scholarly journals Automatic Camera-Trap Classification Using Wildlife-Specific Deep Learning in Nilgai Management

2021 ◽  
Author(s):  
Matthew Kutugata ◽  
Jeremy Baumgardt ◽  
John A. Goolsby ◽  
Alexis E. Racelis
Keyword(s):  
Deep Learning ◽  
Wildlife Conservation ◽  
Low Cost ◽  
Camera Trap ◽  
Camera Traps ◽  
Small Scale ◽  
Similar Species ◽  
Conservation Projects ◽  
Methodological Considerations ◽  
High Degree

Abstract Camera traps provide a low-cost approach to collect data and monitor wildlife across large scales but hand-labeling images at a rate that outpaces accumulation is difficult. Deep learning, a subdiscipline of machine learning and computer science, can address the issue of automatically classifying camera-trap images with a high degree of accuracy. This technique, however, may be less accessible to ecologists or small-scale conservation projects, and has serious limitations. In this study, we trained a simple deep learning model using a dataset of 120,000 images to identify the presence of nilgai Boselaphus tragocamelus, a regionally specific nonnative game animal, in camera-trap images with an overall accuracy of 97%. We trained a second model to identify 20 groups of animals and one group of images without any animals present, labeled as “none,” with an accuracy of 89%. Lastly, we tested the multigroup model on images collected of similar species, but in the southwestern United States, resulting in significantly lower precision and recall for each group. This study highlights the potential of deep learning for automating camera-trap image processing workflows, provides a brief overview of image-based deep learning, and discusses the often-understated limitations and methodological considerations in the context of wildlife conservation and species monitoring.

Deep neural networks for animal object detection and recognition in the wild

2019 ◽  
Author(s):  
◽  
Hayder Yousif
Keyword(s):  
Deep Learning ◽  
Animal Species ◽  
Object Segmentation ◽  
Camera Trap ◽  
Background Modeling ◽  
Camera Traps ◽  
Natural Scene ◽  
Feature Maps ◽  
Object Proposals ◽  
Animal Detection

[ACCESS RESTRICTED TO THE UNIVERSITY OF MISSOURI AT REQUEST OF AUTHOR.] 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 images for each detection, and are deployed in arrays of dozens or hundreds of sites, often resulting in millions of images per study. The task of converting images 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 offer the first widely available computer vision tool for processing camera trap images. Our results show that the tool is accurate and results in substantial time savings for processing large image datasets, thus improving our ability to monitor wildlife across large scales with camera traps. In this dissertation, we have developed new image/video processing and computer vision algorithms for efficient and accurate object detection and sequence-level classiffication from natural scene camera-trap images. This work addresses the following five major tasks: (1) Human-animal detection. We develop a fast and accurate scheme for human-animal detection from highly cluttered camera-trap images using joint background modeling and deep learning classification. Specifically, first, We develop an effective background modeling and subtraction scheme to generate region proposals for the foreground objects. We then develop a cross-frame image patch verification to reduce the number of foreground object proposals. Finally, We perform complexity-accuracy analysis of deep convolutional neural networks (DCNN) to develop a fast deep learning classification scheme to classify these region proposals into three categories: human, animals, and background patches. The optimized DCNN is able to maintain high level of accuracy while reducing the computational complexity by 14 times. Our experimental results demonstrate that the proposed method outperforms existing methods on the camera-trap dataset. (2) Object segmentation from natural scene. We first design and train a fast DCNN for animal-human-background object classification, which is used to analyze the input image to generate multi-layer feature maps, representing the responses of different image regions to the animal-human-background classifier. From these feature maps, we construct the so-called deep objectness graph for accurate animal-human object segmentation with graph cut. The segmented object regions from each image in the sequence are then verfied and fused in the temporal domain using background modeling. Our experimental results demonstrate that our proposed method outperforms existing state-of-the-art methods on the camera-trap dataset with highly cluttered natural scenes. (3) DCNN domain background modeling. We replaced the background model with a new more efficient deep learning based model. The input frames are segmented into regions through the deep objectness graph then the region boundaries of the input frames are multiplied by each other to obtain the regions of movement patches. We construct the background representation using the temporal information of the co-located patches. We propose to fuse the subtraction and foreground/background pixel classiffcation of two representation : a) chromaticity and b) deep pixel information. (4) Sequence-level object classiffcation. We proposed a new method for sequence-level video recognition with application to animal species recognition from camera trap images. First, using background modeling and cross-frame patch verification, we developed a scheme to generate candidate object regions or object proposals in the spatiotemporal domain. Second, we develop a dynamic programming optimization approach to identify the best temporal subset of object proposals. Third, we aggregate and fuse the features of these selected object proposals for efficient sequence-level animal species classification.

