Litcius/Paper detail

Development of a Module for Analyzing Milling Defects Using Computer Vision Defects Using Computer Vision

Ilya Kovalev, Nadezhda Chervonnova, Ramilya Nezhmetdinova

20212021 International Russian Automation Conference (RusAutoCon)12 citationsDOI

Abstract

The article proposes an approach to the use of flaw detection of workpieces using computer vision. The architecture of the system nondestructive testing developed is presented and its main components are described. The module for analyzing milling defects consists of an active mode based on the OpenCV library and a deep analysis mode using the Keras library. The article shows practical tests of the active mode to determine various types of rejects during milling of test workpieces. The sequence of steps for processing photographs of processed workpieces using the Sobel operator and Gaussian filters is described. The article also discusses the basic preparatory operations that must be taken into account before conducting experiments on the detection of defects using computer vision.

Topics & Concepts

Sobel operatorComputer scienceComputer visionArtificial intelligenceMode (computer interface)Machine visionImage processingOperator (biology)Engineering drawingEdge detectionEngineeringImage (mathematics)Operating systemTranscription factorBiochemistryChemistryGeneRepressorEngineering Technology and MethodologiesAdvanced machining processes and optimizationManufacturing Process and Optimization