6 as well. In automobile industry, mostly every parts of the car are assembled directly using
industrial robot as human need to be the second eye for the robot. However, the inspection of windscreen car gluing process still using human eye to classified the gluing quality. Not
only that, the correction for defects done manually by human labor in ensuring the product can be reproduced as a quality product. This will effect to the quality of product as well as
time consuming in achieving target of each product. Human also has a limitation on focusing a specific task for a long time, so the better solution is applying automation in
replacing human to provide more productive and effective system concurrent to the objective of the third Industrial Master Plan IMP3. By implementing proper system for
gluing process, helps in improving quality of the car in terms of aerodynamic. This is because there is no gap between car and windscreen car in preventing drag, wind noise,
noise emission and undesired lift forces that causes of aerodynamic instability.
1.3 Literature Review
In this section, the literatures are differentiated from two points of view. One is on the vision system and another is on the industrial applications, which focus mainly on
installing car windscreen in automation industry.
1.3.1 Machine Vision
First and foremost, the first literature review was entitled as “Intelligent Shape Recognition for Comple
x Industrial Task” Yang and Sengupta, 1988 states that applications of current machine vision system in industry demand highly constrained
environment and also limit the number of objects involved. This paper proposes an intelligent shape representation and recognition system that can handle a large class of
7 objects under less constraint situations. They investigated on intelligent integration of
different shape representation schemes and generation of the best shape recognition strategy using global shape properties. The proposed scheme effectively incorporates
model-driven top-down and bottom-up approaches of shape analysis. By analyzing global shape properties, the essential features and their degrees of importance are determined
quickly. In the representation phase, objects are described by using these essential features, in recognition phase, searching for the best candidate is restricted to the models
represented by these features and the observed shape is matched to the candidate’s model in the order of importance of the essential features. Systems are being developed for two
and three dimensional shape separately since they are exploiting different visual data photometric and range, respectively.
This two literature reviews gives a lot of idea in implementation Region of Interest in this res
earch. They entitled as “Flexible approach for Region of Interest Creation for Shape-
Based Matching in Vision System” by Teck et al., 2009 and “Implementation of Shape-
Based Matching Vision System in Flexible Manufacturing System” by Teck et al., 2010 that helps a lot in this research. The creation of Flexible Region of Interest ROI
provides better application that extracts the main features of the image that using as Model Template. Most important thing, the data extraction through this creates a unique data for
each model defects that easily to identify defects in matching application. The better creation of ROI gives better processing time in this system.
Besides that, the other literature review that being revised in this project was entitled as “HALCON Application for Shape-Based Matching” Xu et al., 2008a that
discussed mostly on how to develop an algorithm in determining the 3D object recognition. Shape-based matching provides an application to fulfil its main task in finding the object
based on single model image and locate them with sub pixel accuracy. In this algorithm,
8 template matching act as a tool to provide a powerful solution in recognizing object from
its model. First, before matching process takes place, the object must be specified and created object as a model in training phase. Then, the model being used in order to find and
localized an object including optimizing the outcome of the matching by minimizing the search space of the object. This unique algorithm helps in handling the changes in
illumination, clutter, various in size of object, and also affect in position and rotation of the object. Not only that, this algorithm may affect the relative movement of parts of the
template. HALCON, machine vision software consists of a huge data collection of library
that can be used for developing many different tasks. This software can easily construct the tasks that are required by using its image library operators. HALCON can be used to
construct most reliable application known as shape based matching application. This application provides superior sub pixels-accurate matching technology finds objects
robustly and accurately in real- time. It’s also able to find objects even if they are rotated,
scaled, perspectively distorted, locally deformed and partially occluded. It can process images with 8 bits or 16 bits and also handles colour or multichannel images. Figure 1
shows the process of the shape based matching used in HALCON system.
9 1.
Acquire Image – Base images that will be used as the reference image is obtain. 2.
Pre-processing – Pre-processing is needed to modify the image so that the further process can be simplified for acquiring a better solution in time consuming.
Common pre-processing includes image filter and also image enhancement. 3.
Determine Training Parameters – The essential parameters required are identified and created as a unique parameter for identifying certain image.
4. Train Model – The model are created based on the parameter develop in step 3, the
speed of the process can be improved by creating Region of Interest ROI and also using image pyramid level. ROI create a region where as only focus on the main of
interest needs to create an image as a model. 5.
Apply Matching – The Test image is the applied for shape matching process. 6.
Results – The results can be shown in sentences or images according to requirement.
Figure 1.3: Basic Process of Shape-Based Matching using HALCON Xu et al., 2008a
10 7.
Destroy Model – The test image is deleted or is overwritten after the results have been obtained. This is important in reducing memory usage.
Region is one of the basic frameworks in determining object recognition. As proposed by Gu et al., 2009
in “Recognition using Regions” stated that each region created, it represents by a rich set of image cues shape, colour and texture that accounts
as a major target in image processing for recognition object. In this paper, the objective is to provide a region database of image by introducing Generalized Hough transform in
detection and segmentation to generate hypothesis of object locations, scales and support, followed by verification classifier. This approach building on a novel region segmentation
algorithm which produces robust overlaid regions. By proposed this approach shown that by creating a specific region able to recognize object robustly but proper region creation
must be applied. As a result, there are two conclusions obtained 1 cue combination significantly boosts recognition performance; 2 region-based voting scheme reduces the
number of candidate bounding boxes by orders of magnitude over standard sliding window scheme due to robust estimation of object scales from region matching.
1.3.2 Defect Inspection System
