國立虎尾科技大學 |

Design of a Computer Vision Based System to Improve the Inspection Process for Testing Taper-Lok Hole Geometry

紀錄類型:	書目-語言資料,印刷品 : Monograph/item
正題名/作者:	Design of a Computer Vision Based System to Improve the Inspection Process for Testing Taper-Lok Hole Geometry/ Becan Lawless.
作者:	Lawless, Becan,
面頁冊數:	1 electronic resource (249 pages)
附註:	Source: Dissertations Abstracts International, Volume: 87-03, Section: B.
Contained By:	Dissertations Abstracts International87-03B.
標題:	Computer science. -
電子資源:	http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=32259446
ISBN:	9798290989365

Design of a Computer Vision Based System to Improve the Inspection Process for Testing Taper-Lok Hole Geometry
Lawless, Becan,

Design of a Computer Vision Based System to Improve the Inspection Process for Testing Taper-Lok Hole Geometry [electronic resource] /Becan Lawless. - 1 electronic resource (249 pages)

Source: Dissertations Abstracts International, Volume: 87-03, Section: B.

Tapered interference-fit fasteners (e.g. Taper-lok) are commonly used by the aerospace industry due to their high fatigue strength. Their inspection relies on a 'true' geometry gauge coated with a thin film of engineer's blue being brought firmly into contact with the tapered hole. The pattern created on the gauge is then used to identify bearing surface area and distribution. It was felt by Airbus that this reliance on inspector perception has a potential for subjectivity factors such as ownership bias. This thesis presents the research, design and build of a scientific rig to reduce Taper-lok inspection subjectivity using computer vision techniques. This process has also increased understanding of the traditional inspection. The hardware development involved creation of a bespoke lighting enclosure for even illumination, and minimal reflections. One key difficulty was the highly reflective nature of the gauges. Software was also written utilising open source computer vision libraries in order to register the correct region with each image taken as the gauge rotates. The program then stacks the isolated regions into a single 'unwrapped' image for subsequent analysis. These images were shown to have a maximum distortion below 1.6%. A questionnaire study utilising unwrapped images was conducted at Airbus establishing a baseline for visual subjectivity. For one in five images the 'pass'/'fail' verdict varied when the image was repeated after a short interval. This study also identified the presence of 'leak paths' to be the main decision driving characteristic. Leak paths present as vertical dark bands of low contact from top to bottom of the contact region, periodically repeating. A direct relationship was established between engineer's blue layer thickness and shade. Shade is not expected to vary above 40 Î¼m; an acceptable limit since a borehole with particularly strong leak paths was measured to have 20.2 Î¼m conicality. Another experiment established a direct relationship between shade and unloaded geometry for a Taper-lok inspection, supporting the use of the process as a means to access expected bearing fit. Computer vision algorithms were developed to quantify unwrapped image characteristics, the most successful of which was 'leak path amplitude'. This value was found by summing each column for the red-channel in the unwrapped RGB image and then plotting a best fit square wave through this data, thus assessing averaged difference between areas of contact and non-contact. Leak path amplitude proved effective at separating images by their borehole of origin, and correlated with the Airbus inspection decisions from the questionnaire.

English

ISBN: 9798290989365Subjects--Topical Terms:

573171
Computer science.
Subjects--Index Terms:

Tapered hole

Design of a Computer Vision Based System to Improve the Inspection Process for Testing Taper-Lok Hole Geometry
LDR:04101nam a22003853i 4500 001 1172948
005 20260622113228.5
006 m o d
007 cr|nu||||||||
008 260803s2020 miu||||||m |||||||eng d
020 $a 9798290989365
035 $a (MiAaPQD)AAI32259446
035 $a (MiAaPQD)Manchester648a0296-f8c3-4a5d-b870-8c29bf3fd31e
035 $a AAI32259446
040 $a MiAaPQD $b eng $c MiAaPQD $e rda
100 1 $a Lawless, Becan, $e author. $3 1503597
245 1 0 $a Design of a Computer Vision Based System to Improve the Inspection Process for Testing Taper-Lok Hole Geometry $c Becan Lawless. $h [electronic resource] /
264 1 $a Ann Arbor : $b ProQuest Dissertations & Theses, $c 2020
300 $a 1 electronic resource (249 pages)
336 $a text $b txt $2 rdacontent
337 $a computer $b c $2 rdamedia
338 $a online resource $b cr $2 rdacarrier
500 $a Source: Dissertations Abstracts International, Volume: 87-03, Section: B.
500 $a Advisors: Heinemann, Robert.
502 $b Ph.D. $c The University of Manchester (United Kingdom) $d 2020.
520 # $a Tapered interference-fit fasteners (e.g. Taper-lok) are commonly used by the aerospace industry due to their high fatigue strength. Their inspection relies on a 'true' geometry gauge coated with a thin film of engineer's blue being brought firmly into contact with the tapered hole. The pattern created on the gauge is then used to identify bearing surface area and distribution. It was felt by Airbus that this reliance on inspector perception has a potential for subjectivity factors such as ownership bias. This thesis presents the research, design and build of a scientific rig to reduce Taper-lok inspection subjectivity using computer vision techniques. This process has also increased understanding of the traditional inspection. The hardware development involved creation of a bespoke lighting enclosure for even illumination, and minimal reflections. One key difficulty was the highly reflective nature of the gauges. Software was also written utilising open source computer vision libraries in order to register the correct region with each image taken as the gauge rotates. The program then stacks the isolated regions into a single 'unwrapped' image for subsequent analysis. These images were shown to have a maximum distortion below 1.6%. A questionnaire study utilising unwrapped images was conducted at Airbus establishing a baseline for visual subjectivity. For one in five images the 'pass'/'fail' verdict varied when the image was repeated after a short interval. This study also identified the presence of 'leak paths' to be the main decision driving characteristic. Leak paths present as vertical dark bands of low contact from top to bottom of the contact region, periodically repeating. A direct relationship was established between engineer's blue layer thickness and shade. Shade is not expected to vary above 40 Î¼m; an acceptable limit since a borehole with particularly strong leak paths was measured to have 20.2 Î¼m conicality. Another experiment established a direct relationship between shade and unloaded geometry for a Taper-lok inspection, supporting the use of the process as a means to access expected bearing fit. Computer vision algorithms were developed to quantify unwrapped image characteristics, the most successful of which was 'leak path amplitude'. This value was found by summing each column for the red-channel in the unwrapped RGB image and then plotting a best fit square wave through this data, thus assessing averaged difference between areas of contact and non-contact. Leak path amplitude proved effective at separating images by their borehole of origin, and correlated with the Airbus inspection decisions from the questionnaire.
546 $a English
590 $a School code: 1543
650 # 4 $a Computer science. $3 573171
653 # # $a Tapered hole
653 # # $a Software
690 $a 0984
710 2 # $a The University of Manchester (United Kingdom). $3 1213791
720 1 $a Heinemann, Robert $e degree supervisor.
773 0 # $t Dissertations Abstracts International $g 87-03B.
790 $a 1543
791 $a Ph.D.
792 $a 2020
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=32259446