Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2016 Mar 30;16(4):453.
doi: 10.3390/s16040453.

Human Detection Based on the Generation of a Background Image and Fuzzy System by Using a Thermal Camera

Eun Som Jeon  1 , Jong Hyun Kim  2 , Hyung Gil Hong  3 , Ganbayar Batchuluun  4 , Kang Ryoung Park  5
Affiliations

Human Detection Based on the Generation of a Background Image and Fuzzy System by Using a Thermal Camera

Eun Som Jeon et al. Sensors (Basel). .

Abstract

Recently, human detection has been used in various applications. Although visible light cameras are usually employed for this purpose, human detection based on visible light cameras has limitations due to darkness, shadows, sunlight, etc. An approach using a thermal (far infrared light) camera has been studied as an alternative for human detection, however, the performance of human detection by thermal cameras is degraded in case of low temperature differences between humans and background. To overcome these drawbacks, we propose a new method for human detection by using thermal camera images. The main contribution of our research is that the thresholds for creating the binarized difference image between the input and background (reference) images can be adaptively determined based on fuzzy systems by using the information derived from the background image and difference values between background and input image. By using our method, human area can be correctly detected irrespective of the various conditions of input and background (reference) images. For the performance evaluation of the proposed method, experiments were performed with the 15 datasets captured under different weather and light conditions. In addition, the experiments with an open database were also performed. The experimental results confirm that the proposed method can robustly detect human shapes in various environments.

Keywords: fuzzy system; generation of background image; human detection; thermal camera image.

