Automatic system for carbonation depth measurement (now manual procedure). The specimen sprayed with a colorimetric solution is placed on a back-illuminated bench with calibration markers. A camera takes a picture, then processed for contaminated perimeter determination through contaminated/non-contaminated areas transition, discarding agglomerations (thanks to convolutional neural network) affecting results. Penetration depth is given as average and maximum values. Measurement is more accurate than EN13295 standard method, since all the perimeter points are considered, and also faster than it.

Patent Status


Priority Number


Priority Date





The contaminants penetration is one of the main degradation causes affecting reinforced concrete structures, causing deterioration and durability issues. Maintenance and repairing operations costs are estimated at 3-4% of the gross national product. The cost of each intervention is up to 2% of the whole structure cost and exponentially increases with the degradation status (De Sitter’s law). Regular periodic inspections play a pivotal role, enabling timely interventions limiting damages and related costs. Potential customers could be engineering firms, consulting companies and measuring systems developers.


Contaminants penetration into a cement-based structure can affect durability and impact on maintenance and repairing costs (3-4% of gross national product). Carbonation is one of the main causes of concrete degradation, especially in those countries characterized by a warm and relatively humid environment, since moisture content accelerates carbonation. In this way, the passive layer protecting the steel rebars is damaged and corrosion can start if steel interacts with moist air, shortening the reinforced concrete structure service life. For carbonation depth measurement, the present methods complies with the EN 13295 standard and consists in taking samples of the material to be examined and sprinkling their exposed surface with a solution containing a colorimetric indicator (phenolphthalein), which will turn red-fuchsia in the non-carbonated areas. However, the transition between the two areas is not clear-cut, so it is up to the operator’s subjectivity and experience to establish the optimal transition perimeter. This leads to poor repeatability of the measurement and to a subjective result. Moreover, the measurement is taken in only few points not considering the whole specimen area, hence the measurement is operator-dependent. In addition, some types of agglomerates do not color anyway, and should not be considered as a carbonated area when taking measurements. Obviously, also the computational time is to be taken into account, being this a manual operation.

Current Technology Limitations

As concerns carbonation, the EN13295 standard suggests spraying a phenolphthalein solution on the target to highlight the presence of carbonated/non-carbonated areas. The phenolphthalein solution changes its colour in relation to the pH of the material: in the non-carbonated part of the specimen, where concrete has still a highly alkaline behaviour (pH > 9), a purple-red coloration is obtained, whereas in the carbonated area of the specimen (pH < 9) no colour change is observed [9]. The standard describes a procedure in which the operator is asked to manually measure the carbonation depth (dk), i.e. “the average distance, measured in mm, from the surface of the concrete or mortar where the CO2 has reduced the alkalinity of the hydrated cement to an extent such that an indicator solution based on phenolphthalein remains colourless”, by using rulers, callipers, etc.

Even if this measurement method is relatively easy to perform, it suffers from subjectivity, due to operator’s experience, colours perception, and manual ability. Hence, the measurement will be characterized by low repeatability and low reproducibility, leading to non-optimal results.

The proposed method aims at developing an automated and objective approach, based on machine vision, for measuring penetration depth of contaminants in concrete (not limiting to carbonation). Artificial intelligence is used to automatically discard aggregates that would lead to misleading results.

Killer Application

The project was initially developed as a laboratory test bench:

  • Camera and diffuse illumination system
  • Bar with fiducial markers for pixel-to-mm conversion
  • Back-light LED illumination plate

Two types of stand-alone systems can be derived:

  1. In the laboratory: specific housing with an illuminated background (with markers or lidar sensors for target-sensor distance measurement), an illuminating system and a camera. Data processing system can be controlled via dashboard/smartphone
  2. In the field: mobile device used on a structure with visible reinforcement bar. The area is illuminated and framed, then processed.

Our Technology and Solution

The proposed measurement system (hardware and software) is targeted to the automated detection of carbonation depth in cement-based specimens, previously treated on an exposed surface with a colorimetric solution. The system includes:

  • An image acquisition device, to frame the whole exposed surface;
  • A first illuminating device projecting a diffused light to the exposed surface, from the same direction of observation as the imaging device;
  • A data processing device for image digitalisation and exposed surface selection. Hence, the perimeter and area of the exposed surface are computed, as well as the area coloured by the colorimetric indicator, then subtracted from the total one to derive the uncoloured portion area. Then, the contaminants penetration is obtained as the ratio between area and perimeter, at first in pixels and then converted into millimeters.

Such a measurement system results to be objective (not operator-dependent), precise and accurate (expanded uncertainty with coverage factor k = 2: 0.08 mm). Both average and maximum penetration depth values are provided, as indicated in the EN 13295 standard for carbonation depth measurement. Moreover, the correlation with the method suggested by the EN 13295 standard is very strong (R2 = 0.96). This technology is suitable for a stand-alone system, which can be deployed in different forms depending on the target application, which can be in laboratory or in field.


  • Objectivity: automatic method, not operator-dependent
  • Repeatability: 0.0017 mm2
  • Accuracy (type A uncertainty): 0.08 mm (k = 2)
  • Execution time: -98% vs standard method
  • No influence of aggregates, automatically identified through VGG16 convolutional neural network
  • Measurement on the whole specimen, not limiting to few points (contrarily to EN 13295 standard)
  • Independence from ambient lighting conditions
  • Applicable to any specimen geometry
  • Independent from rotations with respect to the imaging system
  • Not-subjective threshold for segmentation (back-illumination system, HSV domain analysis)


In order to commercialise the proposed system in the form of a portable cabinet or as integrated in a mobile device (e.g. smartphone, PDA – Portable Data Assistant), possible directions of work could be the following ones:

  • Contracts with customers of the service;
  • Creation of a spin-off;
  • Search of EU fundings within Fast Track to Innovation actions.

Potential customers could be engineering firms, consulting companies and measuring systems developers.

Further research is necessary to develop the innovation as a stand-alone system (both portable cabinet or mobile device with specific application installed).

Review the Technology