Automated Pothole Detection System

Potholes (also known as chuckholes) on the road surface are detested by all motorists and pedestrians alike but for different reasons. It is well known fact that the officials responsible for maintaining roads usually turn a blind eye to this problem, sometimes until the road in question becomes no longer travelable. On the other hand, whoever has tried to collect information regarding the surface condition of heavily traveled roads is well aware of its Herculean nature. Such a task may call for number of persons to travel along roads for several days inspecting and recording surface irregularities. Besides, the drudging nature of this task forces participants to taint the collected data with errors rendering them inaccurate.

The main objective of this research is to devise an automated information gathering and analyzing technique that saves time, manpower and money while increasing the accuracy compared to the conventional method of pothole detection. In the information gathering phase, a vehicle with a sensors mounted on it travels along the road in question recording its vibration patterns. Subsequently in the analyzing phase, these data will be scanned by an algorithm to detect potholes along the path traversed earlier by the vehicle. The effectiveness of this process depends on the ability of the detection algorithm to pinpoint the potholes accurately. For this purpose, the algorithm has to identify vibration peaks that are caused exclusively by potholes on the road. However, disturbances such as the noise caused by the vibration of the vehicle’s engine interfere with the signal output generated by the mounted sensor. Therefore the data has to be preprocessed (i.e. filtered) before the algorithm can use it to detect potholes. Consequently the filtering technique also has a significant bearing on the accuracy of the final outcome of this process. Experiments conducted thus far have shown that the employed methods of filtering and pothole detection are capable of meeting the rigorous requirements towards achieving a reasonable accuracy. It is also possible to build a low-cost, road-surface-monitoring kit employing these techniques to define and evaluate an index, which gives an indication how travelable a road is. Such an index can be used to continuously caution motorists and alert the road maintenance crews to take actions as soon as it falls below a predetermined level.


Contact Us

Mailing Address

University of Colombo School of Computing
UCSC Building Complex,
35 ,Reid Avenue, Colombo 7

T: +94 -11- 2581245/ 7

F: +94-11-2587239

E: info[at]

Subscribe to Our mailing list

Social Links

Subscribe on Youtube:

Like our Facebook: