Using Machine Learning to detect defects on the steel surface
- 0 Collaborators
Surface quality is the essential parameter for steel sheet. In a steel industry manual review for defects is most troublesome and dull assignment and consequently, it is hard to guarantee the surety of imperfection free steel surface. To guarantee prompt prerequisites of clients, vision-based automatic steel surface investigation strategies have been observed to be exceptionally powerful and prevalent amid the most recent two decades. ...learn more
Project status: Published/In Market
Intel Technologies
AI DevCloud / Xeon
Overview / Usage
To provide an effective and robust approach to detect and classify the metal defects using
computer vision and machine learning.
The solution to the problem is to use image pre-processing techniques such as filtering
and then extracting the features from the image which will be used to train machine learning
models from which we can get which type of defect the steel plate has. This solution can even
be used in real time application.
Methodology / Approach
The project starts with loading the images and extracting texture features such as contrast,
dissimilarity, homogeneity, energy, and asymmetry. The features with the label are then
given to test train split function which is already present in scikit-learn library. The train-test
split function splits data and labels. The data is split 80% for training and 20% for testing.
Then the 80% data was given for training different classifiers and then the testing was
done on 20% of the data. The model which gives the highest accuracy was then selected as
the main model.
Technologies Used
Intel AI DevCloud, Intel Distribution for Python