Deep Learning and Computer Vision Techniques for Enhanced Quality Control in Manufacturing Processes

IEEE Access, Год журнала: 2024, Номер 12, С. 121449 - 121479

Опубликована: Янв. 1, 2024

Язык: Английский

---

Crop Type Identification Using High-Resolution Remote Sensing Images Based on an Improved DeepLabV3+ Network

Remote Sensing, Год журнала: 2023, Номер 15(21), С. 5088 - 5088

Опубликована: Окт. 24, 2023

Remote sensing technology has become a popular tool for crop classification, but it faces challenges in accurately identifying crops areas with fragmented land plots and complex planting structures. To address this issue, we propose an improved method identification high-resolution remote images, achieved by modifying the DeepLab V3+ semantic segmentation network. In paper, typical area Jianghuai watershed is taken as experimental area, Gaofen-2 satellite images high spatial resolutions are used data source. Based on original model, CI OSAVI vegetation indices added to input layers, MobileNet V2 backbone Meanwhile, upper sampling layer of network added, attention mechanism ASPP layers. The accuracy verification results shows that MIoU PA model test set reach 85.63% 95.30%, IoU F1_Score wheat 93.76% 96.78%, rape 74.24% 85.51%, respectively. significantly better than other related models. proposed paper can extract distribution information from images. This provides new technical approach application rape.

Язык: Английский

---

Deep learning for rapid crop damage assessment after cyclones

Natural Hazards, Год журнала: 2025, Номер unknown

Опубликована: Фев. 7, 2025

Язык: Английский

---

Applications of machine learning and deep learning in agriculture: A comprehensive review

Green Technologies and Sustainability, Год журнала: 2025, Номер unknown, С. 100199 - 100199

Опубликована: Март 1, 2025

Язык: Английский

---

Temporal Optimisation of Satellite Image‐Based Crop Mapping: A Comparison of Deep Time Series and Semi‐Supervised Time Warping Strategies

IET Computer Vision, Год журнала: 2025, Номер 19(1)

Опубликована: Янв. 1, 2025

ABSTRACT This study presents a novel approach to crop mapping using remotely sensed satellite images. It addresses the significant classification modelling challenges, including (1) requirements for extensive labelled data and (2) complex optimisation problem selection of appropriate temporal windows in absence prior knowledge cultivation calendars. We compare lightweight Dynamic Time Warping (DTW) method with heavily supervised Convolutional Neural Network ‐ Long Short‐Term Memory (CNN‐LSTM) high‐resolution multispectral optical imagery (3 m/pixel). Our integrates effective practical preprocessing steps, augmentation data‐driven strategy window, even presence numerous classes. findings demonstrate that DTW, despite its lower demands, can match performance CNN‐LSTM through our steps while significantly improving runtime. These results both DTW achieve deployment‐level accuracy underscore potential as viable alternative more resource‐intensive models. The also prove effectiveness windowing runtime study, no planting timeframes.

Язык: Английский

---

Fine-Scale (10 m) Dynamics of Smallholder Farming through COVID-19 in Eastern Thailand

Remote Sensing, Год журнала: 2024, Номер 16(6), С. 1035 - 1035

Опубликована: Март 14, 2024

This study aims to understand the spatiotemporal changes in patterns of tropical crop cultivation Eastern Thailand, encompassing periods before, during, and after COVID-19 pandemic. Our approach involved assessing efficacy high-resolution (10 m) Sentinel-2 dense image time series for mapping smallholder farmlands. We integrated harmonic regression random forest map a diverse array types between summer 2017 2023, including durian, rice, rubber, eucalyptus, oil palm, pineapple, sugarcane, cassava, mangosteen, coconut, other crops. The results revealed an overall accuracy 85.6%, with several exceeding 90%. High-resolution imagery demonstrated particular effectiveness situations involving intercropping, popular practice simultaneously growing two or more plant species same patch land. However, we observed overestimation majority studied cash crops, primarily those located young plantations open tree canopies grass-covered ground surfaces. adverse effects pandemic were specific labor-intensive rubber but limited short term. No discernible impact was noted across entirety timeframe. In comparison, financial gain climate change appeared be pivotal influencing farmers’ decisions regarding cultivation. Traditionally dominant crops such as rice palm have witnessed decline cultivation, reflecting decade-long trend price drops preceding Conversely, Thai durian has seen significant upswing even over pandemic, which ironically served catalyst prompting farmers adopt e-commerce meet surging demand, particularly from China.

