Edge detection in machine vision using a simple L1 norm template matching algorithm

doi:10.1016/0031-3203(90)90064-R

Pattern Recognition

Volume 23, Issue 5, 1990, Pages 411-421

https://doi.org/10.1016/0031-3203(90)90064-R Get rights and content

Abstract

Template matching based on the sum of absolute errors (the L₁ norm) is an effective means of edge detection in certain controlled imaging environments where the form of the edges to be detected is known. The algorithm employs a 1-D edge profile as the template. Edges are detected by computing the L₁ norm of an error vector obtained by subtracting an edge template from the image data. This paper evaluates the performance of the L₁ norm template matching algorithm and draws comparisons to classical correlation matching. The L₁ norm template matching algorithm has potential for integrated circuit and printed circuit board inspection, and other inspection applications where the lighting and view aspect can be controlled.

References (15)

P. Kaufmann et al.
Visual inspection using linear features
Pattern Recognition
(1984)
W.A. Perkins
Using circular symmetry and intensity profiles for computer vision inspection
B.K.P. Horn
A problem in computer vision: Orienting Silicon Integrated Circuit Chips for lead bonding
W.K. Pratt
Digital Image Processing
R.T. Chin et al.
Automated visual inspection: a survey
IEEE Trans. Pattern Anal. Machine Intell.
(1982)
W.M. Silver
Normalized correlation search in alignment, gauging, and inspection
M.D. Levine
Vision in Man and Machine

There are more references available in the full text version of this article.

Cited by (11)

