Methods for Edge Detection on CCD or CMOS Linear Sensor Arrays enabling Subpixel Resolution

Christian Mentin, Eugen Brenner

Research output: Contribution to conferenceAbstractResearchpeer-review

Abstract

Optical high-precision measuring devices are used where fast and contactless measurements of object surfaces and edges with sub-micrometer resolution is needed. Wafer inspection in the semiconductor industry is an important application. The proposed measurement setup relies on exposure of an object’s edge onto a light sensitive detector with a point light source. The X/Y position of the edge, or geometrical object parameters like thickness or diameter could be determined from the position of sequentially projected edges. When using narrowband point light sources for exposing an object, a diffraction pattern is generated, which can be seen by a photosensitive detector. Thresholding is commonly used for determining the projected edge position on the sensor. Since its lack of performance, additional information contained in the diffraction pattern can be used by other methods to achieve a much higher resolution than the pixel size of the detector is. Other Algorithms outperform thresholding quite well. A prediction error for the theoretical projected edge position on the sensor of less than 200nm could be achieved, still when using a modelled image sensor pixel size of 3.8µm. To compare the different algorithms, and additionally, take the influence of variables such as measurement noise, sensor pixel size, and the pixel value ADC resolution into account, a simulation framework was set up.

Conference

Conference2nd International Conference on Functional Integrated nano Systems
CountryAustria
CityGraz
Period27/06/1629/06/16

Fingerprint

Sensor arrays
Edge detection
Charge coupled devices
Pixels
Detectors
Diffraction patterns
Light sources
Sensors
Image sensors
Inspection
Semiconductor materials
Industry

Keywords

  • Algorithm design and analysis; High-resolution imaging; Image sensors; Image edge detection; Optical diffraction

ASJC Scopus subject areas

  • Electrical and Electronic Engineering

Cite this

Mentin, C., & Brenner, E. (2016). Methods for Edge Detection on CCD or CMOS Linear Sensor Arrays enabling Subpixel Resolution. Abstract from 2nd International Conference on Functional Integrated nano Systems, Graz, Austria.

Methods for Edge Detection on CCD or CMOS Linear Sensor Arrays enabling Subpixel Resolution. / Mentin, Christian; Brenner, Eugen.

2016. Abstract from 2nd International Conference on Functional Integrated nano Systems, Graz, Austria.

Research output: Contribution to conferenceAbstractResearchpeer-review

Mentin, C & Brenner, E 2016, 'Methods for Edge Detection on CCD or CMOS Linear Sensor Arrays enabling Subpixel Resolution' 2nd International Conference on Functional Integrated nano Systems, Graz, Austria, 27/06/16 - 29/06/16, .
Mentin C, Brenner E. Methods for Edge Detection on CCD or CMOS Linear Sensor Arrays enabling Subpixel Resolution. 2016. Abstract from 2nd International Conference on Functional Integrated nano Systems, Graz, Austria.
Mentin, Christian ; Brenner, Eugen. / Methods for Edge Detection on CCD or CMOS Linear Sensor Arrays enabling Subpixel Resolution. Abstract from 2nd International Conference on Functional Integrated nano Systems, Graz, Austria.
@conference{8792ee0eae5f48aaa4d4b6987bde0f20,
title = "Methods for Edge Detection on CCD or CMOS Linear Sensor Arrays enabling Subpixel Resolution",
abstract = "Optical high-precision measuring devices are used where fast and contactless measurements of object surfaces and edges with sub-micrometer resolution is needed. Wafer inspection in the semiconductor industry is an important application. The proposed measurement setup relies on exposure of an object’s edge onto a light sensitive detector with a point light source. The X/Y position of the edge, or geometrical object parameters like thickness or diameter could be determined from the position of sequentially projected edges. When using narrowband point light sources for exposing an object, a diffraction pattern is generated, which can be seen by a photosensitive detector. Thresholding is commonly used for determining the projected edge position on the sensor. Since its lack of performance, additional information contained in the diffraction pattern can be used by other methods to achieve a much higher resolution than the pixel size of the detector is. Other Algorithms outperform thresholding quite well. A prediction error for the theoretical projected edge position on the sensor of less than 200nm could be achieved, still when using a modelled image sensor pixel size of 3.8µm. To compare the different algorithms, and additionally, take the influence of variables such as measurement noise, sensor pixel size, and the pixel value ADC resolution into account, a simulation framework was set up.",
keywords = "Algorithm design and analysis; High-resolution imaging; Image sensors; Image edge detection; Optical diffraction",
author = "Christian Mentin and Eugen Brenner",
year = "2016",
month = "6",
day = "28",
language = "English",
note = "2nd International Conference on Functional Integrated nano Systems ; Conference date: 27-06-2016 Through 29-06-2016",

}

TY - CONF

T1 - Methods for Edge Detection on CCD or CMOS Linear Sensor Arrays enabling Subpixel Resolution

AU - Mentin,Christian

AU - Brenner,Eugen

PY - 2016/6/28

Y1 - 2016/6/28

N2 - Optical high-precision measuring devices are used where fast and contactless measurements of object surfaces and edges with sub-micrometer resolution is needed. Wafer inspection in the semiconductor industry is an important application. The proposed measurement setup relies on exposure of an object’s edge onto a light sensitive detector with a point light source. The X/Y position of the edge, or geometrical object parameters like thickness or diameter could be determined from the position of sequentially projected edges. When using narrowband point light sources for exposing an object, a diffraction pattern is generated, which can be seen by a photosensitive detector. Thresholding is commonly used for determining the projected edge position on the sensor. Since its lack of performance, additional information contained in the diffraction pattern can be used by other methods to achieve a much higher resolution than the pixel size of the detector is. Other Algorithms outperform thresholding quite well. A prediction error for the theoretical projected edge position on the sensor of less than 200nm could be achieved, still when using a modelled image sensor pixel size of 3.8µm. To compare the different algorithms, and additionally, take the influence of variables such as measurement noise, sensor pixel size, and the pixel value ADC resolution into account, a simulation framework was set up.

AB - Optical high-precision measuring devices are used where fast and contactless measurements of object surfaces and edges with sub-micrometer resolution is needed. Wafer inspection in the semiconductor industry is an important application. The proposed measurement setup relies on exposure of an object’s edge onto a light sensitive detector with a point light source. The X/Y position of the edge, or geometrical object parameters like thickness or diameter could be determined from the position of sequentially projected edges. When using narrowband point light sources for exposing an object, a diffraction pattern is generated, which can be seen by a photosensitive detector. Thresholding is commonly used for determining the projected edge position on the sensor. Since its lack of performance, additional information contained in the diffraction pattern can be used by other methods to achieve a much higher resolution than the pixel size of the detector is. Other Algorithms outperform thresholding quite well. A prediction error for the theoretical projected edge position on the sensor of less than 200nm could be achieved, still when using a modelled image sensor pixel size of 3.8µm. To compare the different algorithms, and additionally, take the influence of variables such as measurement noise, sensor pixel size, and the pixel value ADC resolution into account, a simulation framework was set up.

KW - Algorithm design and analysis; High-resolution imaging; Image sensors; Image edge detection; Optical diffraction

M3 - Abstract

ER -