To search, Click below search items. |
| All Published Papers Search Service |
|
|
|
Title |
Features Extraction Techniques and Support Vector Classifier for Air Particulate Matters Levels Classification |
|
Author |
Rayan Awni Matloob, Mohammed Ahmed Shakir |
| Citation |
Vol. 26 No. 5 pp. 106-116 |
|
Abstract |
Smog is a grave environmental issue, especially in urban areas with high tiers of air pollution from traffic, industry, and other sources. It comprises a mix of gases and particles, including PM2.5, which is particulate matter with a diameter of fewer than 2.5 micrometres. In this work, two schemes are separately proposed. The first scheme applies principal component analysis (PCA) to select some suitable image features. The second scheme employs three wavelet transform filters separately for this purpose. Both approaches use a support vector machine classifier (SVC) for the classification stage. These two schemes classify the images into different classes based on their PM2.5 concentration tiers. The used data are 1990 images in total, randomly divided into (80% for the training process, and the remaining images are used to test the schemes and assess their estimation accuracy). The experimental results demonstrate that the first proposed algorithm achieved (87.15%) accuracy, while the second scheme achieved three accuracy results, 90.43%, 87.91%, and 85.90 % for Ricker, Haar, and Daubechies filters respectively. |
|
Keywords |
Air quality, particulate matter (PM2.5), Image Classification, principal component analysis (PCA), wavelet transform, support vector classifier (SVC). |
|
URL |
