2D Defect detection
General Description
The 2D Defect Detection module is used for surface quality inspection of the hood’s surfaces, ensuring detection of defects that can occur during the manufacturing process. The system employs image processing techniques and AI models to analyze and flag the positions of the visual anomalies. The flagged defects are automatically stored in the database and the operator is notified of their existence on the UI. The robustness of the detection is aided by the high-quality images that are captured by the Basler a2A5320-23ucBAS camera and enhanced by the diffuse lighting provided by the illumination panel.
| Resource | Link |
|---|---|
| Source code | Link to source code |
| Demo Video | Link to video |
✔️ Generate a GIF with screenshots or images that represent the main functionalities of the component and add it to the it img subfolder, and link it in the image below.
Contact
The following table includes contact information of the main developers in charge of the component:
| Name | Organisation | |
|---|---|---|
| Yaşar Kurt | kurt@teknopar.com.tr |  |
| Ezgi Şanli | sanli@teknopar.com.tr | |
| Amr Okasha | okasha@teknopar.com.tr | |
License
Proprietary
Technical Foundations
✔️ Describe the technical foundations of your component. Describe in detail the models, algorithms, or methods you use. If the documentation is available in a different source, for instance, an open access paper, provide here a high-level overview and links to external references.
Code blocks
You can enter block codes in Markdown cells surrounding the text you want to display as code using triple quotes (````) and adding the language name. For instance, the following code block will display some python code styled using Python language syntax:
# This is a comment
print("Hello, world!")
Tables
You can add a table like this in your document:
| This | is |
|------|------|
| a | table|
will display the following table:
| This | is |
|---|---|
| a | table |
The table cells are delimited by the (|) character. The first row and first column are used to determine the width of the table, so it is important to insert spaces if you want to control the table width and text alignment across the table. Also, note that the first row is used as headers, and that there is a separator between the header and the table body.
Note that the text in the table is automatically aligned to the left. You can control the alignment of the text using colons (:) after the column delimiter. For instance, the code:
```markdown
| Right aligned |centered aligned |
|----------------------------------:|:------------------------:|
| This is a right aligned column | and this one is centered |
```
see:
| Right aligned | centered aligned |
|---|---|
| This is a right aligned column | and this one is centered |
Lists
You can build itemized lists using the - character, and nested sublists using the tabulator or insert some spaces before the - character. For instance, the code:
- Item list number one
- Sublist number one
- This is a sublist item
- Sublist number two
- This is a sublist item
- And this is another
- Item list number two
- This is another sublist item
- Here, just one more sublist:
- One item
- And one more!
Creates the following list:
- Item list number one
- Sublist number one
- This is a sublist item
- Sublist number two
- This is a sublist item
- And this is another
- Sublist number one
- Item list number two
- This is another sublist item
- Here, just one more sublist:
- One item
- And one more!
You can also create numbered nested lists just using numbers before a '.' character. For instance:
1. Here we go
1. Sublist item 1
2. Sublist item 2
2. Here we go again
1. Sublist item 1
2. Sublist item 2
Will create this sublist: 1. Here we go 1. Sublist item 1 2. Sublist item 2 2. Here we go again 1. Sublist item 1 2. Sublist item 2
LaTeX equations
You can include mathematical expressions using LaTex notation wrapped using the dollar symbol as a delimiter. You can insert equations inline (in the same line) as text like:
This is an inline equation $e^{i\pi} + 1 = 0$
````
This is an inline equation $e^{i\pi} + 1 = 0$
You can also use the dollar symbol to include equations standalone in a new line. For instance, the code:
```markdown
$$e^x=\sum_{i=0}^\infty \frac{1}{i!}x^i$$
Will introduce an equation in a new line, like this:
$$e^x=\sum_{i=0}^\infty \frac{1}{i!}x^i$$
Latex cheat sheet
You can find a comprehensive cheat sheet for LaTeX equations here. You can also use Overleaf to find tutorials and edit equations online. However, we hope this table contains the most common LaTeX commands and symbols.
