The present work proposes a simple procedure to evaluate the efficiency of a given image processing technique applied to weld radiographs. Specifically, special attention is given to the problem of how to evaluate whether or not a threshold value was suitably set in order to preserve relevant image features. In short, a quantitative comparison is made between the images of welds revealing defects before and after processing. Six weld radiographs of rather different image quality were chosen to be analyzed using this proposed procedure.