Determination of the composition of recycled aggregates using a deep learning-based image analysis

Abstract Recycled aggregates (RA) are obtained by crushing inert construction and demolition waste. Their composition is variable and is currently obtained by a time-consuming manual sorting. Our work makes use of deep learning, especially convolutional neural networks (CNN), to determine this composition in near real time and in an automated way. A labelled database was created for learning of the CNNs. It consists of approximately 36,000 images of individual grains classified according to their nature. After training, our best-performing CNN reaches a validation accuracy of 97% for classifying images of grains. It is based on a Residual Network that we customised in order to improve its performance. Moreover, we evaluated the mass of the grains by assuming that grains of a given nature have a constant form and density. Our approach was compared with manual sorting. There was less than 2% of difference in mass for most RA natures tested.

[1]  Bogdan Grigore Cazacliu,et al.  Usage of air jigging for multi-component separation of construction and demolition waste. , 2017, Waste management.

[2]  Michael S. Bernstein,et al.  ImageNet Large Scale Visual Recognition Challenge , 2014, International Journal of Computer Vision.

[3]  P. Devillers,et al.  Influence of initial saturation degree of recycled aggregates on fresh cement paste characteristics: consequences on recycled concrete properties , 2018 .

[4]  José Luis Duarte Ribeiro,et al.  Modeling of mechanical properties and durability of recycled aggregate concretes , 2012 .

[5]  Nitish Srivastava,et al.  Dropout: a simple way to prevent neural networks from overfitting , 2014, J. Mach. Learn. Res..

[6]  Min-hong Zhang,et al.  Effects of the Parent Concrete Properties and Crushing Procedure on the Properties of Coarse Recycled Concrete Aggregates , 2013 .

[7]  J. de Brito,et al.  Durability and shrinkage of concrete with CDW as recycled aggregates: Benefits from superplasticizer’s incorporation and influence of CDW composition , 2018 .

[8]  George Wardeh,et al.  Mix Design and Recycled Aggregates Effects on the Concrete’s Properties , 2018 .

[9]  Julio R. Valdes,et al.  Aggregate production: Fines generation during rock crushing , 2007 .

[10]  C. Poon,et al.  Properties of concrete prepared with low-grade recycled aggregates , 2012 .

[11]  Régis Sebben Paranhos,et al.  Stratification in air jigs of concrete/brick/gypsum particles , 2016 .

[12]  Xiaoping Ji,et al.  Performance of cement-stabilised crushed brick aggregates in asphalt pavement base and subbase applications , 2016 .

[13]  Alessandro Marradi,et al.  C&D Waste For Road Construction:Long Time Performance Of Roads ConstructedUsing Recycled Aggregate ForUnbound Pavement Layers , 2006 .

[14]  L. V. Eycken,et al.  Evaluation of Uniform Color Spaces Developed after the Adoption of CIELAB and CIELUV , 1994, Color Research & Application.

[15]  Jian Sun,et al.  Identity Mappings in Deep Residual Networks , 2016, ECCV.

[16]  Jian Sun,et al.  Deep Residual Learning for Image Recognition , 2015, 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[17]  Geoffrey E. Hinton,et al.  Learning representations by back-propagating errors , 1986, Nature.

[18]  E. Ghorbel,et al.  Relationships between recycled concrete aggregates characteristics and recycled aggregates concretes properties , 2016 .

[19]  Léon Bottou,et al.  The Tradeoffs of Large Scale Learning , 2007, NIPS.

[20]  Wenzhong Zhu,et al.  Effect of the constituents (asphalt, clay materials, floating particles and fines) of construction and demolition waste on the properties of recycled concretes , 2015 .

[21]  Lawrence D. Jackel,et al.  Handwritten Digit Recognition with a Back-Propagation Network , 1989, NIPS.

[22]  Gang Sun,et al.  Squeeze-and-Excitation Networks , 2017, 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition.

[23]  J. Serra,et al.  An overview of morphological filtering , 1992 .

[24]  James S Lai,et al.  Three-Dimensional Image Analysis of Aggregate Particles from Orthogonal Projections , 1996 .

[25]  R. D. T. Filho,et al.  A novel mix design methodology for Recycled Aggregate Concrete , 2016 .

[26]  P. Turcry,et al.  Reuse of recycled crushed concrete fines as mineral addition in cementitious materials , 2016 .

[27]  Enzo Martinelli,et al.  Alternative processing procedures for recycled aggregates in structural concrete , 2014 .

[28]  Pierre Soille,et al.  Morphological Image Analysis: Principles and Applications , 2003 .

[29]  C. H. Sampaio,et al.  The potential of using air jigging to sort recycled aggregates , 2014 .

[30]  Geoffrey E. Hinton,et al.  Deep Learning , 2015, Nature.

[31]  Vivian W. Y. Tam,et al.  Mitigation of carbon dioxide by accelerated sequestration in concrete debris , 2020, Renewable and Sustainable Energy Reviews.

[32]  Akh Kwan,et al.  Particle size distribution analysis of coarse aggregate using digital image processing , 1998 .