Scribd Logo
Upload

Welcome to Scribd!

  • 9 views

    MADA11

    Uploaded by

    Himanshu Jaiswal
    This document provides an overview of an automated image processing system used to classify marble slabs on a conveyor belt based on quality. The key steps of the classification process are: 1) images of marble slabs on the conveyor are acquired using a camera inside a light-controlled cabinet; 2) features are extracted from the images; 3) a classifier categorizes the marble into different quality groups based on the features; 4) the conveyor system routes the marble accordingly based on the classification. The system standardizes image capture and uses hierarchical clustering of color, texture, and vein patterns to classify marble into four groups representing different aesthetic qualities.

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd

    MADA11

    Uploaded by

    Himanshu Jaiswal
    9 views35 pages
    This document provides an overview of an automated image processing system used to classify marble slabs on a conveyor belt based on quality. The key steps of the classification process are: 1) images of marble slabs on the conveyor are acquired using a camera inside a light-controlled cabinet; 2) features are extracted from the images; 3) a classifier categorizes the marble into different quality groups based on the features; 4) the conveyor system routes the marble accordingly based on the classification. The system standardizes image capture and uses hierarchical clustering of color, texture, and vein patterns to classify marble into four groups representing different aesthetic qualities.

    Original Title

    MADA11_167654591927146057363ee0f7f2e1ce

    Copyright

    © © All Rights Reserved

    Available Formats

    PDF, TXT or read online from Scribd

    Did you find this document useful?

    Is this content inappropriate?

    This document provides an overview of an automated image processing system used to classify marble slabs on a conveyor belt based on quality. The key steps of the classification process are: 1) images of marble slabs on the conveyor are acquired using a camera inside a light-controlled cabinet; 2) features are extracted from the images; 3) a classifier categorizes the marble into different quality groups based on the features; 4) the conveyor system routes the marble accordingly based on the classification. The system standardizes image capture and uses hierarchical clustering of color, texture, and vein patterns to classify marble into four groups representing different aesthetic qualities.

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd
    Download as pdf or txt
    9 views35 pages

    MADA11

    Uploaded by

    Himanshu Jaiswal
    This document provides an overview of an automated image processing system used to classify marble slabs on a conveyor belt based on quality. The key steps of the classification process are: 1) images of marble slabs on the conveyor are acquired using a camera inside a light-controlled cabinet; 2) features are extracted from the images; 3) a classifier categorizes the marble into different quality groups based on the features; 4) the conveyor system routes the marble accordingly based on the classification. The system standardizes image capture and uses hierarchical clustering of color, texture, and vein patterns to classify marble into four groups representing different aesthetic qualities.

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd
    Download as pdf or txt
    of 35

    Introduction to Image Processing

    Image Pre-Processing
    Once an image is acquired it is generally processed to eliminate
    errors
    Two categories:
    • Geometric correction
    • Radiometric correction
    Introduction to Image Processing

    Geometric Correction
    Sources of distortion
    • Variations in altitude
    • Variations in velocity
    • Earth curvature
    • Relief displacement
    • Atmospheric refraction
    • Skew distortion from earth’s eastward rotation
    Introduction to Image Processing

    Geometric Correction
    • Raw digital images contain two types of geometric distortions:
    systematic and random
    • Systematic sources are understood and can be corrected by
    applying formulas
    • Random distortions, or ‘residual unknown systematic distortions’
    are corrected using multiple regression of ground control points
    that are visible from the image
    Introduction to Image Processing

    Radiometric Correction
    Radiance measured at a given point is influenced by:
    • Changes in illumination
    • Atmospheric conditions (haze, clouds)
    • Angle of view
    • Instrument response characteristics
    • Elevation of the sun (seasonal change in sun angle)
    • Earth-sun distance variation
    Introduction to Image Processing

    Image enhancement
    • Improving image quality, particularly contrast
    It includes a number of methods used for enhancing subtle radiometric
    differences so that the eye can easily perceive them
    Two types: point and local operations
    • Point: modify brightness value of a given pixel independently
    • Local: modify pixel brightness based on neighborhood brightness
    values
    Introduction to Image Processing

