An automated solid waste bin level detection system using Gabor wavelet filters and multi-layer perception

This paper examines the application of advanced computer image processing techniques integrated with communication technologies such as radio frequency identification (RFID), global positioning system (GPS), general packet radio system (GPRS), and geographic information system (GIS) with a camera for solving the problem of solid waste collection and automated bin level detection. A new method of bin level detection is proposed that uses a Gabor wavelet filter as a feature extractor. Taking advantage of the desirable characteristics of Gabor filters, such as mask size and frequency selectivity, we have designed four filters corresponding to different frequency values for the extraction of waste features from images of the bin. The feature vector output based on Gabor filters is used as the input of the classification algorithm, which is a feed-forward back propagation (BP) network. The effectiveness of the proposed method is demonstrated on a large number of images. Both the mask size and the frequency of the Gabor filter allow it to serve as a feature extractor, and the bin level detector with the Gabor wavelet filter in conjunction with the BP acts as a classifier to provide a robust solution for solid waste automated bin level detection, collection and management.

► Bin level detection using Gabor wavelet filter as a feature extractor.
► Filters are designed for the extraction of waste features from images of the bin.
► The feature vector output on Gabor filters is used as an input classification algorithm.
► Gabor filter with feed-forward back propagation network is act as classifier.