Computer Vision based Volume Estimation of Potholes using ZED Stereo Camera
G Sruthi Devi1, DVR Mohan2, D Srinivasa Reddy3, Kiran Vegesna4, Gopala Krishna Murthy Nookala5
1G. Sruthi Devi, Department of Electronics & Communication Engineering, SRKR Engineering College, Bhimavaram (Andhra Pradesh), India.
2D.V.R. Mohan, Department of Electronics & Communication Engineering, SRKR Engineering College, Bhimavaram (Andhra Pradesh), India.
3D. Srinivas Reddy, Department of Electronics & Communication Engineering and Technology Centre, SRKR Engineering College, Bhimavaram (Andhra Pradesh), India.
4Kiran Vegesna, Department of Information Technology and Technology Centre, SRKR Engineering College, Bhimavaram (Andhra Pradesh), India.
5Gopala Krishna Murthy Nookala, Department of Information Technology, and Technology Centre, SRKR Engineering College, Bhimavaram (Andhra Pradesh), India.
Manuscript received on 12 May 2019 | Revised Manuscript received on 06 June 2019 | Manuscript Published on 15 June 2019 | PP: 180-185 | Volume-8 Issue-1S3 June 2019 | Retrieval Number: A10330681S319/2019©BEIESP
Open Access | Editorial and Publishing Policies | Cite | Mendeley | Indexing and Abstracting
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC-BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Abstract: Potholes asphalt state is essential when developing path maintenance platforms. So, potholes and different types of cracks required to be noticed and fixed in due time. Material required to seal the potholes are frequently performed manually, which is labor-exhaustive, time-consuming, slow and inaccurate. Calculating volume of the pothole is challenging because of its asymmetrical shape. The assessment of material required to fill the potholes is one of the significant tasks since shortage of precise amount of material will lead to low-grade compaction while an excess quantity can lead to over banding at the edge and material wastage which need to be transported from a different location. The proposed approach to attain potholes volume in real world footage is relied on stereo vision. NVIDIA GPU and Stereolab’s ZED camera hardware have been used to accomplish this task. From visualization of recorded video, depth information of every pixel in each frame is represented by 3D point clouds from 2D image and meshes are generated. Using convex hull algorithm key metrological parameters of pothole is calculated with respect to depth and results are obtained.
Keywords: Pavement, Potholes, ZED Camera, Convex Hull, Metrology.
Scope of the Article: Computer Vision