scholarly journals Fauna diversity in the southern part of the Kon Ka Kinh National Park, Gia Lai province

2021 ◽  
Vol 43 (4) ◽  
pp. 139-151
Author(s):  
Nguyen Ai Tam ◽  
Nguyen Van Tay ◽  
Nguyen Thi Kim Yen ◽  
Ha Thang Long
Keyword(s):  
National Park ◽  
Wildlife Conservation ◽  
Animal Species ◽  
Infrared Camera ◽  
Camera Trap ◽  
Iucn Red List ◽  
Camera Traps ◽  
Survey Results ◽  
Set Up ◽  
First Time

Kon Ka Kinh National Park (KKK NP) is a priority zone for biodiversity protection in Vietnam as well as ASEAN. In order to survey the current fauna species diversity in the southern part of the KKK NP, we conducted camera trapping surveys in 2017, 2018, and 2019. 28 infrared camera traps were set up on elevations between 1041 to 1497 meters. In total, there were 360 days of survey using camera trap. As result, we recorded a total of 27 animal species of those, five species are listed in the IUCN Red List of Threatened Species (IUCN, 2020). The survey results showed a high richness of wildlife in the southern park region, and it also revealed human disturbance to wildlife in the park. The first-time camera trap was used for surveying wildlife diversity in the southern region of the KKK NP. Conducting camera trap surveys in the whole KKK NP is essential for monitoring and identifying priority areas for wildlife conservation in the national park.

scholarly journals Deep learning enabled smart mats as a scalable floor monitoring system

2020 ◽  
Vol 11 (1) ◽  
Author(s):  
Qiongfeng Shi ◽  
Zixuan Zhang ◽  
Tianyiyi He ◽  
Zhongda Sun ◽  
Bingjie Wang ◽  
...  
Keyword(s):  
Deep Learning ◽  
Monitoring System ◽  
Data Analytics ◽  
Low Cost ◽  
Computational Cost ◽  
Sensory Information ◽  
Small Scale ◽  
Large Power ◽  
Smart Building ◽  
Sensory Data

Abstract Toward smart building and smart home, floor as one of our most frequently interactive interfaces can be implemented with embedded sensors to extract abundant sensory information without the video-taken concerns. Yet the previously developed floor sensors are normally of small scale, high implementation cost, large power consumption, and complicated device configuration. Here we show a smart floor monitoring system through the integration of self-powered triboelectric floor mats and deep learning-based data analytics. The floor mats are fabricated with unique “identity” electrode patterns using a low-cost and highly scalable screen printing technique, enabling a parallel connection to reduce the system complexity and the deep-learning computational cost. The stepping position, activity status, and identity information can be determined according to the instant sensory data analytics. This developed smart floor technology can establish the foundation using floor as the functional interface for diverse applications in smart building/home, e.g., intelligent automation, healthcare, and security.

scholarly journals An IoT System Using Deep Learning to Classify Camera Trap Images on the Edge

Computers ◽  
2022 ◽  
Vol 11 (1) ◽  
pp. 13
Author(s):  
Imran Zualkernan ◽  
Salam Dhou ◽  
Jacky Judas ◽  
Ali Reza Sajun ◽  
Brylle Ryan Gomez ◽  
...  
Keyword(s):  
Deep Learning ◽  
Camera Trap ◽  
Mobile App ◽  
Camera Traps ◽  
Raspberry Pi ◽  
Adjusted Rand Index ◽  
Area Network ◽  
Wide Area Network ◽  
Neural Network Models ◽  
Average Current

