IT Problem Solving: Object Sorting with an Armed Robot - ITECH1101

Desklib provides past papers and solved assignments. This project details the design of an AI-powered robotic arm for object sorting.

ITECH1101
IT Problem Solving
Part 1: Design Documentation
Sorting of objects with Armed Robot
Student Name:
Student ID:

Table of Contents
Introduction................................................................................................................................2
Design Documentation...............................................................................................................3
Brief Description....................................................................................................................3
Machine Description..............................................................................................................3
Algorithm...............................................................................................................................4
Robotics Armed.....................................................................................................................9
UML Diagram......................................................................................................................10
Challenges Faced.................................................................................................................10
Feasible Solution..................................................................................................................10
Conclusion................................................................................................................................12
References................................................................................................................................13
List of Figures
Figure 1: Whole system framework...........................................................................................4
Figure 2: Feature extraction should be done before classification.............................................5
Figure 3: Original Image of the objects.....................................................................................5
Figure 4: Binary Images of the objects......................................................................................6
Figure 5: Detection of the edges................................................................................................7
Figure 6: Dilation of the images.................................................................................................7
Figure 7: Filling of the images...................................................................................................8
Figure 8: BLOBs analysis..........................................................................................................8
Figure 9: Use case Diagram.....................................................................................................10
List of Table
Table 1: Cycle no. and time is taken by the Robotic Arm.........................................................9
1

Introduction
This report is creating for the assignment based on the IT problem-solving. We have selected
the problem for the sorting of the different objects with the help of the armed robots. The
brief description of the machine is providing in the report along with the algorithm for the
sorting of the objects, working of the armed robot for placing the objects and the time taken
by the robot also going to see in the same report. We will also create the UML diagram which
shows the working of the machine for sorting of the projects. There will be also given the
challenges faced while sorting the objects and also the solution for solving the same problem.
2

Design Documentation
Brief Description
Sorting of different objects with the help of armed Robot
Control of the Robotic arm for sorting the objects in different places or according to its colour
by using vision sensors needed the processing algorithm for recognizing and detecting the
targeted objects. For the sorting of the objects, it is firstly important to detect the objects and
this can be done by the features of the extraction algorithm. Next is identifying the extracted
image of the object it is for this it will send to the classifiers for recognizing the different
types of the objects (Bernal, 2012). The final step after identifying the objects is to place the
objects in its respective category with the help of the armed robots.
Machine Description
The below figure shows the how algorithm for the image processing works. The dimensions
of the workplace are the x – y coordinates (Cabre, 2013). The standard workplace dimensions
for the objects are as:
Image’s width in pixel = (37.88) wS
Image’s height in pixel = (37.88) hS
Where,
hS is the height of workplace in the centimetres
wS is the width of workplace in the centimetres
3

Figure 1: Whole system framework
(Source: Kumar, 2014)
1. At very first the image will be taken from the workplace and resizing of the image
will take place according to the dimension of the workplace.
2. After resizing of the image, it will send to the features extraction.
3. Complying to test standards of classifiers.
4. Then after complying to test standards, it will send to the Classifier.
5. After sending to the classifiers, the objects will classify according to the coordinates.
6. And lastly, the sorting of the objects will perform by the Robotics arm.
Algorithm
The feature extraction plays a very vital role in the furnishes of the raw images (Farahmand,
2016). Shown below is the process which should be followed for the determination of the
objects for the sorting done by the robotic arm and for his features extraction should be done
before the classification of the object:
 Cluttered Scene
4

 Features Extraction
 The crop of the objects and its location filed
 Conversion of the objects to the Grayscale and after that Resizing image to the 20 x
20 pixels.
 Lastly, send to the classifiers for the testing.
Figure 2: Feature extraction should be done before classification
(Source: Kumar, 2014)
Algorithm:
1. At first, the image is read and after that converted to the grayscale. By removing the
saturation and the hue formation grayscale and preserving the luminance, conversion can
be done. The coloured images or the RGB images shown below are in the 3D
(dimensions). Now, for the conversion of this 3D into the 2D grayscale image shown
below equation should be used:
0.38 R (red) + 0.68 G (green) + 0.11 B (blue)
Figure 3: Original Image of the objects
(Source: Kumar, 2014)
5

2. After converting the original image into the grayscale image, it will again convert to
the binary image. Otsu’s is the process through which the grayscale images are
converted into the binary images.
Figure 4: Binary Images of the objects
(Source: Kumar, 2014)
3. Detection of the Edge
Once the grayscale images converted into the binary images its detection of the Edges
is applied. The detection of the edges can be done by using the “Canny methods”.
This method finds the edges of the images by means of the local maxima (Lippiello,
2013). Canny methods include the process like:
 Smoothing of the images
 Findings Gradients
 Suppression of the Non-maximum
 Double Thresholding
 Using Hysteresis for the edge tracking
6

Figure 5: Detection of the edges
(Source: Kumar, 2014)
4. Dilation of the Images
With the help of the dilation process, the images formed by the edge detection process
can enlarge the edges of the images created.
Figure 6: Dilation of the images
(Source: Kumar, 2014)
7

5. Filling of the Images
The dilated images are filled in such a manner that the edges outline of the dilated
images looks more visible and more clear.
Figure 7: Filling of the images
(Source: Kumar, 2014)
6. Analysis of the BLOB
Once the detection of the images completed now its turn to apply the rectangular
boxes around the detected objects.
Figure 8: BLOBs analysis
(Source: Kumar, 2014)
8

Robotics Armed
An experiment was done to test the number of the cycles completed by the robotic arm for
picking the objects in seconds. Firstly, the movements of the robotics armed are done from
the standby then the object was picked by the robotic armed, transportation of the objects
from one location to the other location, dropping of the objects by the armed robots and lastly
the back again to the standby position calculation is done as:
Table 1: Cycle no. and time is taken by the Robotic Arm
Cycle no. of Robotic Arm Time Taken by Robotic Arm (second)
1 14, 30
2 14, 19
3 14, 36
4 14, 13
5 14, 56
6 14, 31
7 14, 19
8 14, 21
Average 14, 28
9

UML Diagram
Figure 9: Use case Diagram
Challenges Faced
The major challenges faced while the recognition of the different types of the objects based
on their sizes and coordinates and also while developments of the image processing with the
help of the classifier are shown below:
 Images Resizing
 Developments of the test subject in the observance with classifier parameters.
 Due to the images in the raster formats their losses in pixel data.
Feasible Solution
The solutions to cure this types the problem mentioned above while recognization of the
different types of the objects based on their sizes and coordinates and also while
developments of the image processing with the help of the classifier is shown below:
10

 Uses of the concepts like SVG (Scalable Vector Graphics) for the formation of the
images.
 Select the centre image approach
 Firstly, writing of the function for enabling the MATLAB to modify and read the vector
files and the. svg/ files (Sanz, 2013).
11