Design And Development Of Road Surface To Camera Image Plane Estimation System For Object Measurement

Hoo, Jian Ea (2014) Design And Development Of Road Surface To Camera Image Plane Estimation System For Object Measurement. Project Report. UTeM, Melaka, Malaysia. (Submitted)

Text (24 Pages) Design And Development Of Road Surface To Camera Image Plane Estimation System For Object Measurement 24 Pages.pdf - Submitted Version Download (1MB)

Abstract

Under the International Road Assessment Programme (iRAP, sponsorship of UK charity organization FIA Foundation for the Automobile and Society), led by AAM, JKR, JKJR, and Malaysia Institute of Road Safety Research (MIROS), a series of road safety assessment activity over 3600 km of road in Malaysia has been carried out. The system (road images acquisition and Graphical User Interface (GUI) for road safeness rating) being used to carry out the road safety assessment was found not adequate by MIROS (poor imaging quality and high cost). Using off-theshelf modules (Industrial grade Internet Protocol enabled Cameras, Microsoft Visual Studio, and open source computer vision software), a Malaysia made cost effective, reliable and robust road survey data collection and analysis system shall be jointly designed and developed to measure the road surface physical length and hence improve the road safety in Malaysia. The main aim of this project is to provide a GUI enabled road surface physical length measurement system to MIROS. The developed GUI consists of the function to capture the camera image and take the physical length between two points on the road surface. The relationship of the road surface and the camera image plane can be determined by using the correspondences on both the planes through the designed calibration process. Planar homography estimation method, using least squares as the standard approach to approximate the solution of an over determined system, is used to linearly minimize the error produced during the calibration process. The GUI facility provided by Microsoft Visual Studio and the linear algebra enabled homography estimation method in the OpenCV are used to develop the system. This system can estimate the physical length between two points on the road plane in the range of 1 0 meters until 22 meters from the camera with the relative error of up to 2%. The physical length measurement between two points can be increased by using more correspondences during planar homography estimation process.

Actions (login required)

View Item