Camera traps deployed in remote locations provide an effective method for ecologists to monitor and study wildlife in a non-invasive way. However, current camera traps suffer from two problems. First, the images are manually classified and counted, which is expensive. Second, due to manual coding, the results are often stale by the time they get to the ecologists. Using the Internet of Things (IoT) combined with deep learning represents a good solution for both these problems, as the images can be classified automatically, and the results immediately made available to ecologists. This paper proposes an IoT architecture that uses deep learning on edge devices to convey animal classification results to a mobile app using the LoRaWAN low-power, wide-area network. The primary goal of the proposed approach is to reduce the cost of the wildlife monitoring process for ecologists, and to provide real-time animal sightings data from the camera traps in the field. Camera trap image data consisting of 66,400 images were used to train the InceptionV3, MobileNetV2, ResNet18, EfficientNetB1, DenseNet121, and Xception neural network models. While performance of the trained models was statistically different (Kruskal–Wallis: Accuracy H(5) = 22.34, p < 0.05; F1-score H(5) = 13.82, p = 0.0168), there was only a 3% difference in the F1-score between the worst (MobileNet V2) and the best model (Xception). Moreover, the models made similar errors (Adjusted Rand Index (ARI) > 0.88 and Adjusted Mutual Information (AMU) > 0.82). Subsequently, the best model, Xception (Accuracy = 96.1%; F1-score = 0.87; F1-Score = 0.97 with oversampling), was optimized and deployed on the Raspberry Pi, Google Coral, and Nvidia Jetson edge devices using both TenorFlow Lite and TensorRT frameworks. Optimizing the models to run on edge devices reduced the average macro F1-Score to 0.7, and adversely affected the minority classes, reducing their F1-score to as low as 0.18. Upon stress testing, by processing 1000 images consecutively, Jetson Nano, running a TensorRT model, outperformed others with a latency of 0.276 s/image (s.d. = 0.002) while consuming an average current of 1665.21 mA. Raspberry Pi consumed the least average current (838.99 mA) with a ten times worse latency of 2.83 s/image (s.d. = 0.036). Nano was the only reasonable option as an edge device because it could capture most animals whose maximum speeds were below 80 km/h, including goats, lions, ostriches, etc. While the proposed architecture is viable, unbalanced data remain a challenge and the results can potentially be improved by using object detection to reduce imbalances and by exploring semi-supervised learning.

scholarly journals How many giraffes are there? A comparison of abundance estimators at Ongava Game Reserve, Namibia

2021 ◽  
Author(s):  
Christophe Bonenfant ◽  
Ken Stratford ◽  
Stephanie Periquet
Keyword(s):  
Population Size ◽  
Detection Probability ◽  
Wildlife Conservation ◽  
Camera Trap ◽  
Biological Data ◽  
Camera Traps ◽  
Population Abundance ◽  
Visit Frequency ◽  
Game Reserve ◽  
Capture Recapture

Camera-traps are a versatile and widely adopted tool to collect biological data in wildlife conservation and management. If estimating population abundance from camera-trap data is the primarily goal of many projects, what population estimator is suitable for such data needs to be investigated. We took advantage of a 21 days camera-trap monitoring on giraffes at Onvaga Game Reserve, Namibia to compare capture-recapture (CR), saturation curves and N-mixture estimators of population abundance. A marked variation in detection probability of giraffes was observed in time and between individuals. Giraffes were also less likely to be detected after they were seen at a waterhole with cameras (visit frequency of f = 0.25). We estimated population size to 119 giraffes with a Cv = 0.10 with the best CR estimator. All other estimators we a applied over-estimated population size by ca. -20 to >+80%, because they did not account for the main sources of heterogeneity in detection probability. We found that modelling choices was much less forgiving for N-mixture than CR estimators. Double counts were problematic for N-mixture models, challenging the use of raw counts at waterholes to monitor giraffes abundance.

scholarly journals MAGI: Multistream Aerial Segmentation of Ground Images with Small-Scale Drones

