The monocular vision algorithms proposed by iLab provides a new solution to efficiently and reliably quantify construction wastes.
Construction is an emission-intensive industry, and construction waste management (CWM) can help alleviate its adverse impact to the environment. Effective CWM requires quantitative information on its composition and amount, which, however, is not easy to quantify because of the bulky, cluttered and mixed nature of the materials.
Typical construction waste dumps, which are bulky and cluttered
To tackle the challenge of construction waste quantification, I, together with my colleagues at iLab, developed a computer vision algorithm that only requires a single camera and other common sensors (e.g., weighbridge, range finders) to estimate the composition and volumes of different materials in a waste dump.
Our proposed monocular vision approach for construction waste quantification
(a) Photos used to calibrate the camera; (b) Reference points used to calculated the camera extrinsic parameters; (c) The installed range finders for waste depth measurement.
Examples of the waste quantification results. First column: the input photo; Second column: the reconstructed 3D geometry of the waste materials; Third column: 3D semantic of the waste materials with estimated volumes.
Relevant Content
For more details, please refer to our papers published on Resources, Conservation & Recycling (RCR):
[1] Lu, W., Chen, J., & Xue, F. (2022). Using computer vision to recognize composition of construction waste mixtures: A semantic segmentation approach. Resources, Conservation and Recycling, 178, 106022. https://doi.org/10.1016/j.resconrec.2021.106022
[2] Chen, J., Lu, W., Yuan, L., Wu, Y., & Xue, F. (2022). Estimating construction waste truck payload volume using monocular vision. Resources, Conservation and Recycling, 177, 106013. https://doi.org/10.1016/j.resconrec.2021.106013