With an ever increasing demand for energy, the renewable energy sector is gaining a lot of importance. Among the various renewable energy sources, wind energy is becoming more attractive compared to its counterparts (solar, biomass, etc.). Some of the main advantages of wind energy are faster payback time and that power generation is possible both during the day and night. The wind energy industry is constantly aiming at larger size rotors for increased power generation and these larger rotor blades demand stronger and more durable materials. Currently, composite materials are extensively used for wind turbine blades. With a large heterogeneity in composite materials and complications in manufacturing processes, defects in composites are inevitable. A variety of defects and damages can be seen in wind turbine blades. Composites with such defects and damage undergo a significant loss in their mechanical properties. For the use of composites as a structural material, a good knowledge of possible defects and their behaviour should be understood. To do so, the behaviour of defects under different load conditions should be assessed. Many techniques such as ultrasound scans, X-ray radiography, etc. are available to detect the defects in composites, whereas studying the damage growth in a composite is still a challenging process. In this research, a method has been developed to use thermography for characterising the delamination under dynamic loading. To demonstrate this method, a test sample with double shear configuration (DSC) and an initial delamination consisting of a Polytetrafluoroethylene (PTFE) insert was developed. The test sample was tested under fatigue loading and an infrared (IR) camera was used to monitor its thermal response and the delamination growth during loading. The data from the thermal camera was processed in two steps, firstly, fast Fourier transform (FFT) was used to transform the raw data from time domain to frequency domain. In the second step, FFT thermographs were further processed using an image segmentation algorithm. Here, the thermal plots are segmented to separate the delaminated and un-delaminated areas. By computing the number of pixels in the delaminated region, the area of delamination was obtained at each cycle and was plotted against the cycles to failure. The strain energy was computed with the help of force and displacement data from the test machine. Such signals allowed computing of the fatigue propagation curves and an understanding of the fatigue behaviour of the test samples. The results from this research were promising as the delamination behaviour reported using this method was in good accordance with a reference visual inspection method. The quantifiable output from this method can be a good starting point to study delamination experimentally and computationally. In future, this technique could be extended to different test types that cannot be quantitatively analysed using the conventional testing methods. The research was also presented at the European Conference on Composite Materials ECCM17, Munich, Germany.