Drones ◽  
2021 ◽  
Vol 5 (4) ◽  
pp. 111
Author(s):  
Danilo Avola ◽  
Daniele Pannone
Keyword(s):  
Deep Learning ◽  
Precision Agriculture ◽  
Low Cost ◽  
Semantic Segmentation ◽  
Search And Rescue ◽  
Small Scale ◽  
Border Control ◽  
Control Precision ◽  
Challenging Tasks ◽  
Deep Learning Model

In recent years, small-scale drones have been used in heterogeneous tasks, such as border control, precision agriculture, and search and rescue. This is mainly due to their small size that allows for easy deployment, their low cost, and their increasing computing capability. The latter aspect allows for researchers and industries to develop complex machine- and deep-learning algorithms for several challenging tasks, such as object classification, object detection, and segmentation. Focusing on segmentation, this paper proposes a novel deep-learning model for semantic segmentation. The model follows a fully convolutional multistream approach to perform segmentation on different image scales. Several streams perform convolutions by exploiting kernels of different sizes, making segmentation tasks robust to flight altitude changes. Extensive experiments were performed on the UAV Mosaicking and Change Detection (UMCD) dataset, highlighting the effectiveness of the proposed method.

scholarly journals Automatically identifying, counting, and describing wild animals in camera-trap images with deep learning

2018 ◽  
Vol 115 (25) ◽  
pp. E5716-E5725 ◽  
Author(s):  
Mohammad Sadegh Norouzzadeh ◽  
Anh Nguyen ◽  
Margaret Kosmala ◽  
Alexandra Swanson ◽  
Meredith S. Palmer ◽  
...  
Keyword(s):  
Neural Networks ◽  
Deep Learning ◽  
Deep Neural Networks ◽  
Data Extraction ◽  
High Volume ◽  
Camera Trap ◽  
Camera Traps ◽  
Deep Convolutional Neural Networks ◽  
Manual Task ◽  
In The Wild

Having accurate, detailed, and up-to-date information about the location and behavior of animals in the wild would improve our ability to study and conserve ecosystems. We investigate the ability to automatically, accurately, and inexpensively collect such data, which could help catalyze the transformation of many fields of ecology, wildlife biology, zoology, conservation biology, and animal behavior into “big data” sciences. Motion-sensor “camera traps” enable collecting wildlife pictures inexpensively, unobtrusively, and frequently. However, extracting information from these pictures remains an expensive, time-consuming, manual task. We demonstrate that such information can be automatically extracted by deep learning, a cutting-edge type of artificial intelligence. We train deep convolutional neural networks to identify, count, and describe the behaviors of 48 species in the 3.2 million-image Snapshot Serengeti dataset. Our deep neural networks automatically identify animals with >93.8% accuracy, and we expect that number to improve rapidly in years to come. More importantly, if our system classifies only images it is confident about, our system can automate animal identification for 99.3% of the data while still performing at the same 96.6% accuracy as that of crowdsourced teams of human volunteers, saving >8.4 y (i.e., >17,000 h at 40 h/wk) of human labeling effort on this 3.2 million-image dataset. Those efficiency gains highlight the importance of using deep neural networks to automate data extraction from camera-trap images, reducing a roadblock for this widely used technology. Our results suggest that deep learning could enable the inexpensive, unobtrusive, high-volume, and even real-time collection of a wealth of information about vast numbers of animals in the wild.

scholarly journals A practical guide for estimating animal density using camera traps: Focus on the REST model

2021 ◽  
Author(s):  
Shun Hongo ◽  
Yoshihiro Nakashima ◽  
Gota Yajima
Keyword(s):  
Wildlife Conservation ◽  
Camera Trap ◽  
Camera Traps ◽  
Video Data ◽  
Camera Placement ◽  
Practical Guide ◽  
Animal Density ◽  
Forested Areas ◽  
Staying Time

Estimating animal density and finding the factors that influence it are central in wildlife conservation and management but challenging to achieve, particularly in forested areas. Camera trapping is a pervasive method in forest mammal survey and a plausible technique to overcome this challenge. This report provides a practical guide for conducting a camera trap survey to estimate the density of forest mammals applying the random encounter and staying time (REST) model. We firstly provide a brief explanation about the structure and assumptions of the REST model. Next, we describe essential points during different steps in planning a survey: determination of objectives, design of camera placement, choice of camera models, camera setting, the layout of the camera station, and list of covariates. We then develop detail-oriented instruction for conducting a survey and analysing the obtained video data. We encourage camera trap surveyors to provide the practised protocols of their surveys, which will be helpful to other camera trappers.

