Automated volume estimation of particles based on arbitrary 2D shadow images applied to recycled aggregates

Reusing of construction and demolition waste (CDW) is essential for conserving natural resources. In order to produce the highest quality of recycled aggregates, it is essential to analyze the characteristics of the building materials and the material composition of the construction waste. As part of developing an automated process for recognizing the materials contained in recycled aggregates, it is necessary to determine the proportion of each material fraction. This involves calculating the mass by using an estimated volume and an assumed bulk density. The aim of this research is to estimate the volume of recycled aggregates. A series of randomly acquired two-dimensional (2D) shadow images were taken during a horizontal throw. No further information on spatial coordinates is available. Therefore, different approaches are being investigated and compared. Features are extracted from each image in the image data series for each particle. These features form the basis for several analytical and heuristic estimation approaches. The, these approaches are combined. Finally, a recommendation is made based on their accuracy. • Volume estimation from 2D shadow images is feasible. • Image-based volume estimation accelerates the processing time. • Information about spatial coordinates is not needed. • Analytical equations based on extracted features are reasonable. • The combination of the analytical approaches serves best for the volume estimation.