Automating sandhill crane counts from nocturnal thermal aerial imagery using deep learning
Abstract Population monitoring is essential to management and conservation efforts for migratory birds, but traditional low‐altitude aerial surveys with human observers are plagued by individual observer bias and risk to flight crews. Aerial surveys that use remote sensing can reduce bias and risk,...
Main Authors: | , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
Wiley
2023-04-01
|
Series: | Remote Sensing in Ecology and Conservation |
Subjects: | |
Online Access: | https://doi.org/10.1002/rse2.301 |
_version_ | 1797842311242055680 |
---|---|
author | Emilio Luz‐Ricca Kyle Landolt Bradley A. Pickens Mark Koneff |
author_facet | Emilio Luz‐Ricca Kyle Landolt Bradley A. Pickens Mark Koneff |
author_sort | Emilio Luz‐Ricca |
collection | DOAJ |
description | Abstract Population monitoring is essential to management and conservation efforts for migratory birds, but traditional low‐altitude aerial surveys with human observers are plagued by individual observer bias and risk to flight crews. Aerial surveys that use remote sensing can reduce bias and risk, but manual counting of wildlife in imagery is laborious and may be cost‐prohibitive. Therefore, automated methods for counting are critical to cost‐efficient application of remote sensing for wildlife surveys covering large areas. We conducted nocturnal surveys of sandhill cranes (Antigone canadensis) during spring migration in the Central Platte River Valley of Nebraska, USA, using midwave thermal infrared sensors. We developed a framework for automated counting of sandhill cranes from thermal imagery using deep learning, assessed and compared the performance of two automated counting models, and quantified the effect of spatial resolution on counting accuracy. Aerial thermal imagery data were collected in March 2018 and 2021; 40 images were analyzed. We applied two deep learning models: an object detection approach, Faster R‐CNN and a recently developed pixel‐density estimation approach, ASPDNet. Model performance was determined using data independent of the training imagery. The effect of spatial resolution was quantified with a beta regression on relative error. Our results showed model accuracy of 9% mean percent error for ASPDNet and 18% for Faster R‐CNN. Most error was related to the undercounting of sandhill cranes. ASPDNet had <50% of the error of Faster R‐CNN as measured by mean percent error, root‐mean‐squared error and mean absolute error. Spatial resolution affected accuracy of both models, with error rate increasing with coarser resolution, particularly with Faster R‐CNN. Deep learning models, particularly pixel‐density estimators, can accurately automate counting of migratory birds in a dense, aggregate setting such as nocturnal roosting sites. |
first_indexed | 2024-04-09T16:45:52Z |
format | Article |
id | doaj.art-7b8cba4962db458eb2189d8830632499 |
institution | Directory Open Access Journal |
issn | 2056-3485 |
language | English |
last_indexed | 2024-04-09T16:45:52Z |
publishDate | 2023-04-01 |
publisher | Wiley |
record_format | Article |
series | Remote Sensing in Ecology and Conservation |
spelling | doaj.art-7b8cba4962db458eb2189d88306324992023-04-22T17:18:04ZengWileyRemote Sensing in Ecology and Conservation2056-34852023-04-019218219410.1002/rse2.301Automating sandhill crane counts from nocturnal thermal aerial imagery using deep learningEmilio Luz‐Ricca0Kyle Landolt1Bradley A. Pickens2Mark Koneff3Institute for Integrative Conservation William & Mary 221 North Boundary Street Williamsburg Virginia 23185 USAUpper Midwest Environmental Sciences Center U.S. Geological Survey 2630 Fanta Reed Road La Crosse Wisconsin 54603 USADivision of Migratory Bird Management U.S. Fish and Wildlife Service 11510 American Holly Drive Laurel Maryland 20708 USABranch of Migratory Bird Surveys | Division of Migratory Bird Management U.S. Fish and Wildlife Service 69 Grove Street Extension Orono Maine 04469 USAAbstract Population monitoring is essential to management and conservation efforts for migratory birds, but traditional low‐altitude aerial surveys with human observers are plagued by individual observer bias and risk to flight crews. Aerial surveys that use remote sensing can reduce bias and risk, but manual counting of wildlife in imagery is laborious and may be cost‐prohibitive. Therefore, automated methods for counting are critical to cost‐efficient application of remote sensing for wildlife surveys covering large areas. We conducted nocturnal surveys of sandhill cranes (Antigone canadensis) during spring migration in the Central Platte River Valley of Nebraska, USA, using midwave thermal infrared sensors. We developed a framework for automated counting of sandhill cranes from thermal imagery using deep learning, assessed and compared the performance of two automated counting models, and quantified the effect of spatial resolution on counting accuracy. Aerial thermal imagery data were collected in March 2018 and 2021; 40 images were analyzed. We applied two deep learning models: an object detection approach, Faster R‐CNN and a recently developed pixel‐density estimation approach, ASPDNet. Model performance was determined using data independent of the training imagery. The effect of spatial resolution was quantified with a beta regression on relative error. Our results showed model accuracy of 9% mean percent error for ASPDNet and 18% for Faster R‐CNN. Most error was related to the undercounting of sandhill cranes. ASPDNet had <50% of the error of Faster R‐CNN as measured by mean percent error, root‐mean‐squared error and mean absolute error. Spatial resolution affected accuracy of both models, with error rate increasing with coarser resolution, particularly with Faster R‐CNN. Deep learning models, particularly pixel‐density estimators, can accurately automate counting of migratory birds in a dense, aggregate setting such as nocturnal roosting sites.https://doi.org/10.1002/rse2.301Computer visiondeep learningsandhill cranethermal imagerywildlife monitoring |
spellingShingle | Emilio Luz‐Ricca Kyle Landolt Bradley A. Pickens Mark Koneff Automating sandhill crane counts from nocturnal thermal aerial imagery using deep learning Remote Sensing in Ecology and Conservation Computer vision deep learning sandhill crane thermal imagery wildlife monitoring |
title | Automating sandhill crane counts from nocturnal thermal aerial imagery using deep learning |
title_full | Automating sandhill crane counts from nocturnal thermal aerial imagery using deep learning |
title_fullStr | Automating sandhill crane counts from nocturnal thermal aerial imagery using deep learning |
title_full_unstemmed | Automating sandhill crane counts from nocturnal thermal aerial imagery using deep learning |
title_short | Automating sandhill crane counts from nocturnal thermal aerial imagery using deep learning |
title_sort | automating sandhill crane counts from nocturnal thermal aerial imagery using deep learning |
topic | Computer vision deep learning sandhill crane thermal imagery wildlife monitoring |
url | https://doi.org/10.1002/rse2.301 |
work_keys_str_mv | AT emilioluzricca automatingsandhillcranecountsfromnocturnalthermalaerialimageryusingdeeplearning AT kylelandolt automatingsandhillcranecountsfromnocturnalthermalaerialimageryusingdeeplearning AT bradleyapickens automatingsandhillcranecountsfromnocturnalthermalaerialimageryusingdeeplearning AT markkoneff automatingsandhillcranecountsfromnocturnalthermalaerialimageryusingdeeplearning |