    Image enhancement
    Three types of manipulation are:
    • Contrast enhancement: Methods include gray level thresholding,
    level slicing and contrast stretching
    • Spatial feature manipulation: Methods include spatial filtering, edge
    enhancement and Fourrier analysis
    • Multi-image manipulation: Methods include multispectral band
    ratioing and differencing, principal components, canonical
    components, vegetative components, decorrelation stretching, others
    Introduction to Image Processing

    Contrast Enhancement
    (Point Operation)

    Most images start with low contrast; these improve it


    • Level slicing re-classes DNs into fewer classes, so differences can
    be more easily seen; colours or grayscale values can be assigned.
    • Contrast Streching is the opposite, where a smaller number of
    values are stretched out over full DN range
    Introduction to Image Processing

    Contrast enhancement
    Here is what spectral histograms look like

    Note that DN is not zero for any of them


    Source: http://www.sci-ctr.edu.sg/ssc/publication/remotesense/process.htm
    Introduction to Image Processing

    Contrast enhancement

    The image on the left is hazy because of atmospheric scattering; the image is improved (right)
    through the use of Gray level thresholding. Note that, If there is more contrast and features
    can be better extracted.
    Source: http://www.sci-ctr.edu.sg/ssc/publication/remotesense/process.htm
    Introduction to Image Processing

    Spatial Feature Enhancement


    (local operation)

    • Spatial filtering/ Convolution: Neighborhood operations, that


    calculate a new value for the center pixel based on the values of its
    neighbors within a window; includes low-pass (emphasizes regional
    spatial trends, demphasizes local variability ) and high-pass
    (emphasizes local spatial variability) filters
    Introduction to Image Processing

    Spatial Feature Enhancement

    Edge Enhancement: This is a convolution method that combines


    elements of both low and high-pass filtering in a way that accentuates
    linear and local contrast features without losing the regional patterns
    • First, a high-pass image is made with local detail
    • Next, all or some of the gray level of the original scene is added
    back
    • Finally, the composite image is contrast stretched
    Introduction to Image Processing

    Image classification

    Spectral pattern recognition procedures classifies a pixel based on its


    pattern of radiance measurements in each band: more common and
    easy to use

    Spatial pattern recognition classifies a pixel based on its relationship


    to surrounding pixels: more complex and difficult to implement

    Temporal pattern recognition: looks at changes in pixels over time to


    assist in feature recognition
    Introduction to Image Processing

    Spectral Classification
    Two types of classification:
    • Supervised: The analyst designates on-screen “training areas” known
    land cover type from which an interpretation key is created, describing
    the spectral attributes of each cover class . Statistical techniques are
    then used to assign pixel data to a cover class, based on what class its
    spectral pattern resembles.
    • Unsupervised: Automated algorithms produce spectral classes based
    on natural groupings of multi-band reflectance values (rather than
    through designation of training areas), and the analyst uses references
    data, such as field measurements, DOQs or GIS data layers to assign
    areas to the given classes
    Introduction to Image Processing

    Spectral Classification
    Unsupervised:
    Spectral class 1
    • Computer groups all pixels
    according to their spectral
    relationships and looks for
    natural spectral groupings of
    Spectral class 2

    pixels, called spectral classes


    • Assumes that data in different
    cover class will not belong to
    same grouping
    • Once created, the analyst
    assesses their utility Source: F.F. Sabins, Jr., 1987, Remote Sensing: Principles and Interpretation.
    CASE STUDY:

    An automated industrial conveyor belt system using image


    processing and hierarchical clustering for classifying marble slabs
    In quality classification tasks, the classification output determines the
    category, or quality group, of a particular item. A typical classification process
    comprises five main steps:

    (i) Locating or recognizing the items on the conveyor belt via some type of a
    sensor such as a camera, scanner,etc.

    (ii) Acquiring the necessary data from the item (i.e.taking pictures,
    measuring the amount of reflected light, electromagnetic wave, or another
    type of signal). The acquisition device is usually located above the conveyor
    belt to view the items orthographically.
    (iii) Processing the data to extract several useful features.

    (iv) Classification of the item using the extracted features and a


    classifier.

    (v) Performing the necessary action following the classification


    result of the classifier.
    Marble quality classification is based on some physical, mechanical, and
    technological properties required by universal standards.

