A dataflow execution engine for automatic visual inspection of production lines

Abstract

Automatic visual inspection is used to control the quality of the assembly steps of a product. By allowing the early detection of faults in the production line, it can have a large economic impact. In this work, we propose an execution environment for supporting the development of automatic visual inspection applications and running them in the production line. The applications are described as graphs, which are executed by the engine. Debugging features for use during development and production are provided. Also, the engine provides means to trace the results, such that a fault detected after production could be traced to the cause. Another key aspect is the seemingly integration of external libraries to extend the modeling capacity of the applications. As an example, we describe the integration of the Halcon library and its use in a test case. Our approach accelerates the creation of complex algorithms for visual inspection by eliminating compilation steps and improving the debugging experience.