Finding areas of motion in camera trap images

Miguel, Agnieszka; Beery, Sara; Flores, Erica; Klemesrud, Loren; Bayrakçısmith, Rana

doi:10.1109/icip.2016.7532575

Cited by 18 publications

(17 citation statements)

References 8 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Cameras with non-parallel sensing directions (d i,j > 0, θ i,j > 0). We started from the Columbia Object Image Library (COIL-100) 2 and Amsterdam Library of Image Objects (ALOI) 3 datasets, which contain images of objects captured at 72 different poses obtained by rotating the object by 5 degrees each time. From each dataset, 6 camera pairs are selected, corresponding to the following angles between the camera sensing directions:…”

Section: Model Fittingmentioning

confidence: 99%

“…where η max is the maximum observed compression and is a parameter to be estimated. Note that the model proposed in equation (2) does not seem to be valid for all tested pixel resolutions: Figure 3 shows how the relationship between the CSA and the compression efficiency varies when increasing the images resolution (at a fixed JPEG quality factor of 10). For very small resolutions (Figure 3(a) and (b)), there is no visible relationship between the CSA and η: both the RMSE and the coefficient of determination R2 calculated after fitting the model in (2) to the CSA data are lower than the ones computed with a simple linear fitting.…”

Section: Model Fittingmentioning

confidence: 99%

“…Being able to gather, process and analyse visual data, VSNs constitute a key technology to implement several monitoring and analysis applications in remote and inaccessible scenarios. As an example, in the field of environmental monitoring, battery-operated wireless cameras can be employed to keep track of the changes in the characteristics of the vegetation or other environmental features over time [1]; in the fields of ecology, conservation biology and wildlife monitoring, VSNs are deployed in remote locations and coupled with additional sensors (e.g., motion or infrared) to capture and possibly analyze images of wild animals [2]. VSNs find application also in the context of Smart Cities: although it is reasonable to assume that most of the cameras deployed in urban environments do not run on batteries and have high-bandwidth communication capabilities, the possibility to quickly deploy many wireless-capable battery-operated cameras opens the way to the implementation of several services (traffic and infrastructure monitoring, vacant parking lot detection, etc.)…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

A prediction-based approach for features aggregation in Visual Sensor Networks

et al. 2019

View full text Add to dashboard Cite

Visual Sensor Networks (VSNs) constitute a key technology for the implementation of several visual analysis tasks. Recent studies have demonstrated that such tasks can be efficiently performed following an operative paradigm where cameras transmit to a central controller local image features, rather than pixel-domain images. Furthermore, features from multiple camera views may be efficiently aggregated exploiting the spatial redundancy between overlapping views. In this paper we propose a routing protocol designed for supporting aggregation of image features in a VSN. First, we identify a predictor able to estimate the efficiency of local features aggregation between different cameras in a VSN. The proposed predictor is chosen so as to minimize the prediction error while keeping the network overhead cost low. Then, we harmonically integrate the proposed predictor in the Routing Protocol for Low-Power and Lossy Networks (RPL) in order to support the task of in-network feature aggregation. We propose a RPL objective function that takes into account the predicted aggregation efficiency and build the routes from the camera nodes to a central controller so that either energy consumption or used network bandwidth is minimized. Extensive experimental results confirm that the proposed approach can be used to increase the efficiency of VSNs.

show abstract

Section: Model Fittingmentioning

confidence: 99%

Section: Model Fittingmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

A prediction-based approach for features aggregation in Visual Sensor Networks

et al. 2019

View full text Add to dashboard Cite

show abstract

“…Analytical approaches include constructing a background model using the median pixel value (Miguel, Beery, Flores, Klemesrud, & Bayrakcismith, 2016), a non-parametric approach where the pixellevel background model is represented by a set of background samples (Barnich & Van Droogenbroeck, 2011) and robust principle component analysis (RPCA) (Candès, Li, Ma, & Wright, 2011). Unfortunately, the success of these efforts has been limited by producing large number of false positives and difficulty in distinguish between animal and human objects.…”

Section: Introductionmentioning

confidence: 99%

Animal Scanner: Software for classifying humans, animals, and empty frames in camera trap images

Yousif

Yuan

Kays

et al. 2019

Ecology and Evolution

View full text Add to dashboard Cite

Camera traps are a popular tool to sample animal populations because they are noninvasive, detect a variety of species, and can record many thousands of animal detections per deployment. Cameras are typically set to take bursts of multiple photographs for each detection and are deployed in arrays of dozens or hundreds of sites, often resulting in millions of photographs per study. The task of converting photographs to animal detection records from such large image collections is daunting, and made worse by situations that generate copious empty pictures from false triggers (e.g., camera malfunction or moving vegetation) or pictures of humans. We developed computer vision algorithms to detect and classify moving objects to aid the first step of camera trap image filtering—separating the animal detections from the empty frames and pictures of humans. Our new work couples foreground object segmentation through background subtraction with deep learning classification to provide a fast and accurate scheme for human–animal detection. We provide these programs as both Matlab GUI and command prompt developed with C++. The software reads folders of camera trap images and outputs images annotated with bounding boxes around moving objects and a text file summary of results. This software maintains high accuracy while reducing the execution time by 14 times. It takes about 6 seconds to process a sequence of ten frames (on a 2.6 GHZ CPU computer). For those cameras with excessive empty frames due to camera malfunction or blowing vegetation automatically removes 54% of the false‐triggers sequences without influencing the human/animal sequences. We achieve 99.58% on image‐level empty versus object classification of Serengeti dataset. We offer the first computer vision tool for processing camera trap images providing substantial time savings for processing large image datasets, thus improving our ability to monitor wildlife across large scales with camera traps.

show abstract

“…The wireless camera sensors can be placed without any constraints, as in WSN 1–3 . However, WVSN is confronted by several crucial challenges and the most serious challenges are energy and bandwidth constraints.…”

Section: Introductionmentioning

confidence: 99%

Energy conserving texture‐based adaptable compressive sensing scheme for WVSN

Lakshmi¹,

Gnanadurai²,

Muthulakshmi

2019

Concurrency and Computation

View full text Add to dashboard Cite

Summary Wireless Sensor Networks (WSN) have gained considerable research interest due to its wider applicability and efficiency. Due to the technological advancement, the WSN is multifaceted and one of the facets is the Wireless Visual Sensor Network (WVSN). The core functionality of WVSN is to handle multimedia data that involves data processing and transference, which, in turn, consumes more energy. Data transference requires more energy and the energy consumption is reduced by the concept of compressive sensing. Understanding the benefits of compressive sensing, this article proposes an adaptable compressive scheme that relies on the texture property of the video frames. The texture property is extracted by means of Local Directional Pattern (LDP). The performance of the proposed approach is evaluated in terms of image quality, time, and energy consumption. The proposed approach outperforms the comparative approaches with reasonable performance.

show abstract

Finding areas of motion in camera trap images

Cited by 18 publications

References 8 publications

A prediction-based approach for features aggregation in Visual Sensor Networks

A prediction-based approach for features aggregation in Visual Sensor Networks

Animal Scanner: Software for classifying humans, animals, and empty frames in camera trap images

Energy conserving texture‐based adaptable compressive sensing scheme for WVSN

Contact Info

Product

Resources

About