With the acceleration of urbanization and the surge in construction waste production, building material recycling and reuse has become a key approach to alleviating resource shortages and reducing environmental pressures. However, traditional recycling models suffer from low efficiency, insufficient classification accuracy, and imbalanced resource allocation. Furthermore, existing recycling systems lack effective integration of data across the entire lifecycle of building materials, making it difficult to achieve accurate classification, efficient traceability, and optimized allocation. This results in low recycling rates and a high risk of secondary pollution. To this end, this paper first constructs a big data collection system for the entire lifecycle of building materials, integrating data from the design, construction, demolition, and recycling stages. Next, a deep learning algorithm is introduced to intelligently classify materials, combined with a genetic algorithm to optimize recycling route planning. Finally, a recycling and reuse supply and demand matching platform is established using Internet of Things technology, forming a closed-loop solution encompassing "data collection - intelligent analysis - optimized application." Experimental data from a pilot construction waste recycling project in a certain city shows that the classification accuracy of the CNN-SVM fusion algorithm remains consistently above 90%, reaching 91.3% in level 5 scenarios, a 39.3 percentage point improvement over manual classification. This demonstrates the feasibility and efficiency of applying intelligent algorithms and big data in this field.
Zuo et al. (Thu,) studied this question.