The Exercise
Image binarisation is a process used to simplify images for further investigation or simply to reduce the size of an image file. It simply replaces each pixel in an image with one of two possible colours (usually either black or white). The object of this coursework is for you to write an extensible Graphical User Interface (GUI) that can allow a user to load image data, in the form of text files, and perform image binarisation and to save the results.
1. The Image data file formats
There are 3 image file formats to consider in this exercise. All of them are text based and the first line in the file is the name of the format used.
i. Greyscale Image
Each line of data contains information about one pixel.
It has the form x; y; v. Where x; y are the coordinates of the pixel and v is the intensity of the pixel. The intensity ranges from 0 (black) all the way to 255 (white).
Below is the first 5 lines of the file GreyImage.txt.
Greyscale Image
61, 85, 2
69, 8, 21
39, 55, 125
46, 37, 92
2. Colour Image
Each line of data contains information about one pixel.
It has the form x; y; r; g; b. Where x; y are the coordinates of the pixel and r; g; b are the red, green and blue components of the colour.
Below is the first 5 lines of the file ColourImage.txt.
Colour Image
117, 69, 76, 85, 66
7, 126, 82, 63, 56
81, 16, 132, 146, 95
13, 178, 30, 16, 39
Your program must be able to read in files using the above 2 formats. Your program must also be able to write a file in the following Binary Image format:
i. Binary Image
This format has the form x; y; b. Where x; y are the coordinates of the pixel and b is either 0 or 1. Below is the first 5 lines of GreyImageBinaryOutput.txt.
Binary Image
61,85,0
69,8,0
39,55,1
46,37,0
2. Binarisation Methods
For a user selected threshold, t. The binarisation process works as follows:
- For Greyscale Images if v < t then the output is 0 else it is 1.
- For Colour Images if r+g+b/3 < t then the output is 0 else it is 1.
Automatically selecting a threshold
Your program is required to suggest a threshold to use. This is calculated in the following way:
- For Greyscale Images t is the average of all the intensity values (v) from the image.
- For Colour Images The intensity of a pixel is r+g+b/3. t is the average of all these computed intensity values from the image.
The suggested threshold is int(t), (since t is likely not to be a whole number).
3. Coding the solution
Your solution must follow the specification below:
Display
For the display use 2 canvas widgets, one to show the input image and one to show the result of the binarisation. The pixels in the image are to be displayed using the create_rectangle method.
Above the leftmost canvas have a label that will display the full file name of any succesfully loaded file. Above the rightmost canvas have another label with content Select Threshold (0-255) an entry box and finally a button.
Have a menu item called File which has sub-items load and save, each link to the relevant file manager window.
Once a file has been succefully loaded, its name will be displayed in the leftmost label. The suggested threshold will be computed and placed in the entry box. When the process button is pressed, the binaristaion will occur using the threshold value in the entry box. The result will be displayed in the rightmost canvas.
To successfully complete this coursework you will need to investigate the functionality available in tkinter, specifically research the canvas widget.
4. Code Provided
A structure to your program is already provided for you to download. You must use the structure provided and follow any instructions provided in these files.
- BinaryConverter.py. This class is your GUI. You can run it and it will create a small window. Read through it carefully. You have been provided with working code to display pixels onto a canvas. The colour a create_rectangle requires is a String that follows a specific format. I have also provided a method called _determineColorValue() for each class of image which will create this String for you.
- GUIconnect.py. An abstract class which is the parent to both ColourImage.py and GreyScaleImage.py. Read the comments in the method definitions carefully. These provide further instructions of what to write.
- ColourImage.py and GreyScaleImage.py are where you put code specific to each file format.
- BinaryImage.py. This class stores the result of the binarisation process and the code for the file format for saving to.
Hint: ColourImage.py, GreyScaleImage.py and BinaryImage.py are missing constructors. You will need to implement them.
Note: You may only import from tkinter. You must not import any other libraries.
Extensibility and Polymorphism
The Graphical User Interface should not be too tightly coded to binarisation approaches and their file formats. In fact we aim to have a situation that if a programmer wishes to include their own file format and binarisation approach they could include it with minimal effort. In addition if the programmer wishes to change the file format for saving the binary image, the code for your GUI (BinaryConverter) should not need to be ammended at all. To achieve this, all image types are implemented as classes that inherit from the abstract class GUIconnect (an abstract class has method definitions but with no usable method bodies). Your Graphical User Interface should communicate with these classes through the methods defined in that abstract class.
Within your Graphical User Interface use polymorphism to make the inclusion of another image file class as simple as possible.
You may always assume that the first line in an image file is a string containing the file type.
Preliminaries
Download fromMoodle BinaryConverter.py, GUIconnect.py, ColourImage.py, GreyScaleImage.py and BinaryImage.py . Place them all in the same directory of your choosing. Read through this code and understand the structure.
Finally download the test data:
GreyImage.txt, ColourImage.txt, GreyImageBinaryOutput.txt, ColourImageBinaryOutput.txt.
- For GreyImage.txt, the suggested threshold should be 97 and the file output after saving the result is in GreyImageBinaryOutput.txt.
- For ColourImage.txt, the suggested threshold should be 70 and the file output after saving the result is in ColourImageBinaryOutput.txt.
Attachment:- Assignment Files.rar