| Symbol | LaTex Command | Description | Example |
|---|---|---|---|
| $x^y$ | ^{x} |
Exponent. The expression in the exponent is passed between curly brackets ({}) | $x^{y+1}$ |
| $x_i$ | _{i} |
Subscript. The expression in the subscript is passed between curly brackets | $x_{i+1}$ |
| $\sum$ | \sum |
Summation. You can add subscripts and exponents to determine the range of the summation | $\sum_{i=0}^{10}{i!}$ |
| $\frac{x}{y}$ | \frac |
Fraction. Pass first the numerator and then the denominator using curly brackets | $\frac{1}{2}$ |
| $\int$ | \int |
Integration. Pass the expression to integrate and the variable to integrate over | $\int x^2 dx$ |
| $\sqrt{x}$ | \sqrt |
Square root. Pass the expression to square root using curly brackets | $\sqrt{x^2}$ |
| $\geq$ | \geq |
Greater than or equal to. | $\geq x^2 dx$ |
| $\leq$ | \leq |
Less than or equal to. | $\leq x^2 dx$ |
| $\neq$ | \neq |
Not equal to. | $\neq x^2 dx$ |
| $\alpha$ | \alpha |
Any greek symbol, using its english name (e.g. alpha). | $\alpha$ |
| $\beta$ | \beta |
Beta is another popular greek symbol. | $\beta$ |
| $\gamma$ | \gamma |
Gamma is another great greek symbol. | $\gamma$ |
| $\omega$ | \omega |
But Omega as in Omega level mutant is my favourite. | $\omega$ |
Matrices and vectors in LaTeX
You can just use brackets to insert vectors in LaTex, which is quite convenient. For instance, the code:
x = [1, 2, 3]
Defines a row vector x with the values 1, 2, 3.
Writing a matrix in LaTex is a bit more complicated. You can use the following syntax:
\begin{bmatrix}
1 & 2 & 3 \\
4 & 5 & 6 \\
7 & 8 & 9
\end{bmatrix}
To write a matrix like:
\begin{bmatrix} 1 & 2 & 3 \ 4 & 5 & 6 \ 7 & 8 & 9 \end{bmatrix}
Note that we write the matrix row-by-row, using two backslashes to separate rows, and the & symbol to separate cells in a row.
Integrated and Open Source Components
Overview
This component depends on the following open-source packages.
- Ultralytics
- OpenCV2
- cv_bridge
Pre-existing Components
Ultralytics
Source
The link to github repo: Ultralytics
Alternatively, if pip is available, the latest Ultralytics package can be downloaded from the commandline:
python -m pip install ultralytics
or:
pip install ultralytics
Description
Ultralytics houses the YOLO family, powerful object detection models.
Modifications
A YOLO model is trained to detect defects on the surfaces of the hoods using the livefeed from the Basler Ace camera mounted on the Universal Robot.
Purpose in AI-PRISM
The detections made by the YOLO model are tracked and their positions integrated into the ROS2 environment where they subsequently stored in the database.
License
GNU Affero General Public License v3.0
OpenCV
Source
The link to github repo: OpenCV
Alternatively, if pip is available, the latest OpenCV package can be downloaded from the commandline:
python -m pip install opencv-python
or:
pip install opencv-python
Description
OpenCV (Open Source Computer Vision Library) is an open source computer vision and machine learning software library. OpenCV was built to provide a common infrastructure for computer vision applications and to accelerate the use of machine perception in the commercial products.
Modifications
No modifications are made to the sub-component.
Purpose in AI-PRISM
OpenCV is used extensively to process the captured frames from the security camera whic are to be used by the YOLO model. It is also used annotate the frames with the detections, and encode them for database storage and UI usage.
cv_bridge
Source
The link to github repo: cv_bridge
Alternatively, if pip is available, the latest cv_bridge package can be downloaded from the commandline:
python -m pip install cv-bridge
or:
pip install cv-bridge
Description
cv_bridge is used to interface ROS2 with OpenCV which is a library designed for computational efficiency and strong focus for real time computer vision applications.
Modifications
No modifications are made to the sub-component.
Purpose in AI-PRISM
cv_bridge is used to capture the frames from the ROS2 ecosystem and allow OpenCV to process them.
License
Apache License 2.0
How to install
✔️ Every AI-PRISM component is installed using the Cluster management service. During the installation process, the user needs to configure a set of high-level parameters. Provide here a description of the parameters that the user needs to fill in.
How to use
Add usage guidelines describing how to use your component.