Platform, Requirements & Dependencies

Currently only Linux operating systems are supported. The system was tested using Ubuntu Linux.

  • Linux 64-bit, GCC 4.8

The library depends on the following packages:

  • ROS Kinetic
  • Boost
  • OpenCV
  • A slightly modified version of the NodeEditor Library

Installation

To build the software use the ROS function catkin_make:

$ cd path/to/ros/workspace
$ catkin_make or catkin build

Run

  • First, the information in the settings.json need to be adjusted. Namely, set the path to templates, camera name, image path for testing, mode (and if needed image path) and case for monitoring.
  • Compile the sources.
  • Start the camera for the real-time mode.
  • Start templatetester: rosrun templatetester templatetester_node.
  • Select the monitoring tab.
  • Start monitoring server: rosrun templatetester monitoring_node_server.
  • To imitate the call of the monitoring service: rosrun templatetester monitoring_client Flow TemplateID/FlowName