The primary goals of picture enhancement are to draw attention to hidden elements in a photograph. Picture enhancement methods may be used in a variety of ways to increase a picture’s visual quality. The objective of this paper is investigating the relative performance of three methods to improve pictures for rapidly processing pictures. These two from these methods are classical methods. They are histogram equalization (HE) and median filter, while the third is an artificial method from machine learning; it is the convolutional neural network (CNN). Moreover, in responding to the humanitarian and economic disasters caused by road cracks in any country choosing an algorithm that can improve the photos’ quality in a reasonable amount of time is essential. The viewing conditions and the imaging modalities all apply a significant role in selecting the appropriate method. The results that were obtained showed that the convolutional neural network method (CNN) is superior to the rest two methods. MATLAB code was applied as programming language for practical application.