Insect pest monitoring with camera-equipped traps: strengths and limitations

Preti, Michele; Verheggen, François; Angeli, Sergio

doi:10.1007/s10340-020-01309-4

Cited by 139 publications

(117 citation statements)

References 64 publications

(101 reference statements)

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…The latest high-performance CNN-based detection and classification models require a large number of images to optimize their numerous parameters. Only a few practical and large datasets currently exist for insect detection because setting many cameras to capture numerous insects in various field backgrounds is costly (Preti et al, 2020). Although we collected hundreds of flea beetle images, merely thirty im-516 ages were considered effective samples for each species.…”

Section: Discussionmentioning

confidence: 99%

“…Insects are some of the most abundant groups of animals 2 in the terrestrial ecosystem. Rapid and accurate identifica-3 tion of insect species is essential for pest management (Park et al, 2020;Preti et al, 2020) and species-level identifica-5 tion in biodiversity monitoring (Hogeweg et al, 2019;Høye 6 et al, 2021). To augment the expertise in insect taxonomy takimoto@c.oka-pu.ac.jp (H. Takimoto) ORCID(s): 0000-0002-8795-7109 (H. Takimoto); 0000-0002-6466-723X (Y. Sato); 0000-0001-7891-5049 (A.J.…”

Section: Introductionmentioning

confidence: 99%

“…geneous background with multiple species and individuals (Martineau et al, 2017). Furthermore, a snapshot or interval camera may not be sufficient to capture the key characteristics of species identification (Preti et al, 2020). A series of images, namely a video, may solve this issue by capturing multiple posing for an individual insect (Høye et al, 2021).…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Using a two-stage convolutional neural network to rapidly identify tiny herbivorous beetles in the field

Takimoto¹,

Sato

Nagano

et al. 2021

Preprint

View full text Add to dashboard Cite

Recently, deep convolutional neural networks (CNN) have been adopted to help beginners identify insect species from field images. However, the application of these methods on the identification of tiny congeneric species moving across heterogeneous background remains difficult. To enable rapid and automatic identification in the field, we customized a method involving real-time object detection of two Phyllotreta beetles. We first performed data augmentation using transformations, syntheses, and random erasing of the original images. We then proposed a two-stage method for the detection and identification of small insects based on CNN, where YOLOv4 and EfficientNet were used as a region proposal network and a re-identification method, respectively. Evaluation of the model revealed that one-step object detection by YOLOv4 alone was not precise (Precision = 0.55) when classifying two species of flea beetles and background objects. In contrast, the two-step CNNs improved the precision (Precision = 0.89) with moderate accuracy (F-measure = 0.55) and acceptable speed (ca. 5 frames per second for full HD images) of detection and identification of insect species in the field. Although real-time identification of tiny insects remains a challenge in the field, our method aids in improving small object detection on a heterogeneous background.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Using a two-stage convolutional neural network to rapidly identify tiny herbivorous beetles in the field

Takimoto¹,

Sato

Nagano

et al. 2021

Preprint

View full text Add to dashboard Cite

show abstract

“…Moreover, we selected the common solid food lures of protein baits for all the species of citrus fly pests, and they were processed as fine particles that were evenly adsorbed to the sticky surface of the yellow paper. The previous works [39,40] have proved that the incorporation of appropriate attractants into the surface of sticky traps will efficiently capture flies.…”

