Development of an image-based analysis method to determine the physical composition of a mixed waste material

Abstract

An experimental study was undertaken to assess the suitability of an image-based approach for determining the physical composition of mixed wastes. Waste samples were investigated at six different waste sorting facilities each visited twice during the study. These materials were physically sorted to examine the composition of these materials; during each of these surveys the waste was also processed to obtain digital images covering an area of 30 m 2, representing approximately 250–500 kg (3× mechanical bucket loads) of mixed waste. The images were processed using ERDAS Imagine software to assess the area covered by each component within the waste material. The composition determined from the image analysis was compared with results from the physical hand sorting. The image analysis results indicated a strong correlation with the physical results (mean r = 0.91), however it was evident that components such as film plastics and paper were being over-estimated by the image analysis approach. This manuscript provides initial results, demonstrating the potential of an image-based method, and discusses further research requirements and future applications of this technique.