    At the same time, the classified marble slabs should reflect attractive
    colour and pattern choices. Important constraints for aesthetic
    appearance are homogeneity, texture, color, distribution of limestone
    The marble specimens used in this study are extracted from a mine in
    Manisa region of Turkey. Although there are not unique criteria for
    classifying marble specimens, colour scheme, homogeneity, size,
    orientation, thickness, and distribution of the filled joints (red–brown
    colored veins) are often used to visually perform the classification by
    human experts.
    Fig. 1. Typical marble
    Slab
    Images
    From
    Four
    Different quality groups:
    (a) Group1,
    (b) Group2,
    (c) Group3, and
    (d) Group4.
    (1) Homogenous limestone (beigecolour) (Fig. 1(a));

    (2) Limestone with filled thin joints (veins) (Fig. 1(b));

    (3) Brecciated limestone (composed of limestone grains of different


    shape and size cemented with cohesive matrix) (Fig. 1(c)).
    Here,cohesive matrix is defined as the collection of joints (veins) that
    are unified to construct a larger area of material; and

    (4) Homogenous cohesive matrix(Fig. 1(d)).


    Fig. 2. Block diagram of the electro-mechanical conveyor
    Belt system for marble classification
    Fig. 3. A look inside the closed and black painted
    Cabinet that is designed for image acquisition.
    The image acquisition part of the system is designed in order to
    standardize capturing of the marble surface images.

    The system consists of a camera (an 8 mega pixel CanonEOS350D


    digital camera with18–55 mm EF-S zoom lens or a CCD web-cam),
    connection cables, light sources, a desktop computer, and a cabinet
    housing all these parts.

    The complete image acquisition system is shown in Fig. 3.

    The 8 mega pixel camera produces high quality images with a resolution
    of 1575x1550.
    Experiments showed that using a resolution of 315x310 was enough to
    process images without affecting the success rate.

    For this purpose, a CCD web-cam was also used.

    The camera was set to have a position perpendicular to the bottom


    surface of the cabinet and the USB connection to the desktop personal
    computer was established via cables.
    The cabinet was used to ensure a fully isolated and uniformly
    illuminated area.

    The florescent light sources were positioned to prevent blazing as


    much as possible that may occur at the surface of marble samples.

    Thus, in this system, fluorescent lamps suitably positioned in the


    closed and black painted cabinet are used for illumination.
    Although the light sources inside the cabinet are located carefully so
    that illumination can be made uniform inside the image acquisition unit,
    some non-uniform illumination remained.

    To compensate for the remaining non- uniformity, image of a white


    paper is captured and then the maximum gray-level intensity value in
    the green channel of this image is calculated.

    The green channel is selected after several experiments on conversion


    of the RGB image to grayscale.
    Fig. 4. (a) Acquired image with non-uniform illumination, (b) correction of
    non-uniform illumination by adding the template image obtained from
    the white paper to the green channel of the acquired image, and (c) the
    resulting gray scale image after illumination correction and background
    cropping.
    Fig. 5. Four synthetic images used to represent cohesive material
    regions in four different quality groups: (a) in typical samples of Group1,
    cohesive material regions are very small in size, (b) in typical samples of
    Group2, cohesive material regions have vein like shapes, (c) in typical
    samples of Group3, cohesive material regions are unified, and (d) in
    typical samples of Group4, cohesive material regions are very large in
    size.
    The numbers of samples for each quality group (from 1 to 4) are
    172, 388, 411, and 187, respectively, comprising a data set of 1158
    samples in total.
    Fig. 6. Hierarchical clustering scheme.
    N is the number of quality groups and
    K is the level index. At the end of the tree,
    each quality group is constructed by
    merging its corresponding clusters at
    each level using union, U, operation.
    Fig. 7. (a) Mechanical structure of the
    system and inclined loading ramp,
    (b) pneumatic pistons, (c) PLC used
    in the system (TWDLCAE40DRF),
    (d) relay circuit, (e) 220V panel, and
    (f) air compressor for the pneumatic
    pistons.
    Fig. 8. Flow chart of the complete system.
    Fig. 9. (a) SCADA interface of the system and
    (b) MATLAB GUI of the system

    You might also like