Principal Investigator
Prof Bharadwaj Amrutur (Chairman, Robert Bosch Centre for Cyber-Physical Systems)
Abstract
Most of the smart city components, as articulated by the 20 candidate cities chosen for Smart City implementation by the Indian Government, can greatly benefit by the use of cameras as sensing devices. Examples include surveillance, smart parking, transit management, etc. This is not surprising as cameras are a very versatile and rich sensor and hence one can extract useful information to enable a diverse set of applications. However, the negative is the high data bandwidth required to transfer data from the cameras to the computing center. Hence it is imperative that one explore ways of analyzing the camera data at the edge so that only useful information is sent to the computing center, thus minimizing bandwidth. At the same time, the whole system needs to be flexible enough to allow programmatic control of what useful information to extract as this depends on the actual application. Hence this forms the basic premise and problem for this proposal.
We will set up a test bed for distributed video analytics on the IISc campus.
We plan to use light poles as the key infrastructure element on which the camera edge nodes and the aggregator server node will be mounted. The edge camera nodes could be made out of commodity compute hardware like the snapdragon flight. The aggregator node could have further compute accelerators like the Xilinx Zynq or the parallella core or the nvidia GPUs. The camera network will be interconnected using either 802.11s or BATMAN-ADV protocols, both of which are based on WiFi.
We will explore concepts in both the algorithms as well as the architecture using a few of the problems from the smart city components like surveillance, smart parking, etc. The setup of using light poles is a new concept and might be a very promising approach for Smart City implementation.
Project Publications
| 1. | Bhargava, Srivatsa; Gorur, Pushkar; Amrutur, Bharadwaj A distributed object detector-tracker aided video encoder for smart camera networks Conference Proceedings of the 11th International Conference on Distributed Smart Cameras (ICDSC), 05.-07.09.17, Stanford (USA), 2017. Abstract | BibTeX | Links:  @conference{Srivatsa2017,
title = {A distributed object detector-tracker aided video encoder for smart camera networks},
author = {Srivatsa Bhargava and Pushkar Gorur and Bharadwaj Amrutur},
doi = {10.1145/3131885.3131920},
year = {2017},
date = {2017-09-07},
booktitle = {Proceedings of the 11th International Conference on Distributed Smart Cameras (ICDSC), 05.-07.09.17, Stanford (USA)},
pages = {69-75},
abstract = {In this paper, we propose a Region of Interest (ROI) modulated H.264 video encoder system, based on a distributed object detector-tracker framework, for smart camera networks. Locations of objects of interest, as determined by detector-tracker are used to semantically partition each frame into regions assigned with multiple levels of importance. A distributed architecture is proposed to implement the object detector-tracker framework to mitigate the computational cost. Further, a rate control algorithm with modified Rate-Distortion(RD) cost is proposed to determine Quantization Parameter(QP) and skip decision of Macro Blocks based on their relative levels of importance. Our experiments show that, the proposed system achieves upto 3x reduction in bitrate without significant reduction in PSNR of ROI(head-shoulder region of pedestrians). We also demonstrate the trade-off between total computational cost and compression possible with the proposed distributed detector-tracker framework. },
keywords = {},
pubstate = {published},
tppubtype = {conference}
}
In this paper, we propose a Region of Interest (ROI) modulated H.264 video encoder system, based on a distributed object detector-tracker framework, for smart camera networks. Locations of objects of interest, as determined by detector-tracker are used to semantically partition each frame into regions assigned with multiple levels of importance. A distributed architecture is proposed to implement the object detector-tracker framework to mitigate the computational cost. Further, a rate control algorithm with modified Rate-Distortion(RD) cost is proposed to determine Quantization Parameter(QP) and skip decision of Macro Blocks based on their relative levels of importance. Our experiments show that, the proposed system achieves upto 3x reduction in bitrate without significant reduction in PSNR of ROI(head-shoulder region of pedestrians). We also demonstrate the trade-off between total computational cost and compression possible with the proposed distributed detector-tracker framework. |