SUMMARY
Data collection is one of the most important and costly steps of pavement management systems. Traditional methods have been widely replaced with automated data collection vehicles due to their advantages such as safety, accuracy, precision, standardization, and repeatability. However, these vehicles are very expensive due to several high-cost sensors mounted on-board which might not be financially efficient. The main goal of this paper is to propose a cost-effective data collection approach utilized to reconstruct the 3D model of a pavement surface which can be utilized to evaluate pavement condition. For this purpose, an inexpensive sensor called Kinect V2 is applied including both cameras and infrared projector to capture depth data. Having calibrated the sensor and captured data, the color images were stitched together. Then, the depth data was added to the stitched images so that the 3D model of pavement was built. This approach makes a significant difference in terms of total cost of data collection for pavement distresses which their main feature is elevation such as roughness and rutting.