Optimally leveraging depth features to enhance segmentation of recyclables from cluttered construction and demolition waste streams

Journal of Environmental Management, Journal Year: 2024, Volume and Issue: 354, P. 120313 - 120313

Published: Feb. 16, 2024

This paper addresses the critical environmental issue of effectively managing construction and demolition waste (CDW), which has seen a global surge due to rapid urbanization. With advent deep learning-based computer vision, this study focuses on improving intelligent identification valuable recyclables from cluttered heterogeneous CDW streams in material recovery facilities (MRFs) by optimally leveraging both visual spatial features (depth). A high-quality RGB-D dataset was curated capture MRF stream complexities often overlooked prior studies, comprises over 3500 images for each modality more than 160,000 dense object instances diverse materials with high resource value. In contrast former studies directly concatenate RGB depth features, introduces new fusion strategy that utilizes computationally efficient convolutional operations at end conventional segmentation architecture fuse colour information. avoids cross-modal interference maximizes use distinct information present two different modalities. Despite clutter diversity objects, proposed RGB-DL achieves 13% increase accuracy 36% reduction inference time when compared direct concatenation features. The findings emphasize benefit incorporating geometrical complement cues. approach helps deal varied nature streams, enhancing automated recognition improve MRFs. This, turn, promotes solid management efficiently concerns.

Language: Английский

---

Sustainable urban development: Evaluating the potential of mineral-based construction and demolition waste recycling in emerging economies

Sustainable Futures, Journal Year: 2024, Volume and Issue: 7, P. 100179 - 100179

Published: March 5, 2024

Widespread illegal disposal of surging construction and demolition waste (CDW) is a prominent threat to recycling, leading resource wastage environmental issues. The lack data major barrier designing an effective management system in emerging economies. Employing mixed-method approach, cross-sectional case study was conducted on Indian city assess the existing investigate root causes low recycling rates limited demand for recycled materials. finds that (a) prevalent due "end-of-pipe" approach system, (b) policy enforcement generators weak, (c) low-rise residential buildings are contributor waste, (d) material awareness, marketing, incentives, high costs, (e) transportation costs 50% higher than cost, (f) segregation production low-value All these factors make unattractive circular economy (CE) unfeasible. Stakeholder centralized trading platform, integrated systems essential promote advance SDGs 9, 11, 12, focusing industry, sustainable cities, responsible consumption production.

Language: Английский

---

Machine learning-based automated waste sorting in the construction industry: A comparative competitiveness case study

Waste Management, Journal Year: 2025, Volume and Issue: 194, P. 77 - 87

Published: Jan. 8, 2025

This article presents a comparative analysis of the circularity and cost-efficiency two distinct construction material recycling processes: ML-based automated sorting (MLAS) conventional technologies. Empirical data was collected from Finnish companies, providing robust foundation for this comparison. Our study examines operational specifics, economic implications, environmental impacts each method, highlighting advantages drawbacks. By leveraging data-driven insights, we aim to illustrate how MLAS can enhance efficiency sustainability compared traditional methods. In our cost modeling over seven-year period, achieved cumulative €12.76 million, significantly lower than CS, which incurred €21.47 underscoring long-term MLAS. The findings underscore potential advanced AI technologies revolutionize waste management practices, offering significant improvements in accuracy, recovery rates, overall cost-effectiveness. provides valuable perspectives stakeholders industries, emphasizing importance integrating innovative achieve higher goals.

Language: Английский

---

Optimal supply chain performance: risk aversion to green innovation

Management Decision, Journal Year: 2024, Volume and Issue: 62(12), P. 3996 - 4020

Published: July 5, 2024

Purpose Although many countries are focusing on the management of construction and demolition waste (CDW) resource utilization, effect risk aversion green innovation-led enterprise performance CDW utilization supply chain is unclear when considering different innovation contexts (green led by building materials remanufacturer or recycler). This study aims to investigate how level affects under based contingency theory. Design/methodology/approach Using Stackelberg game theory, this establishes a decision model consisting remanufacturer, recycler production unit investigates influences chain. Findings The conclusions as follows. (1) For enterprise, risk-averse behaviour always detrimental his own profits. (2) follower, profits negatively correlated with in case small investment coefficient. If coefficient high, opposite result obtained. (3) When low, total decrease increases. profit shows an inverted U-shaped trend respect degree enterprise. Originality/value first construct context enterprises chain, which provides new perspective operation. explore operation mechanism evidence from prove At same time, examines interactive effects cost members, expanding theory research contingencies affecting performance. will guide members rationally face risks achieve optimal

Language: Английский

---

FE-YOLO: A Lightweight Model for Construction Waste Detection Based on Improved YOLOv8 Model

Buildings, Journal Year: 2024, Volume and Issue: 14(9), P. 2672 - 2672

Published: Aug. 27, 2024

Construction waste detection under complex scenarios poses significant challenges due to low accuracy, high computational complexity, and large parameter volume in existing models. These are critical as accurate efficient is essential for effective management the construction industry, which increasingly focused on sustainability resource optimization. This paper aims address accuracy of detection, models scenarios. For this purpose, an improved YOLOv8-based algorithm called FE-YOLO proposed paper. replaces C2f module backbone with Faster_C2f integrates ECA attention mechanism into bottleneck layer. Also, a custom multi-class dataset created evaluation. achieves mAP@50 92.7% dataset, up by 3% compared YOLOv8n. Meanwhile, count floating-point operations scaled down 12% 13%, respectively. Finally, test conducted publicly available dataset. The results demonstrate excellent performance generalization robustness.

Language: Английский

---

Change detection network for construction housekeeping using feature fusion and large vision models

Automation in Construction, Journal Year: 2025, Volume and Issue: 172, P. 106038 - 106038

Published: Feb. 6, 2025

Language: Английский

---

ShARP-WasteSeg: A shape-aware approach to real-time segmentation of recyclables from cluttered construction and demolition waste

Waste Management, Journal Year: 2025, Volume and Issue: 195, P. 231 - 239

Published: Feb. 9, 2025

Language: Английский

---

A Deep Learning Approach for Cost-Effective and Environmentally Sustainable Waste Transportation Systems in Developing Countries

Journal of Cleaner Production, Journal Year: 2025, Volume and Issue: unknown, P. 145314 - 145314

Published: March 1, 2025

Language: Английский

---

Automated recognition of contaminated construction and demolition wood waste using deep learning

Resources Conservation and Recycling, Journal Year: 2025, Volume and Issue: 219, P. 108278 - 108278

Published: April 4, 2025

Language: Английский

---

Mechanisms Driving Technological Innovation Behavior in Construction and Demolition Waste Remanufactured Products via Self-Determination Theory

Environmental Technology & Innovation, Journal Year: 2025, Volume and Issue: unknown, P. 104196 - 104196

Published: April 1, 2025

Language: Английский

---