Learned Manufacturing Inspection Inferences from Image Recognition Capabilities

Many complex electromechanical assemblies that are essential to some vital function of certain products can be time consuming to inspect to a sufficient level of certainty. Examples include subsystems of machine tools, robots, aircraft, and automobiles. A model-based definition with manufacturing tolerance specifications can address any design-related severe failure modes. However, out of tolerance conditions can occur due to either random common cause variability or undetected nonstandard results, such as foreign object debris. Application of various image recognition techniques potentially can save time by some automation of inspections. However, some of the most meaningful 3D recognitions may not be sufficiently reliable, and it can be an extensive process to train the recognition of all possible anomalies comprehensively enough for inspection certainty. This paper introduces a schema and method that can learn the likelihood that a specific autonomously inspected feature will be within tolerance specifications. These learned manufacturing inspection inferences from image recognition capabilities (LeMIIIRC) may be performed by accepting data inputs that can be obtained during the image recognition training process followed by machine learning of the likely results. The fundamental method is demonstrated by a realistic example with hypothetical manufacturing data.