Bugs and bytes: Entomological biomonitoring through the integration of deep learning and molecular analysis for merged community and network analysis DOI Creative Commons
Mukilan Deivarajan Suresh, Tong Xin, S. M. Cook

et al.

Agricultural and Forest Entomology, Journal Year: 2024, Volume and Issue: unknown

Published: Nov. 25, 2024

Abstract Insects play a vital role in ecosystem functioning, but some parts of the world, their populations have declined significantly recent decades due to environmental change, agricultural intensification and other anthropogenic drivers. Monitoring insect is crucial for understanding mitigating biodiversity loss, especially agro‐ecosystems where focus on pests beneficial insects gaining momentum context sustainable food systems. Biomonitoring has long played an important providing early warnings vectored pathogens assessing agro‐ecosystem management. However, identification invertebrates by taxonomists time‐consuming fraught with numerous challenges, particularly when it comes real‐time monitoring. Recent technological advances both computational image recognition molecular ecology enhanced biomonitoring efficiency accuracy, reducing labour efforts, leading unprecedented volumes data generated. This perspective article examines significance further potential deep learning image‐based recognition, aided complementary technologies, advancing entomological biomonitoring. Using sticky traps as example, we discuss each step workflow required create ecological networks using multimodal learning, identify challenges inherent this method survey techniques. In order become mainstream global biomonitoring, access long‐term, standardised comprehending dynamics, structure function population declines.

Language: Английский

ITF-WPI: Image and text based cross-modal feature fusion model for wolfberry pest recognition DOI
Guowei Dai, Jingchao Fan, Christine Dewi

et al.

Computers and Electronics in Agriculture, Journal Year: 2023, Volume and Issue: 212, P. 108129 - 108129

Published: Aug. 10, 2023

Language: Английский

Citations

33
A novel multi-label pest image classifier using the modified Swin Transformer and soft binary cross entropy loss DOI
Qingwen Guo, Chuntao Wang,

Deqin Xiao

et al.

Engineering Applications of Artificial Intelligence, Journal Year: 2023, Volume and Issue: 126, P. 107060 - 107060

Published: Sept. 6, 2023

Language: Английский

Citations

27
A Decision Support System for Crop Recommendation Using Machine Learning Classification Algorithms DOI Creative Commons
Murali Krishna Senapaty, Abhishek Ray,

Neelamadhab Padhy

et al.

Agriculture, Journal Year: 2024, Volume and Issue: 14(8), P. 1256 - 1256

Published: July 30, 2024

Today, crop suggestions and necessary guidance have become a regular need for farmer. Farmers generally depend on their local agriculture officers regarding this, it may be difficult to obtain the right at time. Nowadays, datasets are available different websites in sector, they play crucial role suggesting suitable crops. So, decision support system that analyzes dataset using machine learning techniques can assist farmers making better choices selections. The main objective of this research is provide quick with more accurate effective recommendations by utilizing methods, global positioning coordinates, cloud data. Here, recommendation personalized, which enables predict crops specific geographical context, taking into account factors like climate, soil composition, water availability, conditions. In regard, an existing historical contains state, district, year, area-wise production rate, name, season was collected 246,091 sample records from Dataworld website, holds data 37 areas India. Also, analysis, offices Rayagada, Koraput, Gajapati districts Odisha Both these were combined stored Firebase service. Thirteen algorithms been applied identify dependencies within To facilitate process, Android application developed Studio (Electric Eel | 2023.1.1) Emulator (Version 32.1.14), Software Development Kit (SDK, SDK 33), Tools. A model has proposed implements SMOTE (Synthetic Minority Oversampling Technique) balance dataset, then allows implementation 13 classifiers, such as logistic regression, tree (DT), K-Nearest Neighbor (KNN), SVC (Support Vector Classifier), random forest (RF), Gradient Boost (GB), Bagged Tree, extreme gradient boosting (XGB classifier), Ada Classifier, Cat Boost, HGB (Histogram-based Boosting), SGDC (Stochastic Descent), MNB (Multinomial Naive Bayes) dataset. It observed performance method 1.00 accuracy, precision, recall, F1-score, ROC AUC (Receiver Operating Characteristics–Area Under Curve) 0.91 sensitivity 0.54 specificity after applying SMOTE. Overall, compared all other classifiers implemented predictions.

Language: Английский

Citations

10
Unlocking the black box: an in-depth review on interpretability, explainability, and reliability in deep learning DOI
Emrullah Şahin, Naciye Nur Arslan, Durmuş Özdemir

et al.

Neural Computing and Applications, Journal Year: 2024, Volume and Issue: unknown

Published: Nov. 18, 2024

Language: Английский

Citations

10
A Spatial Analysis of Urban Streets under Deep Learning Based on Street View Imagery: Quantifying Perceptual and Elemental Perceptual Relationships DOI Open Access

Haozun Sun,

Hong Xu, Hao He

et al.

