Automatic inspection planning for optimizing the surface coverage in industrial inspection
Abstract
Optical product inspection plays an important role in today's industrial manufacturing. Therefore, design of optimized solutions corresponding to the industrial requirements are essential for efficient product quality assurance. To configure an inspection setup one requires to determine the position and orientation of the cameras and illuminations as well as the optical configurations. This problem is commonly known as inspection planning. Today's optical inspection setups are mostly being designed based on trial and error requiring a lot of engineering experience and experimental work. As the design space is high dimensional, the empirical designs typically lead to suboptimal solutions and compromise between contrary requirements. In today's industry we are missing a generic automatic method to translate the inspection requirements into optimized inspection solutions. In this report we propose an optimization framework to automatically propose optimized setup solutions, by minimizing the number of acquisitions which fulfil the inspection requirements. As an example, we consider maximizing the surface coverage for the inspection of a cylinder head in a laser triangulation setup. We characterize the design space and propose different approaches to solve the problem. We finally demonstrate the planning results which successfully cover hard to reach areas on the object.
