Texture analysis using computer vision

• Published in: Computers in industry Vol. 16; no. 1; pp. 25 - 34
• Main Authors: Damodarasamy, Sampath, Raman, Shivakumar
• Format: Journal Article
• Language: English
• Published: Amsterdam Elsevier B.V 01.04.1991; Elsevier; Elsevier Sequoia S.A
• Subjects: Applied sciences; Computer vision; Exact sciences and technology; Experiments; Industrial engineering; Industrial metrology. Testing; Inspections; Measurement; Mechanical engineering. Machine design; Methods; Personal computers; Surface inspection; Surface metrology; Surface roughness; Vision systems; Computer vision; Surface roughness; Surface inspection; Surface metrology; Check; Metrology; Surface conditions
• ISSN: 0166-3615; 1872-6194
• DOI: 10.1016/0166-3615(91)90005-T

Surfaces of industrial parts need to be specified based on their utility and application environment. Surface characterization is hence very vital for design, manufacturing and inspection. Current techniques of surface measurement use surface profilometers, coordinate measuring machines and some optical techniques to estimate the nature of the surfaces. With the advent of automation surface characterization needs to be totally computerized so that the task of inspection (of surfaces) is greatly simplified. In the present paper a methodology is presented that uses a computer vision system to characterize the nature of the surface. In doing so, optical principles and lighting are discussed. Simple experiments were conducted using precalibrated gauge blocks to verify the feasibility of the method. A simple discussion is then presented with details on how to use this system for inspection. The advantages of using a vision system over other techniques is also adequately discussed.