Road roughness poses challenges for the government due to its complexity and costly measurement tools. Our study proposes a novel, automated model to assess and rank road roughness without human interaction or expenses. By using images taken from a drone, our model shows the pattern of roads from the captured image using Gray-Level Co-occurrence Matrix (GLCM) features Homogeneity (H) and Energy (E) and then takes the spikes of its distributions then takes these spikes to get optimal value K for Kmean to segment the image, the result of first model enter to second model that makes that apply the Local Binary Pattern (LBP) Algorithm, To show the pattern more accurately, the excess elements in the image are deleted, and the remaining noise in the image is also deleted. After that, the image enters CNN to get the outcomes by classifying it into which category this roughness belongs. Achieving a high precision of 91. 94%, our model ensures a precise classification of road roughness categories, offering cost-effective and reliable solutions for road assessment.