PubMed Disclaimer

Figures

Figure 1
Figure 1
Overall procedure of the proposed method.
Figure 2
Figure 2
Flow chart of generating a background image (model).
Figure 3
Figure 3
Examples of generating the background image from database I: (a) preliminary background image obtained by median value from the sequence of images; (b) extracted candidate human area by binarization; (c) extracted human areas by labeling, size filtering and a morphological operation; and (d) the final background image.
Figure 4
Figure 4
The first example for generating a background image from database III: (a) preliminary background image obtained by median value from the sequence of images; (b) extracted candidate human area by binarization; (c) extracted human areas by labeling, size filtering and a morphological operation; and (d) the final background image.
Figure 5
Figure 5
The second example for obtaining a background image from database VIII: (a) preliminary background image obtained by median value from the sequence of images; (b) extracted candidate human area by binarization; (c) extracted human areas by labeling, size filtering and a morphological operation; and (d) the final background image.
Figure 6
Figure 6
Fuzzy system for the proposed method to extract adaptive threshold for ROI extraction.
Figure 7
Figure 7
Membership functions for fuzzy system to extract adaptive threshold for ROI extraction: (a) average value of the background image; (b) sum of difference values between background and input image; and (c) obtaining the output optimal threshold.
Figure 8
Figure 8
The first example for output of membership functions for fuzzy system: outputs by (a) F1 and (b) F2.
Figure 9
Figure 9
The first example for output optimal threshold based on the COG defuzzification method.
Figure 10
Figure 10
The second example for output of membership functions for fuzzy system: outputs by (a) F1 and (b) F2.
Figure 11
Figure 11
The second example for output optimal threshold based on the COG defuzzification method.
Figure 12
Figure 12
Example of a difference image: (a) input image; (b) background image; (c) difference image.
Figure 13
Figure 13
Example of a difference image: (a) input image; (b) background image; (c) difference image.
Figure 14
Figure 14
Separation of the candidate region within an input image based on the horizontal histogram: (a) input image and detected candidate region; (b) detected candidate region and its horizontal histogram; and (c) the division result of the candidate region.
Figure 15
Figure 15
Separation of the candidate region within an input image based on the vertical histogram: (a) input image and detected candidate region; (b) detected candidate region and its vertical histogram; and (c) the division result of the candidate region.
Figure 16
Figure 16
Separation of the candidate region within an input image based on the width, height, size and ratio: (a) input image and detected candidate region; (b) detected candidate region and its vertical histogram; and (c) the division result of the candidate region.
Figure 16
Figure 16
Separation of the candidate region within an input image based on the width, height, size and ratio: (a) input image and detected candidate region; (b) detected candidate region and its vertical histogram; and (c) the division result of the candidate region.
Figure 17
Figure 17
The examples of separation of one detected box into three (a,c) or four (b,d) ones by our method.
Figure 18
Figure 18
Separation of the candidate region within an input image: (a) input image and candidate human regions; (b) result of connecting separated regions.
Figure 19
Figure 19
Example of different sizes of human areas because of camera viewing direction.
Figure 20
Figure 20
Example of procedures for detecting human regions: (a) input image; (b) binarized image by background subtraction; (c) result of connecting separated candidate regions; and (d) Final result of detected human area.
Figure 21
Figure 21
Examples of databases: (a) database I; (b) database II; (c) database III; (d) database IV; (e) database V; (f) database VI; (g) database VII; (h) database VIII; (i) database IX; (j) database X; (k) database XI; (l) database XII; (m) database XIII; (n) database XIV; and (o) database XV.
Figure 21
Figure 21
Examples of databases: (a) database I; (b) database II; (c) database III; (d) database IV; (e) database V; (f) database VI; (g) database VII; (h) database VIII; (i) database IX; (j) database X; (k) database XI; (l) database XII; (m) database XIII; (n) database XIV; and (o) database XV.
Figure 22
Figure 22
Comparisons of preliminary background images with database I. The left figure (a) is by [33,34,35] and right figure (b) is by the proposed method, respectively.
Figure 23
Figure 23
Comparisons of created background images with database I. The left-upper, right-upper, left-lower and right-lower figures are by [24,27,28,29,30,31,33,34,35] and the proposed method, respectively.
Figure 24
Figure 24
Comparisons of created background images with database III. The left-upper, right-upper, left-lower and right-lower figures are by [24,27,28,29,30,31,33,34,35] and the proposed method, respectively.
Figure 25
Figure 25
Detection results with database (I–XV). Results of images in: (a) Database I; (b) Database II; (c) Database III; (d) Database IV; (e) Database V; (f) Database VI; (g) Database VII; (h) Database VIII; (i) Database IX; (j) Database X; (k) Database XI; (l) Database XII; (m) Database XIII; (n) Database XIV; and (o) Database XV.
Figure 26
Figure 26
Detection error cases: (a) result of the proposed method with database X; (b) result of the proposed method with database XIII.
See this image and copyright information in PMC

Similar articles

See all similar articles

Cited by

See all "Cited by" articles

References

    1. Ge J., Luo Y., Tei G. Real-Time Pedestrian Detection and Tracking at Nighttime for Driver-Assistance Systems. IEEE Trans. Intell. Transp. Syst. 2009;10:283–298.
    1. Prioletti A., Mogelmose A., Grisleri P., Trivedi M.M., Broggi A., Moeslund T.B. Part-Based Pedestrian Detection and Feature-Based Tracking for Driver Assistance: Real-Time, Robust Algorithms, and Evaluation. IEEE Trans. Intell. Transp. Syst. 2013;14:1346–1359. doi: 10.1109/TITS.2013.2262045. - DOI
    1. Källhammer J.-E. Night Vision: Requirements and Possible Roadmap for FIR and NIR Systems; Proceedings of the SPIE—The International Society for Optical Engineering; Strasbourg, France. 6 April 2006; p. 61980F.
    1. Mehralian S., Palhang M. Pedestrian Detection Using Principal Components Analysis of Gradient Distribution; Proceedings of the Iranian Conference on Machine Vision and Image Processing; Zanjan, Iran. 10–12 September 2013; pp. 58–63.
    1. Jiang Y., Ma J. Combination Features and Models for Human Detection; Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition; Boston, MA, USA. 7–12 June 2015; pp. 240–248.

Publication types

LinkOut - more resources

  NODES
INTERN 1
twitter 2