Hierarchical Classification of Low Resolution Thermal Images for Occupancy Estimation

Liam Walmsley-Eyre, Rachel Cardell-Oliver

Research output: Chapter in Book/Conference paperConference paper

1 Citation (Scopus)

Abstract

Occupancy estimators (sensors which can accurately estimate the number of people occupying a space) hold great potential for reducing the power usage of lighting and heating, ventilation, and air conditioning (HVAC) systems. In this paper we use low-resolution thermal sensors for occupancy estimation, due to their high temporal and spatial resolution and low invasiveness. We extend the connected component analysis and frame-classification approach taken in prior work with several new features designed to provide interesting information about the frame and components, and also examine an alternative approach that classifies the connected components individually. This is done by creating a prototype system used to collect and label data from scenes. We found that the new features significantly improve accuracy, but that the connected component classification approach is no better than frame classification for typical scenes with few occupants. In scenes with many or close occupants, We found that connected component classification produced more accurate predictions.

Original languageEnglish
Title of host publicationProceedings: 2017 IEEE 42nd Conference on Local Computer Networks Workshops LCN Workshops 2017
PublisherIEEE, Institute of Electrical and Electronics Engineers
Pages9-17
Number of pages9
ISBN (Electronic)9781509065837
DOIs
Publication statusPublished - 14 Nov 2017
Event42nd IEEE Conference on Local Computer Networks Workshops, LCN Workshops 2017 - Singapore, Singapore
Duration: 9 Oct 201712 Oct 2017

Conference

Conference42nd IEEE Conference on Local Computer Networks Workshops, LCN Workshops 2017
CountrySingapore
CitySingapore
Period9/10/1712/10/17

Fingerprint

conditioning
Sensors
air conditioning
Air conditioning
Ventilation
ventilation
sensors
Labels
temporal resolution
estimators
Lighting
illuminating
Heating
spatial resolution
prototypes
Hot Temperature
heating
high resolution
estimates
predictions

Cite this

Walmsley-Eyre, L., & Cardell-Oliver, R. (2017). Hierarchical Classification of Low Resolution Thermal Images for Occupancy Estimation. In Proceedings: 2017 IEEE 42nd Conference on Local Computer Networks Workshops LCN Workshops 2017 (pp. 9-17). [8110199] IEEE, Institute of Electrical and Electronics Engineers. https://doi.org/10.1109/LCN.Workshops.2017.59
Walmsley-Eyre, Liam ; Cardell-Oliver, Rachel. / Hierarchical Classification of Low Resolution Thermal Images for Occupancy Estimation. Proceedings: 2017 IEEE 42nd Conference on Local Computer Networks Workshops LCN Workshops 2017. IEEE, Institute of Electrical and Electronics Engineers, 2017. pp. 9-17
@inproceedings{d2e9dda259bc4383b7d0d4f6c25f9d1a,
title = "Hierarchical Classification of Low Resolution Thermal Images for Occupancy Estimation",
abstract = "Occupancy estimators (sensors which can accurately estimate the number of people occupying a space) hold great potential for reducing the power usage of lighting and heating, ventilation, and air conditioning (HVAC) systems. In this paper we use low-resolution thermal sensors for occupancy estimation, due to their high temporal and spatial resolution and low invasiveness. We extend the connected component analysis and frame-classification approach taken in prior work with several new features designed to provide interesting information about the frame and components, and also examine an alternative approach that classifies the connected components individually. This is done by creating a prototype system used to collect and label data from scenes. We found that the new features significantly improve accuracy, but that the connected component classification approach is no better than frame classification for typical scenes with few occupants. In scenes with many or close occupants, We found that connected component classification produced more accurate predictions.",
keywords = "infrared imaging, machine learning, occupancy detection, occupancy estimation",
author = "Liam Walmsley-Eyre and Rachel Cardell-Oliver",
year = "2017",
month = "11",
day = "14",
doi = "10.1109/LCN.Workshops.2017.59",
language = "English",
pages = "9--17",
booktitle = "Proceedings: 2017 IEEE 42nd Conference on Local Computer Networks Workshops LCN Workshops 2017",
publisher = "IEEE, Institute of Electrical and Electronics Engineers",
address = "United States",

