Development of a low-cost video imaging system for pavement evaluation

Graduation date: 1998The evaluation of pavement condition is an important part of pavement\ud management. To evaluate a pavement, a distress survey has been performed mainly by\ud manual field inspections. Several automatic pavement evaluation systems have been\ud developed to overcome the drawbacks of field inspections. Automated evaluation\ud systems, however, imply their own limitations in terms of cost, technical problems, and\ud adaptability for pavement management.\ud The main purpose of this research is to develop a low-cost automatic pavement\ud video imaging system. The secondary purpose is the development of techniques to\ud process the collected video images.\ud A low-cost video image-collection system and an in-office system were\ud developed. A video test was implemented on a selected route including various pavement\ud types and several variables. As a result of the test, seven loop tests provided acceptable\ud results to allow image analysis. By using the video camera with fast shutter-speed, it was\ud decided that the survey vehicle could drive at high speed (65mph) while maintaining\ud good picture quality.\ud To evaluate the performance of the system, video and field inspections were\ud performed using two approaches: the Oregon Department of Transportation (ODOT) and\ud Metropolitan Transportation Commissions (MTC) approaches. The inspections were\ud conducted on 107 sample sections. Also, sample still images were digitized for analysis.\ud To conduct a video inspection, the Global Positioning System (GPS) technique was\ud applied for conversion of video mileage to real field mileage.\ud The results of video and field inspections were compared using statistical\ud analyses. The ODOT approach shows a good correlation between video and field\ud inspection for AC sections. In particular, patching and non-load crack indices provide\ud good correlation. The MTC-PMS analysis showed strong linear relationships between\ud video and field inspections. The analysis of crack indices from digitized images shows\ud poor repeatability for each test loop. Using general linear model analysis, variable effects\ud on crack indices were tested.\ud The cost for development and operation of the system was estimated as well as\ud cost for an enhanced prototype system. Discussions on various aspects of the developed\ud system are provided. Finally, summary and conclusion are included as well as\ud recommendations for future system development