Characterization of Texture Properties of Pavement Images as Aid to Automated Comprehensive Pavement Evaluation

摘要: Automated analysis of pavement images collected by survey vehicles equipped with digital cameras is increasingly becoming popular in evaluation. The majority the existing analytical techniques used image processing are focused on detecting cracks through automation. In these techniques, it usually assumed that have darker pixel intensities compared to background gray-scale from 0 255. However, since surfaces also contain other dark features has become quite a challenge differentiate extraneous surface. Hence authors feel would be beneficial characterize optical texture identify such prior attempting detect cracks. Tools for characterizing surface at least three significant applications; (1) they can applied automatically evaluate respect distress types like rutting, faulting, spalling and loss skid-resistance, (2) monitor deterioration traffic environmental conditions, (3) as indicators quality imaging conditions adequate illumination etc. This paper focuses identifying definitive tools properties concrete images. particular this study were Florida Department Transportation’s Multi Purpose Survey Vehicle. Digital binarizing moments, quantitative measures autocorrelation functions Laws' measures, filtering Fourier power spectral density, well statistically based principal component been work Results obtained most above enabled effective characterization specific Based preliminary findings anticipated extension asphalt pavements well, envision development general set not only improve reliability automated surveys but expand its scope include multiple types.