Paper Info
Title
Automatic hump detection and 3D view generation from a single road image
Abstract
Driving support systems aid in improving the road safety. There are various safety measures on the road like traffic signals, sign boards and humps to avoid possible accidents. This paper presented for road safety has two phases: The first phase includes, the development of an algorithm to detect hump from a single image captured from the camera fixed at low height to the moving vehicle. In the second phase, we generate the 3D views automatically to view the details on the road. The images have to be captured at a constant time period. The idea is to save memory by storing images instead of the video stream. Added to this, an automatic 3D view generation up to the detected hump can be visualized to know the details of the road before reaching the hump by generating the frames from the captured image. This is achieved in a four step process: First, the input image is processed to estimate the vanishing point. Second, the hump is detected using the dilation and erosion process. Third, the distance to the hump from the view point is computed, depending on which the speed of the vehicle can be reduced. Finally, the frames are automatically generated up to the detected hump to view the details on the road and act accordingly before reaching the hump using a novel approach. The proposed method has been experimented with around 400 images. The efficacy of the proposed method is presented by comparing it with the results of the depth of the hump computed manually. Also, the efficiency of the 3D view generation is shown by comparing the first frame with other frames resulting with the corresponding mean values of the difference.
Year
DOI
Venue
2014
10.1109/ICACCI.2014.6968590
ICACCI
Keywords
Field
DocType
video signal processing,dilation process,road segmentation,road safety,traffic signals,constant time period,single road image,camera,sign boards,driving support systems,moving vehicle,3d view generation,dilation and erosion,image sensors,automatic hump detection,video stream,object detection,erosion process,driver information systems,hough transform,vanishing point,hump detection,image motion analysis
Computer vision,Feature detection (computer vision),Computer graphics (images),Computer science,Image processing,Artificial intelligence
Conference
Citations 
PageRank 
References 
0
0.34
13
Authors
2
Name
Order
Citations
PageRank
Geetha Kiran A100.68
Murali S200.34