Concrete Cube Crack Detection using Machine Learning and Image Processing

This research proposes a cutting-edge solution to enhance the accuracy and sustainability of concrete cube testing, a crucial process in modern construction. Current practices result in 80-90% of cubes being inaccurately selected, which leads to reduced concrete strength, increased material costs, and excessive cement usage—contributing to global warming. To address these challenges, the study introduces an Industry 4.0 solution that leverages image processing and Convolutional Neural Networks (CNNs) to detect and assess cracks in concrete cubes. By analysing crack patterns, this approach will precisely determine the extent of damage, offering a more reliable and efficient method for evaluating cube quality. The goal is to improve cube selection accuracy, reduce cement waste, and promote environmental sustainability in concrete production. This advanced system promises to revolutionize concrete testing by providing a cost-effective, sustainable, and highly accurate solution, ultimately enhancing the performance and ecological footprint of construction materials.