249
© Springer International Publishing AG 2017
R. Díaz-Delgado et al. (eds.), The Roles of Remote Sensing in Nature Conservation,
DOI 10.1007/978-3-319-64332-8_13
Improving the Accuracy of Bird Counts Using
Manual and Automated Counts in ImageJ:
An Open-Source Image Processing Program
Clive Hurford
Abstract Bird counts estimated ‘by eye’ are subject to high levels of observer variation. Comparison of bird numbers estimated by experienced ornithologists and
through manual counting in Image J indicated that observers typically underestimate by more than 30 %, with the median ranging from −13% to −57%. Image J is
an open source software package that provides both manual and automated options
for counting birds, whether on the ground or in flight. The manual approach has the
least preparation time, and will generate accurate counts of 1000 birds in only
c. 20 min. This technique is particularly useful for counting birds in breeding colonies, where the complexity of the background in an image will compromise the
accuracy of automated approaches. The manual counter provides markers for counting up to eight species simultaneously and is the preferred option for images containing up to 3000 birds. The automated counter is best suited for estimating bird
numbers when large groups occur against relatively plain backgrounds. However,
the automated counter will not differentiate between species and will typically
underestimate the number of birds in an image, as it is object-based and overlapping
individuals will count as one bird. Conversely, birds with strongly contrasting plumage patterns will be overestimated.”
Keywords Variation in observer bird count estimates • Object based image analysis (OBIA) • ImageJ • Open-source image processing software • Manual and automated bird counts
C. Hurford (*)
Natural Resources Wales, Cardiff, Wales, UK
e-mail: churford5@gmail.com
© Springer International Publishing AG 2017
R. Díaz-Delgado et al. (eds.), The Roles of Remote Sensing in Nature Conservation,
DOI 10.1007/978-3-319-64332-8_13
Improving the Accuracy of Bird Counts Using
Manual and Automated Counts in ImageJ:
An Open-Source Image Processing Program
Clive Hurford
Abstract Bird counts estimated ‘by eye’ are subject to high levels of observer variation. Comparison of bird numbers estimated by experienced ornithologists and
through manual counting in Image J indicated that observers typically underestimate by more than 30 %, with the median ranging from −13% to −57%. Image J is
an open source software package that provides both manual and automated options
for counting birds, whether on the ground or in flight. The manual approach has the
least preparation time, and will generate accurate counts of 1000 birds in only
c. 20 min. This technique is particularly useful for counting birds in breeding colonies, where the complexity of the background in an image will compromise the
accuracy of automated approaches. The manual counter provides markers for counting up to eight species simultaneously and is the preferred option for images containing up to 3000 birds. The automated counter is best suited for estimating bird
numbers when large groups occur against relatively plain backgrounds. However,
the automated counter will not differentiate between species and will typically
underestimate the number of birds in an image, as it is object-based and overlapping
individuals will count as one bird. Conversely, birds with strongly contrasting plumage patterns will be overestimated.”
Keywords Variation in observer bird count estimates • Object based image analysis (OBIA) • ImageJ • Open-source image processing software • Manual and automated bird counts
C. Hurford (*)
Natural Resources Wales, Cardiff, Wales, UK
e-mail: churford5@gmail.com