LAUSR

Flood is one of the devastating natural disaster than anything else. It is a harmful event that can risk human life, damage homes, and have huge economic impacts. Flooding creates garbage and solid waste which includes dead animals, waste products, etc. and this can increase the possibilities of spreading disease and worsening water and sanitation problems in an area and hence the need to warrant a rapid response. Several approaches of waste classification have been proposed by various researchers but only some few researches concentrate on classifying flood waste. In this study, a hybrid flood waste classification model using a 3D-wavelet transform (3D-DWT) and Support Vector Machine (SVM) was developed to address these challenges. 3D-DWT transforms the preprocessed input data by providing a time–frequency representation of the signal in different time periods in the 3D wavelet time domain, and also provides important information about the physical structure of the data and extracts the features from the main signal which serves as input to the SVM classification. The image dataset from Kaggle was classified into recyclable or non-recyclable. A total of 400 images were used to test the model to evaluate the performance and an accuracy of 85.25% and 86.00% for SVM + 2D and SVM + 3D models respectively after the model was tested with adequate times iteration. Comparing with related works which only uses the 2D-DWT and a single model, an hybrid 3D-DWT and SVM was used where 3D-DWT decomposes the images, while prediction is done using SVM thereby improving the accuracy of the system.