Fig 1.
Foreground masks of an image showing a honeybee on a carpet of flowers obtained using background subtraction.
The KNN background subtractor [48] was used to obtain foreground masks when the background is (a) constant; (b) wind-blown. Moving objects are shown in white pixels, the honeybee is circled.
Fig 2.
Hybrid Detection and Tracking (HyDaT) algorithm overview and components.
Fig 3.
Detecting an insect with background subtraction.
(a) Honeybee and flower shown at pixel resolution typical of that we employed for our study; (b) Binary image extracted using KNN background subtractor [48]; Resulting image with (c) median filter; (d) erosion-based morphological filter (centroid indicated).
Fig 4.
An example of an insect occluded under foliage.
Scatterplot shows the variation of insect visible body area before occlusion, and the corresponding least squares polynomial fit. Pixel intensity in the greyscale image represents the amount of change detected in the foreground.
Fig 5.
Experimental setup for recording videos.
Fig 6.
Number of image region changes per frame in test videos.
Box plot showing the distribution of number of image regions with greater than one pixel change per frame in test videos. The filled red diamond indicates the mean number of region changes per frame.
Fig 7.
Trajectories for a single honeybee in test videos.
Tracks were extracted using HyDaT from seven test video files.
Table 1.
A quantitative comparison of HyDaTs’ tracking performance against a stand-alone deep learning-based model (YOLOv2) [49] of honeybees foraging in Scaevola.
Table 2.
Occlusion detection algorithm performance and field of view (FoV) exit estimate for an 8:15 minute video of honeybees recorded in Scaevola.
Fig 8.
HyDaT algorithm tracking honeybee movement.
(a) Scaevola and (b) Lamb’s-ear. Red indicates recorded positions.
Fig 9.
Data analysis of honeybees foraging in Scaevola (N = 47) and Lamb’s-ear (N = 90).
(a) Honeybee trajectories, (b) Location heat-maps, and (c) Visibility duration for Scaevola and Lamb’s-ear. Honeybee (d) Speed distribution, (e) turn-angle distribution in Scaevola. In (b) the heat-map scale shows the aggregate of durations honeybees spent in a region. Bin size of the heat-map is the average area covered by a honeybee in pixels. In (c) recorded time is divided into durations the honeybee spends on the flower carpet (visible), under the carpet (occluded), and un-estimated, based on the output of the occlusion identification algorithm. The red dashed line shows the mean foraging time of honeybees within the field of view of the camera.