Classification of water contamination developed by 2-D Gabor wavelet analysis and support vector machine based on...

Area of Science:

• Analytical Chemistry
• Environmental Science
• Spectroscopy

Background:

• Traditional water quality testing methods face challenges including long analysis times, low sensitivity, reagent dependency, and waste generation.
• These limitations impede frequent monitoring and rapid detection of pollutants in urban water systems.
• Effective identification of specific pollutants is crucial for maintaining water supply safety.

Purpose of the Study:

• To develop an advanced method for identifying specific pollutants in urban water supplies.
• To overcome the limitations of conventional water quality analysis techniques.
• To enhance the efficiency and accuracy of water quality monitoring.

Main Methods:

• Utilized three-dimensional (3D) fluorescence spectroscopy for water quality monitoring.
• Developed an identification method employing two-dimensional (2D) Gabor wavelets and support vector machine (SVM) multi-classification.
• Applied Delaunay triangulation for spectral pre-processing to remove scattering, and 2D Gabor wavelets with block statistics for feature extraction.

Main Results:

• The proposed method effectively describes 3D fluorescence spectral characteristics, outperforming principal component analysis.
• Achieved high classification accuracy for pollutants, particularly those with overlapping spectral peaks.
• Demonstrated the efficacy of 2D Gabor wavelets and block statistics in spectral feature description.

Conclusions:

• The combination of 3D fluorescence spectroscopy, 2D Gabor wavelets, and SVM offers a powerful tool for accurate and efficient water pollutant identification.
• This advanced technique addresses the shortcomings of traditional methods, enabling high-frequency water quality monitoring.
• The method shows significant promise for real-time or near-real-time assessment of urban water quality.