In the paper a method using active thermography and a neural algorithm for material defect characterization is presented. Experimental investigations are conducted with the stepped heating method, so-called time-resolved infrared radiometry, for the test specimen made of a material with low thermal diffusivity. The results of the experimental investigations were used in simulations of artificial neural networks. Simulations are performed for three datasets representing three stages of the heating process occurring in the investigated sample. In this work, the simulation research aimed to determine the accuracy of defect depth estimation with the use of the mentioned algorithm is descibed

In the paper a method for correction of heating non-homogeneity applied in defect detection with the use of active thermography is presented. In the method an approximation of thermal background with second- and third-order surfaces was used, what made it possible to remove partially the background. In the paper the simulation results obtained with the abovementioned method are presented. An analysis of the influence of correction of heating non-homogeneity on the effectiveness of defect detection is also carried out. The simulations are carried out for thermograms obtained on the basis of experiments on a test sample with simulated defects, made of a material of low thermal diffusivity.

Active thermography is an efficient tool for defect detection and characterization as it does not change the properties of tested materials. The detection and characterization process involves heating a sample and then analysing the thermal response. In this paper, a long heating pulse was used on samples with a low thermal diffusivity and artificially created holes of various depths. As a result of the experiments, heating and cooling curves were obtained. These curves, which describe local characteristics of the material, are recognized using a classification tree and divided into categories depending on the material thickness (hole depths). Two advantages of the proposed use of classification trees are: an in-built mechanism for feature selection and a strong reduction in the dimensions of the pattern. Based on the experimental study, it can be concluded that classification trees are a useful tool for the thinning detection of homogeneous material.

In recent years, a significant development of technologies related to the control and communication of mobile robots, including Unmanned Aerial Vehicles, has been noticeable. Developing these technologies requires having the necessary hardware and software to enable prototyping and simulation of control algorithms in laboratory conditions. The article presents the Laboratory of Intelligent Mobile Robots equipped with the latest solutions. The laboratory equipment consists of four quadcopter drones (QDrone) and two wheeled robots (QBot), equipped with rich sensor sets, a ground control station with Matlab-Simulink software, OptiTRACK object tracking system, and the necessary infrastructure for communication and security. The paper presents the results of measurements from sensors of robots monitoring various quantities during work. The measurements concerned, among others, the quantities of robots registered by IMU sensors of the tested robots (i.e., accelerometers, magnetometers, gyroscopes and others).