Sustainability, Journal Year: 2023, Volume and Issue: 15(20), P. 14798 - 14798

Published: Oct. 12, 2023

Measuring the human perception of urban street space and exploring elements that influence this have always interested geographic information planning fields. However, most traditional efforts to investigate are based on manual, usually time-consuming, inefficient, subjective judgments. This shortcoming has a crucial impact large-scale spatial analyses. Fortunately, in recent years, deep learning models gained robust element extraction capabilities for images achieved very competitive results semantic segmentation. In paper, we propose Street View imagery (SVI)-driven approach automatically measure six perceptions areas, including “safety”, “lively”, “beautiful”, “wealthy”, “depressing”, “boring”. The model was trained millions people’s ratings SVIs with high accuracy. First, paper maps distribution spaces within third ring road Wuhan (appearing as later). Secondly, constructed multiple linear regression “street constituents–human perception” by segmenting common constituents from SVIs. Finally, analyzed various objects positively or negatively correlated perceptual indicators model. experiments elucidated subtle weighting relationships between different dimensions they affect, helping identify visual factors may cause an area be involved. findings suggested motorized vehicles such “cars” “trucks” can affect which is previous studies. We also examined perceptions, “safety” “wealthy”. discussed “perceptual bias” issue cities. enhance understanding researchers city managers psychological cognitive processes behind human–street interactions.

Language: Английский

Citations

15
Pest Localization Using YOLOv5 and Classification Based on Quantum Convolutional Network DOI Creative Commons
Javeria Amin, Muhammad Almas Anjum,

Rida Zahra

et al.

Agriculture, Journal Year: 2023, Volume and Issue: 13(3), P. 662 - 662

Published: March 13, 2023

Pests are always the main source of field damage and severe crop output losses in agriculture. Currently, manually classifying counting pests is time consuming, enumeration population accuracy might be affected by a variety subjective measures. Additionally, due to pests’ various scales behaviors, current pest localization algorithms based on CNN unsuitable for effective management To overcome existing challenges, this study, method developed classification pests. For purposes, YOLOv5 trained using optimal learning hyperparameters which more accurately localize region plant images with 0.93 F1 scores. After localization, classified into Paddy pest/Paddy without proposed quantum machine model, consists fifteen layers two-qubit nodes. The network from scratch parameters that provide 99.9% accuracy. achieved results compared recent methods, performed same datasets prove novelty model.

Language: Английский

Citations

13
Different transfer learning approaches for insect pest classification in cotton DOI
Raul Toscano-Miranda, José Aguilar, William Hoyos

et al.

Applied Soft Computing, Journal Year: 2024, Volume and Issue: 153, P. 111283 - 111283

Published: Jan. 18, 2024

Language: Английский

Citations

6
Segmentation and Detection of Crop Pests using Novel U‐Net with Hybrid Deep Learning Mechanism DOI
Nagaveni Biradar,

Girisha Hosalli

Pest Management Science, Journal Year: 2024, Volume and Issue: 80(8), P. 3795 - 3807

Published: March 20, 2024

In India, agriculture is the backbone of economic sectors because increasing demand for agricultural products. However, production has been affected due to presence pests in crops. Several methods were developed solve crop pest detection issue, but they failed achieve better results. Therefore, proposed study used a new hybrid deep learning mechanism segmenting and detecting

Language: Английский

Citations

6
Can artificial intelligence be integrated into pest monitoring schemes to help achieve sustainable agriculture? An entomological, management and computational perspective DOI Creative Commons
Daniel J. Leybourne,

Nasamu Musa,

Po Yang

et al.

Agricultural and Forest Entomology, Journal Year: 2024, Volume and Issue: unknown

Published: May 16, 2024

Abstract Recent years have seen significant advances in artificial intelligence (AI) technology. This advancement has enabled the development of decision support systems that farmers with herbivorous pest identification and monitoring. In these systems, AI supports through detection, classification quantification pests. However, many under fall short meeting demands end user, shortfalls acting as obstacles impede integration into integrated management (IPM) practices. There are four common restrict uptake AI‐driven systems. Namely: technology effectiveness, functionality field conditions, level computational expertise power required to use run system mobility. We propose criteria need meet order overcome challenges: (i) The should be based on effective efficient AI; (ii) adaptable capable handling ‘real‐world’ image data collected from field; (iii) Systems user‐friendly, device‐driven low‐cost; (iv) mobile deployable multiple weather climate conditions. likely represent innovative transformative successfully integrate IPM principles tools can farmers.

Language: Английский

Citations

5
Application of a multi-layer convolutional neural network model to classify major insect pests in stored rice detected by an acoustic device DOI Creative Commons
Carlito Balingbing,

Sascha Kirchner,

Hubertus Siebald

et al.

Computers and Electronics in Agriculture, Journal Year: 2024, Volume and Issue: 225, P. 109297 - 109297

Published: Aug. 9, 2024

Language: Английский

Citations

5