Язык: Английский

---

Attention Mechanism-Combined LSTM for Grain Yield Prediction in China Using Multi-Source Satellite Imagery

Sustainability, Год журнала: 2023, Номер 15(12), С. 9210 - 9210

Опубликована: Июнь 7, 2023

Grain yield prediction affects policy making in various aspects such as agricultural production planning, food security assurance, and adjustment of foreign trade. Accurately predicting grain is great significance ensuring global security. This paper based on the MODIS remote sensing image data products from 2010 to 2020, adds band information vegetation index temperature form composite a dataset. Aiming at lack models for large-scale forecasting need human intervention traditional models, this proposes estimation model deep learning. First, cropping mapping techniques are used process generate training samples. Then channel spatial attention mechanism (convolutional block module, CBAM) added extract different bands improve efficiency model. Long short-term memory (LSTM) neural networks obtain feature time dimension. Finally, national-scale constructed. After study, it was found that LSTM using combination multi-source satellite images an can effectively predict China. Furthermore, proposed tested 2018 2020 showing average R2 0.940 RMSE 80,020 tons, indicating Chinese better. The extracts directly data, solves problem small-scale research imprecise end-to-end manner.

Язык: Английский

---

High-Resolution National-Scale Mapping of Paddy Rice Based on Sentinel-1/2 Data

Remote Sensing, Год журнала: 2023, Номер 15(16), С. 4055 - 4055

Опубликована: Авг. 16, 2023

Rice has always been one of the major food sources for human beings, and monitoring planning cultivation areas to maintain security achieve sustainable development is critical this crop. Traditional manual ground survey methods have recognized as being laborious, while remote-sensing technology can perform accurate mapping paddy rice due its unique data acquisition capabilities. The recently emerged Google Earth Engine (GEE) cloud-computing platform was found be capable storing computing resources required rapid processing massive quantities data, thereby revolutionizing traditional analysis patterns offering advantages large-scale crop mapping. Since phenology depends on local climatic conditions, considering vast expanse China with outstanding geospatial heterogeneity, a zoning strategy proposed in study separate monsoon climate zone into two regions based Qinling Mountain–Huaihe River Line (Q-H Line), discrepant basic algorithms adopted separately map mid-season nationwide. For northern regions, optical indices calculated Sentinel-2 images, growth spectral profiles constructed identify phenological periods, mapped using One-Class Support Vector Machine (OCSVM); southern microwave sequences Sentinel-1 Random Forest (RF). By applying methodological system, at 10 m spatial resolution GEE entire Chinese region 2021. According accuracy evaluation coefficients publicly released statistical yearbook relative error each province limited 10%, overall exceeded 85%. results could indicate that more accurately efficiently China-wide scale relatively few samples methods. adjusting parameters, time interval also further extended. powerful competence used large scale, help governments ascertain distribution across country short-term period, which would well suited meeting increasingly efficient fine-grained decision-making management requirements.

Язык: Английский

---

Optimizing crop classification in precision agriculture using AlexNet and high resolution UAV imagery

Technology in Agronomy, Год журнала: 2024, Номер 4(1), С. 0 - 0

Опубликована: Янв. 1, 2024

The rapid advancement of artificial intelligence, coupled with the utilization aerial images from Unmanned Aerial Vehicles (UAVs), presents a significant opportunity to enhance precision agriculture for crop classification. This is vital meet rising global food demand. In this study, effectiveness an 8-layer AlexNet, Convolutional Neural Network (CNN) variant was investigated automatic A DJI Mavic UAV employed capture high-resolution mixed-crop farm while adopting iterative training approach both AlexNet and conventional CNN model. Comparison based on performance done between these models across various epochs assess impact model's performance. Findings study consistently demonstrated that outperformed throughout all epochs. achieved its highest at 60 epochs, validation accuracies 62.83% 46.98%, respectively. contrast, reached peak 99.25% 71.81% 50 but exhibited slight drop due overfitting. Remarkably, strong positive correlation AlexNet's observed, unlike in CNN. research also highlighted potential generalize classification accuracy datasets beyond domain, caution implement early stopping mechanisms prevent findings reinforces role deep learning remotely sensed data agriculture.

Язык: Английский

---

Attention Mechanism Combined Lstm for Grain Yield Prediction in China Using Multi-Source Satellite Imagery

Опубликована: Май 25, 2023

Grain yield prediction affects policy making in various aspects such as agricultural production planning, food security assurance, and adjustment of foreign trade. Accurately predicting grain is great significance ensuring global security. This paper based on the MODIS remote sensing image data products from 2010 to 2020, adds band information vegetation index temperature form composite a set. Aiming at lack models for large-scale forecasting need human intervention traditional models, this proposes estimation model deep learning. First, cropping mapping techniques are used process generate training samples. Then channel spatial attention mechanism (Convolutional Block Attention Module, CBAM) added extracting different bands improve efficiency model. Long Short-Term Memory (LSTM) neural networks also obtain feature time dimension. Finally, national-scale constructed. The proposed was tested 2018 2020 showing an average R2 0.940 RMSE 80,020 tons, indicating that it can predict Chinese better. extracts directly data, solves problem small-scale research imprecise end-to-end manner.

Язык: Английский

---