Automated image analysis for differentiation of brick and masonry waste types using machine learning methods

The amount of construction and demolition waste increases with time and the building sector is responsible for a large share of the total waste production. The impact on the environment and the primary material consumption by future generations needs to be reduced. This research aims to decrease the use of primary materials in the building sector by promoting recycling of brick construction and demolition waste at the industrial level with the aid of computer vision and artificial intelligence. An image data set of different recycled brick materials was prepared with the help of an image acquisition system. The data was pre-processed, segmented, significant features were selected and different machine learning (ML) and deep learning (DL) classification models were trained. Classic ML classification methods such as random forest and support vector machine and pre-trained DL models VGG16 and ResNet50 were in python 3.7 implemented. The received recognition rates showed that both classical classifiers and convolutional neural network are able to generate excellent results.