Potholes not only make roadways less aesthetically pleasing, but furthermore present a risk to the safety of drivers, cyclists, and vehicles because of potholes, which can result in flat tires, bent rims, and even accidents if cars swerve to avoid them abruptly. This project suggests a way to combine texture and grayscale feature processing. The primary tool used in this procedure to achieve quick and precise marijuana identification is an industrial camera. To extract valuable information from digital pictures, a combination of image processing techniques is employed, such as texture filters, image grayscale, morphology, and extraction of the maximum connected domain. A machine learning model that differentiates between potholes and longitudinal, transverse, and complicated fractures is built using the library for support vector machine (LIBSVM). The technique is tested with data gathered from pastoral and agricultural regions.