Mobile Real-Time Grasshopper Detection and Data Aggregation Framework

Abstract

insects of the family Orthoptera: Acrididae including grasshoppers and locust devastate crops and ecosystems around the globe. The effective control of these insects requires large numbers of trained extension agents who try to spot concentrations of the insects on the ground so that they can be destroyed before they take flight. This is a challenging and difficult task. No automatic detection system is yet available to increase scouting productivity, data scale and fidelity. Here we demonstrate MAESTRO, a novel grasshopper detection framework that deploys deep learning within RBG images to detect insects. MAeStRo uses a state-of-the-art two-stage training deep learning approach. the framework can be deployed not only on desktop computers but also on edge devices without internet connection such as smartphones. MAeStRo can gather data using cloud storage for further research and in-depth analysis. In addition, we provide a challenging new open dataset (GHCID) of highly variable grasshopper populations imaged in inner Mongolia. the detection performance of the stationary method and the mobile App are 78 and 49 percent respectively; the stationary method requires around 1000 ms to analyze a single image, whereas the mobile app uses only around 400 ms per image. The algorithms are purely data-driven and can be used for other detection tasks in agriculture (e.g. plant disease detection) and beyond. This system can play a crucial role in the collection and analysis of data to enable more effective control of this critical global pest.

The sluggard has no locusts: from persistent pest to irresistible icon

“…If positive affirmation is the first step toward creation, then imagining locust damage may serve an important practical purpose: it is the wellspring of creative anti-locust technologies. Recent technologies include remote monitoring tools, computational models, citizen science cell phone apps, and a fungus (Metarhizium anisopliae)-derived toxin marketed under the name Green Muscle (Enserink, 2004;van Huis et al, 2007;Cressman, 2008;Chudzik et al, 2020). The importance of these tools is expected to increase in lockstep with increasing cyclone trends in the northern Indian Ocean (Murakami et al, 2020), suggesting that desert locusts will only continue to invade the fertile fields of our imagination as they transform into a metaphor for climate change.…”

High Throughput Data Acquisition and Deep Learning for Insect Ecoinformatics

“…Several recent studies extended these machine learning approaches to deal with basic and applied issues in insect ecology (Høye et al, 2021). DNN-based software for insect identification in images has been developed in the applicative context of agricultural pest control, aiming to support farmers and extension workers in identifying insect pests (e.g., Cheng et al, 2017;Nieuwenhuizen et al, 2018;Zhong et al, 2018;Liu et al, 2019;Chudzik et al, 2020;Roosjen et al, 2020). Other researchers developed deep learning models to assist insect identification for biodiversity monitoring projects (Hansen et al, 2020 for beetles;Buschbacher et al, 2020 for bees).…”

“…These include measuring the phenotypic similarity among Müllerian mimics in butterflies (Cuthill et al, 2019) and bees (Ezray et al, 2019), and exploring altitudinal trends in color variation of moths (Wu et al, 2019). The sources of images for most of these projects are either museum specimens (Cuthill et al, 2019;Hansen et al, 2020) or photos taken by field surveyors (Ezray et al, 2019;Buschbacher et al, 2020;Chudzik et al, 2020). These data sources have limited utility for broad scale surveys of insect populations in the wild, which are needed to facilitate both basic and applied studies of insect population dynamics of multiple taxa.…”

Sensor‐based outdoor monitoring of insects in arable crops for their precise control

“…The development of outdoor mobile sensors for the automatic detection of insects in arable crops is a challenging task and suffers from a serious challenges, including recognition of the tiny pest and classification of a great diversity of pests. Generally, sensor methods for target detection mostly work well on the laboratory scale but are not ready to efficiently and quickly record data outdoors at different observation scales 4 . In this paper, we review sensor‐based outdoor monitoring methods on different scales which have the potential for automated insect pest detection in arable crops with regard to precise pest control.…”

“…Generally, sensor methods for target detection mostly work well on the laboratory scale but are not ready to efficiently and quickly record data outdoors at different observation scales. 4 In this paper, we review sensor-based outdoor monitoring methods on different scales which have the potential for automated insect pest detection in arable crops with regard to precise pest control.…”