Cracks in roads is a major problem that affect the quality of the roads. Roads tend to crack with the passage of time as well as due to natural disasters such as earthquakes. Current techniques rely mainlyon visual inspection which drains away precious resources.The solution introduced here will automatically acquire images of roads which will then be processed and locations with cracks will be identified using image processing.The process is two staged: first an automated rover, that can be operated remotely, will run on the road and collect images of the whole street. These images, which will be saved in an SD card, will be transferred to a PC and the convolutional neural network based crack detection system can identify the ones that have a crack. Through the GPS coordinates obtained by the rover, areas of the cracked roads will be identified. The user of this product can then use this information to take further action.The product gives results that are 90.4% accurate. The rover is easy to control and does not require a highly skilled technician. The software can run on most PCs and does not require very high computation power