scholarly journals Deep learning models for the prediction of small-scale fisheries catches: finfish fishery in the region of “Bahía Magadalena-Almejas”

2018 ◽  
Vol 75 (6) ◽  
pp. 2088-2096 ◽  
Author(s):  
Ricardo Alberto Cavieses Núñez ◽  
Miguel Ángel Ojeda Ruiz de la Peña ◽  
Alfredo Flores Irigollen ◽  
Manuel Rodríguez Rodríguez
Keyword(s):  
Deep Learning ◽  
Short Term Memory ◽  
Low Cost ◽  
Small Scale ◽  
Learning Models ◽  
Small Scale Fisheries ◽  
Relevant Role ◽  
Important Challenge ◽  
Memory Network ◽  
Using Data

Abstract Globally, over 80% of fisheries are at maximum sustainable levels or overexploited. However, small-scale fisheries (SSFs) in developing countries play a relevant role in coastal communities’ development with important impacts on the economy. The SSFs are normally multi-specific and due to the lack of data, studying them by simulation poses an important challenge especially forecasting models. These models are necessary to support management decisions or develop sustainable fisheries; therefore, models based on Deep Learning were proposed to forecast SSFs catch, using data from official catch landing reports (OCLRs), satellite images, and oceanographic data. The finfish fishery in Bahía Magdalena-Almejas (México) was used for the present study. According to an analysis of OCLRs, the target species of major importance in the fishery were identified and selected for the model. The proposed deep learning models used two artificial neural networks structures: non-linear autoregressive neural network and long-short term memory network, which were designed to assess and forecast monthly catch levels of Paralabrax nebulifer and Caulolatilus princeps. Models with a performance efficiency of R &gt; 0.8, MSE &lt; 300 were found, which indicate that the models are applicable in SSF with poor data and multi-specific fishery contexts, at low cost.

scholarly journals Use of a novel camera trapping approach to measure small mammal responses to peatland restoration

2021 ◽  
Vol 67 (1) ◽  
Author(s):  
Nick A. Littlewood ◽  
Mark H. Hancock ◽  
Scott Newey ◽  
Gorm Shackelford ◽  
Rose Toney
Keyword(s):  
Animal Welfare ◽  
Small Mammals ◽  
Low Cost ◽  
Camera Trap ◽  
Camera Traps ◽  
Plantation Forestry ◽  
Survey Method ◽  
Flash Intensity ◽  
Peatland Restoration ◽  
Blanket Bog

AbstractSmall mammals, such as small rodents (Rodentia: Muroidea) and shrews (Insectivora: Soricidae), present particular challenges in camera trap surveys. Their size is often insufficient to trigger infra-red sensors, whilst resultant images may be of inadequate quality for species identification. The conventional survey method for small mammals, live-trapping, can be both labour-intensive and detrimental to animal welfare. Here, we describe a method for using camera traps for monitoring small mammals. We show that by attaching the camera trap to a baited tunnel, fixing a close-focus lens over the camera trap lens, and reducing the flash intensity, pictures or videos can be obtained of sufficient quality for identifying species. We demonstrate the use of the method by comparing occurrences of small mammals in a peatland landscape containing (i) plantation forestry (planted on drained former blanket bog), (ii) ex-forestry areas undergoing bog restoration, and (iii) unmodified blanket bog habitat. Rodents were detected only in forestry and restoration areas, whilst shrews were detected across all habitat. The odds of detecting small mammals were 7.6 times higher on camera traps set in plantation forestry than in unmodified bog, and 3.7 times higher on camera traps in restoration areas than in bog. When absolute abundance estimates are not required, and camera traps are available, this technique provides a low-cost survey method that is labour-efficient and has minimal animal welfare implications.

Sign in / Sign up

Export Citation Format

Share Document