Integrating Deep Learning Features Extraction and Image Processing for Visual Pattern Analysis in Air Quality Index (AQI) Classification

This paper presents a hybrid approach that combines deep learning with traditional image processing to identify and classify visual patterns related to air quality index (AQI) levels in digital images. This method integrates the VGG16 model, famous for deep visual feature extraction, with classical image processing techniques such as color and texture analysis. In this study, a set of outdoor image data representing six different AQI categories ranging from “Good” to “Severe” are utilized. The results of the combination of VGG16 features and traditional image processing techniques show an improvement in accuracy and pattern detection ability compared to using only a single method. This finding provides new insights into image-based air quality monitoring. The results of this discovery should encourage a more effective environmental monitoring system as well as facilitate the detection of extreme or abnormal atmospheric conditions