Object detection and tracking in Precision Farming: a systematic review

Computers and Electronics in Agriculture, Journal Year: 2024, Volume and Issue: 219, P. 108757 - 108757

Published: Feb. 23, 2024

Object Detection and Tracking have gained importance in recent years because of the great advances image video analysis techniques accurate results these technologies are producing. Moreover, they successfully been applied to multiple fields, including agricultural domain since offer real-time monitoring status crops animals while counting how many present within a field/barn. This review aims current literature on field Precision Farming. For that, over 300 research articles were explored, from which 150 last five systematically reviewed analysed regarding algorithms implemented, belong to, difficulties faced, limitations should be tackled future. Lastly, it examines potential issues that this approach might have, for instance, lack open-source datasets with labelled data. The findings study indicate critical enhance Farming pave way robotization sector provide insights crop animal management, optimize resource allocation. Future work focus optimal acquisition prior Tracking, along consideration biophysical environment farming scenarios.

Language: Английский

---

The neuroethology of ant navigation

Current Biology, Journal Year: 2025, Volume and Issue: 35(3), P. R110 - R124

Published: Feb. 1, 2025

Unlike any other group of animals, all ant species are social: individual ants share the food they gather with their nestmates and as a consequence must repeatedly leave nest to find then return home it. These back-and-forth foraging trips have been studied for about century much our growing understanding strategies underlying animal navigation has come from these studies. One important strategy that use keep track where on trip is 'path integration', in which continuously update 'home vector' gives estimated distance direction nest. As path integration accumulates errors, it cannot be relied bring precisely home: such precision accomplished by using views acquired before start foraging. Further learning scaffolded vectors or remembered vectors, guide route help useful experienced way. Many rely olfaction well vision guidance full details paths revealed how mix innate learnt multisensory cues. Wood ants, we focus this review, take an oscillating along pheromone trail sample odours, but acquire visual information only at peaks troughs oscillations. To provide working model neural basis multimodal navigational outline anatomy functioning major central brain areas circuits - complex, mushroom bodies lateral accessory lobes involved coordination behaviour olfactory patterns. Because brains not yet well-studied, work done notably, Drosophila, silkworm moths bees derive plausible circuitry can deliver ants' strategies.

Language: Английский

---

vmTracking enables highly accurate multi-animal pose tracking in crowded environments

PLoS Biology, Journal Year: 2025, Volume and Issue: 23(2), P. e3003002 - e3003002

Published: Feb. 10, 2025

In multi-animal tracking, addressing occlusion and crowding is crucial for accurate behavioral analysis. However, in situations where generate complex interactions, achieving pose tracking remains challenging. Therefore, we introduced virtual marker (vmTracking), which uses markers individual identification. Virtual are labels derived from conventional markerless tools, such as DeepLabCut (maDLC) Social LEAP Estimate Animal Poses (SLEAP). Unlike physical markers, exist only within the video attribute features to individuals, enabling consistent identification throughout entire while keeping animals reality. Using these cues, annotations were applied videos, was conducted with single-animal (saDLC) SLEAP’s method. vmTracking minimized manual corrections annotation frames needed training, efficiently tackling crowding. Experiments multiple mice, fish, human dancers confirmed vmTracking’s variability applicability. These findings could enhance precision reliability of methods used analysis naturalistic social behaviors animals, providing a simpler yet more effective solution.

Language: Английский

---

Goal Directed Movement in Insects

Current Opinion in Insect Science, Journal Year: 2025, Volume and Issue: unknown, P. 101374 - 101374

Published: April 1, 2025

Language: Английский

---

Neuromorphic sequence learning with an event camera on routes through vegetation

Science Robotics, Journal Year: 2023, Volume and Issue: 8(82)

Published: Sept. 13, 2023

For many robotics applications, it is desirable to have relatively low-power and efficient onboard solutions. We took inspiration from insects, such as ants, that are capable of learning following routes in complex natural environments using constrained sensory neural systems. Such capabilities particularly relevant applications agricultural robotics, where visual navigation through dense vegetation remains a challenging task. In this scenario, route likely high self-similarity be subject changing lighting conditions motion over uneven terrain, the effects wind on leaves increase variability input. used bioinspired event camera terrestrial robot collect sequences along outdoor applied algorithm for spatiotemporal memory closely based known circuit insect brain. show method plausible support recognition more robust than SeqSLAM when evaluated repeated runs same or with small lateral offsets. By encoding spiking network running neuromorphic computer, our model can evaluate familiarity real time footage.