Section: Processing Of Yellow Sticky Trapmentioning

confidence: 99%

A Motor-Driven and Computer Vision-Based Intelligent E-Trap for Monitoring Citrus Flies

et al. 2021

View full text Add to dashboard Cite

Citrus flies are important quarantine pests in citrus plantations. Electronic traps (e-traps) based on computer vision are the most popular types of equipment for monitoring them. However, most current e-traps are inefficient and unreliable due to requiring manual operations and lack of reliable detection and identification algorithms of citrus fly images. To address these problems, this paper presents a monitoring scheme based on automatic e-traps and novel recognition algorithms. In this scheme, the prototype of an automatic motor-driven e-trap is firstly designed based on a yellow sticky trap. A motor autocontrol algorithm based on Local Binary Pattern (LBP) image analysis is proposed to automatically replace attractants in the e-trap for long-acting work. Furthermore, for efficient and reliable statistics of captured citrus flies, based on the differences between two successive sampling images of the e-trap, a simple and effective detection algorithm is presented to continuously detect the newly captured citrus flies from the collected images of the e-trap. Moreover, a Multi-Attention and Multi-Part convolutional neural Network (MAMPNet) is proposed to exploit discriminative local features of citrus fly images to recognize the citrus flies in the images. Finally, extensive simulation experiments validate the feasibility and efficiency of the designed e-trap prototype and its autocontrol algorithm, as well as the reliability and effectiveness of the proposed detection and recognition algorithms for citrus flies.

show abstract

“…This technique have gained popularity because it allows for non‐intrusive observations at large spatiotemporal scales and constant sampling effort (Rovero & Zimmermann, 2016; Wearn & Glover‐Kapfer, 2019). Recently, camera trapping of insects has become an active field of research and development but important technical limitations still persist (Pegoraro et al., 2020; Preti et al., 2021). First, although it has been reported that the passive thermal infrared motion sensor of commercial CT systems can be activated by large flying insects (Houlihan et al., 2019; Johnson & Raguso, 2016), most ectotherms, such as reptiles, amphibians and invertebrates, do not trigger motion sensors (Hobbs & Brehme, 2017).…”

Section: Introductionmentioning

confidence: 99%

PICT: A low‐cost, modular, open‐source camera trap system to study plant–insect interactions

et al. 2021

View full text Add to dashboard Cite

1. Commercial camera traps (CTs) commonly used in wildlife studies have several technical limitations that restrict their scope of application. They are not easily customizable, unit prices sharply increase with image quality and importantly, they are not designed to record the activity of ectotherms such as insects. Those developed for the study of plant-insect interactions are yet to be widely adopted as they rely on expensive and heavy equipment.2. We developed PICT (plant-insect interactions camera trap), an inexpensive (<100 USD) do-it-yourself CT system based on a Raspberry Pi Zero computer designed to continuously film animal activity. The system is particularly well suited for the study of pollination, insect behaviour and predator-prey interactions. The focus distance can be manually adjusted to under 5 cm. In low light conditions, a near-infrared light automatically illuminates the subject. Frame rate, resolution and video compression levels can be set by the user. The system can be remotely controlled using either a smartphone, tablet or laptop via the onboard Wi-Fi. PICT can record up to 72-hr day and night videos at >720p resolution with a 110-Wh power bank (30,000 mAh).Its ultra-portable (<1 kg) waterproof design and modular architecture is practical in diverse field settings. We provide an illustrated technical guide detailing the steps involved in building and operating a PICT and for video post-processing.3. We successfully field-tested PICT in a Central African rainforest in two contrasting research settings: an insect pollinator survey in the canopy of the African ebony Diospyros crassiflora and the observation of rare pollination events of an epiphytic orchid Cyrtorchis letouzeyi.4. PICT overcomes many of the limitations commonly associated with CT systems designed to monitor ectotherms. Increased portability and image quality at lower costs This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited.

show abstract

Insect pest monitoring with camera-equipped traps: strengths and limitations

Cited by 139 publications

References 64 publications

Using a two-stage convolutional neural network to rapidly identify tiny herbivorous beetles in the field

Using a two-stage convolutional neural network to rapidly identify tiny herbivorous beetles in the field

A Motor-Driven and Computer Vision-Based Intelligent E-Trap for Monitoring Citrus Flies

PICT: A low‐cost, modular, open‐source camera trap system to study plant–insect interactions

Contact Info

Product

Resources

About