3D Point Cloud from Close-Range Photogrammetry for Defect Characterisation of Rubberised Concrete

In the world of building materials, engineers are constantly searching for ways to make concrete more flexible and sustainable. One promising approach involves mixing recycled rubber from old tires into the concrete. This creates a composite material that can bend and stretch without breaking as easily as traditional concrete, offering a way to reuse waste while improving the safety of structures. However, this flexibility comes with a unique challenge: when these rubberized concrete beams are stressed, they do not crack in simple, straight lines. Instead, the cracks twist, turn, and follow a complex, jagged path through the material, creating a rough and uneven surface that is difficult to measure. Understanding exactly how these cracks form and grow is essential for designing better mixtures, but seeing and measuring these tiny, twisting fractures has long been a difficult task for scientists working in laboratories.

For decades, researchers have relied on high-tech laser scanners to map the surfaces of structures. These devices, known as terrestrial laser scanners, work by firing a beam of light at an object and measuring the time it takes to bounce back. While these machines are excellent for mapping large buildings or bridges, they struggle with the tiny details found in a laboratory setting. The laser beam itself is often too wide to fit into the narrowest gaps of a micro-crack. When the beam hits a rough surface with a tiny void, the light simply bridges over the gap, filling in the missing detail and making the crack look smaller or smoother than it really is. It is like trying to measure the depth of a thin hairline crack with a ruler that is thicker than the crack itself; the tool physically cannot reach the bottom, leaving the true shape of the damage hidden.

To solve this problem, a team of researchers at the University of New South Wales turned to a different kind of technology: close-range photogrammetry. Instead of using lasers, this method uses standard digital cameras to take hundreds of overlapping photographs of an object from many different angles. By analyzing the slight shifts in how the object appears in each photo, sophisticated software can reconstruct a highly detailed three-dimensional model of the surface. The researchers applied this technique to rubberized concrete beams that had been tested to the point of failure. They captured images using both a high-end professional camera and a modern smartphone, then fed the photos into software that stitched them together into a dense cloud of millions of points, creating a digital twin of the cracked surface.

The results showed that this camera-based approach could see details that the laser scanners missed. The professional camera, equipped with a large sensor, produced a model with a resolution fine enough to distinguish features smaller than a millimeter. When the researchers measured a specific crack in their digital model, the width came out to approximately 2.3 millimeters, which matched almost perfectly with a physical measurement taken with a caliper. In contrast, the smartphone camera, while still useful, produced a model with slightly more noise and less precision, showing that the quality of the camera sensor matters significantly when trying to capture such fine details. The study confirmed that for the sub-millimeter cracks typical of rubberized concrete, a high-resolution camera is a far more effective tool than a standard laser scanner.

Beyond just seeing the cracks, the team developed a way to isolate them from the rest of the surface. Because the cracks are often darker than the surrounding concrete, the software could use color information to highlight the damaged areas. By converting the 3D data into a format that emphasized these tonal differences, they were able to separate the actual fracture points from the background noise. This allowed them to create a clean, structured map of the crack's path, revealing the complex, winding geometry that traditional two-dimensional images often fail to capture. This level of detail is crucial because the shape of the crack tells engineers how the material is behaving under stress, information that is vital for improving the mixture design.

The researchers also used this method to measure how much the concrete moved during the test. By comparing the 3D model of the beam before the test with the model after it broke, they could calculate the exact amount of displacement across the entire surface. The data showed that the area around the main crack moved an average of 2.2 millimeters, with some parts shifting up to 2.9 millimeters. This ability to track movement across the whole surface, rather than just at a single point, provides a much richer picture of how the material deforms. It proves that this photogrammetric workflow is not only capable of finding tiny defects but can also quantify the damage with high precision.

This study demonstrates that close-range photogrammetry offers a flexible and accessible alternative to expensive laser systems for laboratory testing. It allows researchers to capture the true, tortuous nature of cracks in rubberized concrete without the physical limitations of laser beams. By establishing a reliable way to generate these high-resolution 3D models, the work lays the foundation for future automated systems that can analyze material performance more deeply. The researchers plan to expand this work by developing fully automated tools to measure the length, width, and depth of every crack automatically, and by exploring how these cracks grow over time using video-based techniques. For now, the study confirms that with the right camera and a systematic approach, scientists can finally see the hidden complexity of how sustainable building materials fail, leading to safer and more durable structures in the future.