An Automatic Framework for Parameter Extraction from Road Images with Potholes
Suwarna Gothane (CMR Technical Campus, JNTUH, India); Milind Sarode (Govt Polytechnic Yavatmal, India)
With the massive improvements in the road transport network in India, the road safety is becoming the major concern of the authorities. The road safety is primarily defined by the motor skills of the driver. Nevertheless, the effects of road conditions on safety cannot be ignored. The authorities responsible for road maintenance are struggling to provide the higher maintenance of the roads to provide better, faster and safe transportation. Thus the demand for the research to automate the detection of the road conditions cannot be disregarded. A number of research attempts are carried out in order to detect the road condition based on the potholes. Nonetheless the detection process is highly time complex and makes the maintenance process delayed. Also, majority of the parallel research outcomes failed to measure multiple potholes in a single image and cannot distinguish the potholes based on the emergency of repair. Thus this work defines a newer dimension of pothole detection for road images, which contains higher number of potholes in a single image and makes the process faster by reducing the change of false detection. The false detection is usually the images in the dataset without having significant damages. Also, this work extracts the parameters for determining the potholes existence as the major outcome. Yet another outcome of this work is to classify the potholes based on the urgency of repair. The final outcome of the work is to automate the detection facility to provide a timely maintenance alert and deliver a better road condition in India.
Journal: International Journal of Simulation: Systems, Science and Technology IJSSST V19
Published: Feb 28, 2018