Abstract
This study aims to utilize remote sensing techniques to evaluate the accuracy of Cloverleaf interchanges in the city of Jeddah, Saudi Arabia. The objectives of the study are: (1) to investigate the use of pixel-based and object-based image classification to extract the Cloverleaf interchanges from a high-resolution satellite image. (2) To study the relationship between the extracted Cloverleaf interchanges from the satellite image and the designed standard for Cloverleaf interchanges. Five images classification methods were investigated to derive the Cloverleaf interchanges from the images. A least square algorithm was used to draw the best fitting line of the derived roads from the classified images. The best fit lines were then compared with the design standard to assess the accuracy of the existing Cloverleaf interchanges. The results showed that Maximum Likelihood presented better classification accuracy to the other used methods by 5% to 20%. The errors of the extracted Cloverleaf interchanges were found to be within 1m to 36m from the design standard.