}

Walmsley-Eyre, L & Cardell-Oliver, R 2017, Hierarchical Classification of Low Resolution Thermal Images for Occupancy Estimation. in Proceedings: 2017 IEEE 42nd Conference on Local Computer Networks Workshops LCN Workshops 2017., 8110199, IEEE, Institute of Electrical and Electronics Engineers, pp. 9-17, 42nd IEEE Conference on Local Computer Networks Workshops, LCN Workshops 2017, Singapore, Singapore, 9/10/17. https://doi.org/10.1109/LCN.Workshops.2017.59

Hierarchical Classification of Low Resolution Thermal Images for Occupancy Estimation. / Walmsley-Eyre, Liam; Cardell-Oliver, Rachel.

Proceedings: 2017 IEEE 42nd Conference on Local Computer Networks Workshops LCN Workshops 2017. IEEE, Institute of Electrical and Electronics Engineers, 2017. p. 9-17 8110199.

Research output: Chapter in Book/Conference paperConference paper

TY - GEN

T1 - Hierarchical Classification of Low Resolution Thermal Images for Occupancy Estimation

AU - Walmsley-Eyre, Liam

AU - Cardell-Oliver, Rachel

PY - 2017/11/14

Y1 - 2017/11/14

N2 - Occupancy estimators (sensors which can accurately estimate the number of people occupying a space) hold great potential for reducing the power usage of lighting and heating, ventilation, and air conditioning (HVAC) systems. In this paper we use low-resolution thermal sensors for occupancy estimation, due to their high temporal and spatial resolution and low invasiveness. We extend the connected component analysis and frame-classification approach taken in prior work with several new features designed to provide interesting information about the frame and components, and also examine an alternative approach that classifies the connected components individually. This is done by creating a prototype system used to collect and label data from scenes. We found that the new features significantly improve accuracy, but that the connected component classification approach is no better than frame classification for typical scenes with few occupants. In scenes with many or close occupants, We found that connected component classification produced more accurate predictions.

AB - Occupancy estimators (sensors which can accurately estimate the number of people occupying a space) hold great potential for reducing the power usage of lighting and heating, ventilation, and air conditioning (HVAC) systems. In this paper we use low-resolution thermal sensors for occupancy estimation, due to their high temporal and spatial resolution and low invasiveness. We extend the connected component analysis and frame-classification approach taken in prior work with several new features designed to provide interesting information about the frame and components, and also examine an alternative approach that classifies the connected components individually. This is done by creating a prototype system used to collect and label data from scenes. We found that the new features significantly improve accuracy, but that the connected component classification approach is no better than frame classification for typical scenes with few occupants. In scenes with many or close occupants, We found that connected component classification produced more accurate predictions.

KW - infrared imaging

KW - machine learning

KW - occupancy detection

KW - occupancy estimation

UR - http://www.scopus.com/inward/record.url?scp=85040587614&partnerID=8YFLogxK

U2 - 10.1109/LCN.Workshops.2017.59

DO - 10.1109/LCN.Workshops.2017.59

M3 - Conference paper

SP - 9

EP - 17

BT - Proceedings: 2017 IEEE 42nd Conference on Local Computer Networks Workshops LCN Workshops 2017

PB - IEEE, Institute of Electrical and Electronics Engineers

ER -

Walmsley-Eyre L, Cardell-Oliver R. Hierarchical Classification of Low Resolution Thermal Images for Occupancy Estimation. In Proceedings: 2017 IEEE 42nd Conference on Local Computer Networks Workshops LCN Workshops 2017. IEEE, Institute of Electrical and Electronics Engineers. 2017. p. 9-17. 8110199 https://doi.org/10.1109/LCN.Workshops.2017.59