An integrated information technology assisted driving system to improve mine trucks-related safety
2010, Safety Science
This paper presents the conceptual model and framework of a novel information technology assisted driving system (ADS) to help improve mine trucks-related safety. The main components of the system include a real-time Google Earth (GE) graphics, GPS, a wireless communication network, a truck driver’s fatigue detection system, and an enhanced visibility display to restore degraded images when operating under severe visibility conditions. Mine trucks are traced and proximity warning is provided using GPS, IP addressable radios for wireless data transmission, and the GE graphic interface. Driver’s fatigue levels are detected and assessed using face tracing, extraction of eye features, and fatigue recognition algorithms. Development of such integrated information technology assisted driving system has a potential to be applied in coal, metal and non-metal surface mining operations.
Computer vision system for automated cell pressure probe operation
2009, Biosystems Engineering
Citation Excerpt :
In most cases (particularly for small diameter capillary probes), the meniscus represents a small proportion of the field of view, making automatic thresholding difficult. A number of researchers have used direct greyscale template matching (Strickland et al., 1990; Hague and Tillett, 2001) to skip the image binarisation step, eliminating the need for automatic threshold determination. The technique has been shown to make the image analysis process more robust to variations in illumination and to reduce the computational burden of the algorithm.
An automated cell pressure probe was designed to allow robust measurements of cellular water relations without the limitations caused by operator skill levels (rapid response time, good hand–eye coordination, and manual dexterity). The system was based upon a real-time (30 Hz) computer vision system that employed a radial basis function, artificial neural network using direct greyscale template matching for meniscus recognition and control. The system did not require any custom-built hardware but used readily available off-the-shelf components and was designed to be highly portable, scalable, and relatively simple to set up. The system was robust in the presence of debris or dust on or in the glass capillary, and for tapered capillary probes, tolerant of variations in contrast due to changes in the velocity or travel direction of the meniscus. The system was capable of controlling the pressure within ±0.0005 MPa and could perform pressure-clamp experiments at ±0.001 MPa or greater pressure steps. The automatic meniscus system recognition could accurately detect the meniscus location with a mean error of under 6 μm and a standard deviation of less than 3.2 μm. Mean errors in automatic volume estimation ranged from 0.07 to 0.6%, depending upon capillary diameter.
An adaptive template-matching method and its application to the boundary detection of brachial artery ultrasound scans
2001, Ultrasound in Medicine and Biology
Citation Excerpt :
When an image profile possesses a deterministic pattern, a template-matching method can be used to identify such a pattern; thus, the boundaries in that pattern are obtained. The template-matching method is shown to maximize the signal-to-noise ratio (SNR) while determining the location of the signal (Jain 1992), providing a method for edge detection (Strickland and Mao 1990; Tagare 1997). Unlike a deterministic signal embedded in a noisy background, the ultrasonic intensity profile of the vessel wall pattern is not invariant because the thickness of the media layer may differ from patient to patient, and the echoes reflected from the interfaces typically vary among types of US machines.
We developed an adaptive template-matching method for vessel wall boundary detection in brachial artery ultrasound (US) scans. The computer-aided identification of lumen-intima and media-adventitia boundaries of the brachial artery US scan is a critical step for flow-mediated vasodilator response measurement, which has become an increasingly important tool in both clinical and research applications. The adaptive template we propose in this paper is a combination of two parameterized Gaussian envelopes that approximates the US pulse envelope. The algorithm for template matching is cast as a constrained nonlinear least squares problem. A best match between the template and the observed US scan line signal results in a set of optimum parameters of the adaptive template, and the location of the boundaries is directly computed from these parameters. We present the results of applying our method to synthetic and clinical data, showing that the adaptive template-matching method provides superior accuracy in detecting the media-adventitia boundaries of both the near wall and the far wall of the vessel to that of conventional edge-detection methods. (E-mail: [email protected])
Template matching of binary targets in grey-scale images: A nonparametric approach
1997, Pattern Recognition
This paper introduces a nonparametric similarity measure, based on the Kolmogorov-Smirnov (KS) statistic, to be used in template-matching problems where a target of binary characteristics is to be located in a grey-scale image. KS statistic yields the best-expected value for a binary-domain similarity measure if the threshold selection to binarize the image had been optimized to take into account the geometric constraints of the template; there is, however, no need to actually binarize the image. Some good properties of a KS-based similarity measure are exposed and compared with the corresponding properties of normalized correlation. A practical algorithm to implement a template matching procedure based on the KS statistic is shown, and its computing time is compared with normalized correlation. A KS-based similarity measure proves to be usually much faster computationally that normalized correlation. Finally, some experimental results are shown.
A real-time edge detector: Algorithm and VLSI architecture
1997, Real-Time Imaging
In this paper we present a very large scale integration (VLSI) architecture of a new edge detection algorithm, which has a very regular computational structure. The new algorithm detects weak edges and produces single-pixel localized edges. Due to its highly pipelined structure, the VLSI implementation of the algorithm outputs one edge-pixel every clock cycle. The VLSI architecture is a complete realization of the algorithm, where no degradation is introduced to the ASIC output when compared to edges produced by the algorithm. The detector is capable of processing video graphic array (VGA) sized images at 30 frames/s at a clock rate of 10 MHz in a stand-alone mode, where no additional glue logic is required. The ASIC was laid out and fabricated using Samsung/s 0.8μm double-metal CMOS process.
Image analysis and computer vision: 1990
1991, CVGIP: Image Understanding
This paper presents a bibliography of over 1600 references related to computer vision and image analysis, arranged by subject matter. The topics covered include architectures; computational techniques; feature detection, segmentation, and image analysis; matching, stereo, and time-varying imagery; shape and pattern; color and texture; and three-dimensional scene analysis. A few references are also given on related topics, such as computational geometry, computer graphics, image input/output and coding, image processing, optical processing, visual perception, neural nets, pattern recognition, and artificial intelligence.

View all citing articles on Scopus

View full text

Edge detection in machine vision using a simple L1 norm template matching algorithm

Abstract

Pattern Recognition

Digital Image Processing

Automated visual inspection: a survey

IEEE Trans. Pattern Anal. Machine Intell.

Normalized correlation search in alignment, gauging, and inspection

Vision in Man and Machine

Edge detection in machine vision using a simple L₁ norm template matching algorithm