|Main Objective: This research position focuses on the research and development of a system able to process images of the surface of flexible road pavements, to allow for the automatic detection and characterization of road surface defects.
Although cracks with linear development are considered the most common pavement surface degradation found by road inspectors, other important surface defects like cracks with alligator pattern, potholes, routing, raveling, among others, represent important data about pavement surface condition to allow for an adequate quantitative and qualitative evaluation of road pavement surface degradation. Images of road surface taken using imaging systems composed of one regular camera (considered as been non-expensive imaging systems) have been used by researchers, mainly on the detection of cracks with linear development. However, these images are 2D data structures, been difficult to detect defects like potholes, routing or raveling when processing them, because the depth information along a profile of the road defect is missing. Light-field cameras have recently become available and the processing of images taken using these imaging devices may allow for an adequate estimation of depth information along a profile of the above mentioned road defects. Therefore, the mainly objectives are: (i) exploring the capabilities of light-field cameras to allow for the creation of a 3D data structure of road pavement surface, to be used as an efficient source of information on the computation of transversal profiles of a road lane; (ii) combining 2D detection data (defected and non-defected pavement surface areas obtained by the analysis of variations in pixels intensities along the entire image) with those transversal profiles, to characterize surface defects into the following types: cracks with alligator pattern; potholes; routing and raveling (conducting a texture analysis, as raveled locations of road pavement surface frequently exhibit greater roughness in comparison to non-raveled ones, due to the presence of larger aggregates); (iii) to develop a user-friendly computational environment including all the developed tasks.
|Reference: IT - Project EEA/50008|
|Name: CrakIT-LF: Automatic detection and characterization of road distresses in images of flexible road pavements|
|Start Date: 01-04-2016|
|End Date: 01-04-2018|
|Team: Henrique José Monteiro Oliveira, Paulo Luis Serras Lobato Correia|
|Groups: Multimedia Signal Processing – Lx|
|Local Coordinator: Henrique José Monteiro Oliveira|