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
. 2023 Apr 7;13(4):e9903.
doi: 10.1002/ece3.9903. eCollection 2023 Apr.

Counting animals in aerial images with a density map estimation model

Yifei Qian  1 , Grant R W Humphries  2 , Philip N Trathan  3   4 , Andrew Lowther  5 , Carl R Donovan  1
Affiliations

Counting animals in aerial images with a density map estimation model

Yifei Qian et al. Ecol Evol. .

Abstract

Animal abundance estimation is increasingly based on drone or aerial survey photography. Manual postprocessing has been used extensively; however, volumes of such data are increasing, necessitating some level of automation, either for complete counting, or as a labour-saving tool. Any automated processing can be challenging when using such tools on species that nest in close formation such as Pygoscelis penguins. We present here a customized CNN-based density map estimation method for counting of penguins from low-resolution aerial photography. Our model, an indirect regression algorithm, performed significantly better in terms of counting accuracy than standard detection algorithm (Faster-RCNN) when counting small objects from low-resolution images and gave an error rate of only 0.8 percent. Density map estimation methods as demonstrated here can vastly improve our ability to count animals in tight aggregations and demonstrably improve monitoring efforts from aerial imagery.

Keywords: abundance estimation; density map estimation; image processing; machine learning.

PubMed Disclaimer

Conflict of interest statement

The authors declare no competing interests.

Figures

FIGURE 1
FIGURE 1
Selected data samples with (a) small object size, (b) almost indistinguishable fore/background objects, and (c) varying illuminations are shown. The study object, penguins, is marked with red dots.
FIGURE 2
FIGURE 2
Left is a random image (penguins are labeled with red dots) picked from the dataset and its corresponding density map is on the right.
FIGURE 3
FIGURE 3
This figure shows the overall structure of our density map estimation model. The backbone extracts features from the input image, and these intermediate features are further fed to two branches to predict density map and segmentation map.
FIGURE 4
FIGURE 4
Some visualization results of the estimated density maps. The three images in each row, from left to right is the input, the Gaussian‐smoothed ground‐truth density map and the prediction. The corresponding count is given in the lower right corner of the density map. The difference between the ground‐truth and the estimated counts is highlighted.
FIGURE A1
FIGURE A1
Visual examples of the Faster‐RCNN model when setting the IoU threshold to 0.3 and the confidence threshold to 0.2. In the left image, the ground‐truth bounding boxes are red, and the prediction boxes are green. The corresponding detailed statistic is given on the right.
See this image and copyright information in PMC

References

    1. Afán, I. , Máñez, M. , & Díaz‐Delgado, R. (2018). Drone monitoring of breeding waterbird populations: The case of the glossy ibis. Drones, 2(4), 42. 10.3390/drones2040042 - DOI
    1. Buchholz, T.‐O. , Prakash, M. , Schmidt, D. , Krull, A. , & Jug, F. (2020). Denoiseg: Joint denoising and segmentation. ECCV 2020 Workshop on BioImage Computing.
    1. Burn, D. M. , Webber, M. A. , & Udevitz, M. S. (2006). Application of airborne thermal imagery to surveys of pacific walrus. Wildlife Society Bulletin, 34(1), 51–58. 10.2193/0091-7648(2006)34[51:AOATIT]2.0.CO;2 - DOI
    1. Butler, R. , & Muller‐Schwarze, D. (1977). Penguin census by aerial photographic analysis at cape crozier, ross Island. Antarctic Journal of the USA, 12, 25–27.
    1. Chabot, D. , & Bird, D. M. (2012). Evaluation of an off‐the‐shelf unmanned aircraft system for surveying flocks of geese.