A new image processing technique was investigated to assess its ability to detect surface flaws on an on-tower wind turbine blade (WTB). The method was tested by varying the parameters of the surface flaws as well as the parameters of the method. It was found that detecting and quantifying cracks as small as hair thickness with computer-based optical inspection is feasible and the orientation of a crack was not sensitive to image processing so that the inspection camera does not need to be set up at a specific angle to detect cracks. It was also found that uneven background illumination was significantly reduced by optimizing the threshold value using the Canny method. In addition, the accuracy of quantifying a crack was improved by reducing noise with the intersection of two processed images from Sobel and Canny methods.