Language: Английский

---

Investigating visual navigation using spiking neural network models of the insect mushroom bodies

Frontiers in Physiology, Journal Year: 2024, Volume and Issue: 15

Published: May 22, 2024

Ants are capable of learning long visually guided foraging routes with limited neural resources. The visual scene memory needed for this behaviour is mediated by the mushroom bodies; an insect brain region important and memory. In a navigation context, bodies theorised to act as familiarity detectors, guiding ants views that similar those previously learned when first travelling along route. Evidence from behavioural experiments, computational studies lesions all support idea. Here we further investigate role in spiking network model complex natural scenes. By implementing these networks GeNN–a library building GPU accelerated networks–we were able test models offline on image database representing through outdoor environment, also online embodied robot. body successfully learnt large series scenes (400 corresponding 27 m route) used memories choose accurate heading directions during route recapitulation both environments. Through analysing our model’s Kenyon cell (KC) activity, demonstrate KC activity directly related respective novelty input images. conducting parameter search found there non-linear dependence between optimal projection neuron (VPN) connection sparsity length time presented stimulus. showed training lower proportions generally produced better accuracy testing entire We comparator algorithms Quanser Q-car robot processing running Nvidia Jetson TX2. On 6.5 route, had mean distance (error) 0.144 ± 0.088 over 5 trials, which was performance comparable standard visual-only algorithms. Thus, have demonstrated biologically plausible ant can navigate environments simulation real world. Understanding basis will provide insight into how circuits tuned rapidly learn behaviourally relevant information inspiration creating bio-mimetic computer/robotic systems low energy requirements.

Language: Английский

---

Trail using ants follow idiosyncratic routes in complex landscapes

Learning & Behavior, Journal Year: 2023, Volume and Issue: 52(1), P. 105 - 113

Published: Nov. 22, 2023

Abstract A large volume of research on individually navigating ants has shown how path integration and visually guided navigation form a major part the ant toolkit for many species are sufficient mechanisms successful navigation. One behavioural markers interaction these is that experienced foragers develop idiosyncratic routes require individual have personal unique visual memories they use to guide habitual between nest feeding sites. The majority ants, however, inhabit complex cluttered environments social pheromone trails often collective recruitment, organisation foragers. We do not know interacts with behaviour along shared in natural environments. thus asked here if wood forage through densely woodlands where travel repeatedly follow same or choose spread paths within trail. recorded three long homing trajectories 20 their woodland habitat. found when highly landscapes. This shows rely route guidance even chemical trail information available. argue cues likely be dominant sensory modality routes. These experiments shed new light insects general, navigate multimodal

Language: Английский

---

Compensation to visual impairments and behavioral plasticity in navigating ants

Proceedings of the National Academy of Sciences, Journal Year: 2024, Volume and Issue: 121(48)

Published: Nov. 19, 2024

Desert ants are known to rely heavily on vision while venturing for food and returning the nest. During these foraging trips, memorize recognize their visual surroundings, which enables them recapitulate individually learned routes in a fast effective manner. The compound eyes crucial such navigation; however, it remains unclear how information from both integrated cope with impairment. Here, we manipulated ants’ system by covering one of two analyzed ability familiar views. Monocular showed an immediate disruption route. However, they were able compensate this nonnatural impairment few hours engaging extensive route-relearning ontogeny, composed more learning walks than what naïve typically do. This relearning process eye forms novel memories, without erasing previous memories acquired eyes. Additionally, having route only unable eyes, even though is available. Together, shows that encoded recalled egocentric fundamentally binocular way, where input as whole must be matched enable recognition. We show kind processing fits neural circuitry.

Language: Английский

---

vmTracking: Virtual Markers Overcome Occlusion and Crowding in Multi-Animal Pose Tracking

bioRxiv (Cold Spring Harbor Laboratory), Journal Year: 2024, Volume and Issue: unknown

Published: Feb. 7, 2024

Abstract In multi-animal tracking, addressing occlusion and crowding is crucial for accurate behavioral analysis. Consequently, we introduced Virtual Marker Tracking (vmTracking), which uses virtual markers individual identification. markers, created from traditional markerless pose tracking tools like DeepLabCut (maDLC) Social LEAP Estimate Animal Poses (SLEAP), attribute features to individuals, enabling consistent identification throughout the entire video without physical markers. Using these as cues, annotations were applied videos, was conducted with single-animal (saDLC) SLEAP’s method. vmTracking minimized manual corrections annotation frames needed training, efficiently tackling crowding. Experiments multiple mice, fish, human dancers confirmed vmTracking’s variability applicability. These findings could enhance precision reliability of methods used in analysis complex naturalistic social behaviors animals, providing a simpler yet more effective solution.

Language: Английский

---

The ethology of foraging in ants: revisiting Tinbergen’s four questions

Frontiers in Ethology, Journal Year: 2024, Volume and Issue: 3

Published: March 5, 2024

Since Tinbergen’s seminal contribution in 1963, ethology has blossomed as a multifaceted research field. Sixty years later, uncountable articles followed the four questions proposed necessary for understanding animal behaviour, and they culminated segmentation of subareas which communicate little among themselves. Foraging ants is one example where this division happened, despite clear need to integrate results obtained from different approaches. We chose subject revise literature, relating main relevant level explanation theoretical framework. Through such revision, we aim foster integration approaches bring light how can clarify understand foraging sixty after initial proposition.

Language: Английский

---