Surface crack defects can be detected by IR thermograpgy due to the high absorption of energy within the crack cavity. It is often difficult to detect the defect in the raw data, since the signal easily drowns in the background. It is therefore important to have good analysis algorithms that can reduce the background and enhance the defect. Here an analysis algorithm is presented which significantly increases the signal to noise ratio of the defects and reduces the image sequence from the camera to one image.