- 标题
- 摘要
- 关键词
- 实验方案
- 产品
-
Melamine Faced Panels Defect Classification beyond the Visible Spectrum
摘要： In this work, we explore the use of images from different spectral bands to classify defects in melamine faced panels, which could appear through the production process. Through experimental evaluation, we evaluate the use of images from the visible (VS), near-infrared (NIR), and long wavelength infrared (LWIR), to classify the defects using a feature descriptor learning approach together with a support vector machine classifier. Two descriptors were evaluated, Extended Local Binary Patterns (E-LBP) and SURF using a Bag of Words (BoW) representation. The evaluation was carried on with an image set obtained during this work, which contained five different defect categories that currently occurs in the industry. Results show that using images from beyond the visual spectrum helps to improve classification performance in contrast with a single visible spectrum solution.
关键词： machine learning,industrial application,infrared
更新于2025-09-10 09:29:36
-
[IEEE 2018 20th International Conference on Transparent Optical Networks (ICTON) - Bucharest (2018.7.1-2018.7.5)] 2018 20th International Conference on Transparent Optical Networks (ICTON) - Machine Learning Based Optimal Modulation Format Prediction for Physical Layer Network Planning
摘要： Physical layer network design and planning process is a cumbersome one. It includes laying out all possible combinations of modulation formats, fiber types, forward error correction codes, channel spacing, etc., conducting exhaustive simulations and lab experiments to come up with carefully tuned engineering rules, and finally using these approximate models to propose transmission feasibility. Besides being cumbersome, there are two fundamental issues in conventional network planning approach, firstly it almost exclusively offers conservative design, leading to resource underutilization, and secondly it’s not scalable – neither from planning viewpoint nor computationally – to next-generation highly granular and flexible networks. Machine learning, an artificial intelligence toolset, may be applied to solve aforementioned issues by allowing data-driven model development, and consequent transmission quality prediction. While network planning is an extensive topic, in this paper, we focus on neural network based modulation format classification, autonomously identifying best possible modulation format for a given link configuration.
关键词： machine learning,analytics,communication networks,optimization,optical fiber communications
更新于2025-09-10 09:29:36
-
[IEEE IGARSS 2018 - 2018 IEEE International Geoscience and Remote Sensing Symposium - Valencia (2018.7.22-2018.7.27)] IGARSS 2018 - 2018 IEEE International Geoscience and Remote Sensing Symposium - Automated Building Energy Consumption Estimation from Aerial Imagery
摘要： This paper presents a methodology for automatically estimating the energy consumption of buildings from aerial imagery using data from Gainesville, Florida. By detecting buildings in the imagery using convolutional neural networks and extracting features from those building annotations, we use only imagery-derived features to estimate building energy consumption using random forests regression. For individual buildings, we achieve a predictive R2 value of 0.26, and with spatial aggregation over an area of 400m×400m our predictive R2 value increases to 0.95. We also explore the sensitivity of these estimates to errors in the building estimation process. Our results indicate that information limited to the size and shape of buildings, provides substantial predictive potential for the energy consumption of buildings.
关键词： energy consumption,machine learning,building detection,aerial imagery
更新于2025-09-10 09:29:36
-
Ensemble Feature Selection for Plant Phenotyping: A Journey From Hyperspectral to Multispectral Imaging
摘要： Hyperspectral imaging is becoming an increasingly popular tool for high-throughput plant phenotyping, because it provides remarkable insights about the health status of plants. Feature selection is a key component in a hyperspectral image analysis, largely because a significant portion of spectral features are redundant and/or irrelevant, depending on the desired application. This paper presents an ensemble feature selection method to identify the most informative spectral features for practical applications in plant phenotyping. The hyperspectral data set contained the images of four wheat lines, each with a control and a salt (NaCl) treatment. To rank spectral features, six feature selection methods were used as the base for the ensemble: correlation-based feature selection, ReliefF, sequential feature selection, support vector machine-recursive feature elimination (SVM-RFE), LASSO logistic regression, and random forest. The best results were achieved by the ensemble of ReliefF, SVM-RFE, and random forest, which drastically reduced the dimension of the hyperspectral data set from 215 to 15 features, while improving the accuracy in classifying the salt-treated vegetation pixels from the control pixels by 8.5%. To transform the hyperspectral data set into a multispectral data set, six wavelengths as the center of broad multispectral bands around the most prominent features were determined by a clustering algorithm. The result of salt tolerance assessment of the four wheat lines using the derived multispectral data set was similar to that of the hyperspectral data set. This demonstrates that the proposed feature selection pipeline can be utilized for determining the most informative features and can be a valuable tool in the development of tailored multispectral cameras.
关键词： hyperspectral imaging,Band selection,multispectral imaging,wheat,ensemble feature selection,salt stress,machine learning,plant phenotyping,classification
更新于2025-09-10 09:29:36
-
[IEEE IGARSS 2018 - 2018 IEEE International Geoscience and Remote Sensing Symposium - Valencia, Spain (2018.7.22-2018.7.27)] IGARSS 2018 - 2018 IEEE International Geoscience and Remote Sensing Symposium - Urban Change Detection for Multispectral Earth Observation Using Convolutional Neural Networks
摘要： The Copernicus Sentinel-2 program now provides multispectral images at a global scale with a high revisit rate. In this paper we explore the usage of convolutional neural networks for urban change detection using such multispectral images. We first present the new change detection dataset that was used for training the proposed networks, which will be openly available to serve as a benchmark. The Onera Satellite Change Detection (OSCD) dataset is composed of pairs of multispectral aerial images, and the changes were manually annotated at pixel level. We then propose two architectures to detect changes, Siamese and Early Fusion, and compare the impact of using different numbers of spectral channels as inputs. These architectures are trained from scratch using the provided dataset.
关键词： convolutional neural networks,multispectral earth observation,Change detection,supervised machine learning
更新于2025-09-10 09:29:36
-
[IEEE IGARSS 2018 - 2018 IEEE International Geoscience and Remote Sensing Symposium - Valencia, Spain (2018.7.22-2018.7.27)] IGARSS 2018 - 2018 IEEE International Geoscience and Remote Sensing Symposium - Worldpop - Fusion of Earth and Big Data for Intraurban Population Mapping
摘要： High resolution estimates of human population distributions are very useful for large-scale or national scale analyses in many fields including epidemiology, healthcare, resource distribution, and development. Population densities have long been estimated using remote sensing data, particularly at large spatial scales. However, the accuracy of population density predictions can be very poor in cities, and this is particularly relevant in urban areas in sub-Saharan Africa. Here we map intra-urban population densities for select African cities by disaggregating census data using random forest techniques with remotely-sensed and geospatial data, including bespoke time-series intra-urban built-up data. We produce maps with up to 83% explained variance and find including built-up density layers in urban population models allows for clear improvements in prediction.
关键词： machine learning,population density,census,built-up,Urban areas,Africa
更新于2025-09-10 09:29:36
-
[IEEE IGARSS 2018 - 2018 IEEE International Geoscience and Remote Sensing Symposium - Valencia, Spain (2018.7.22-2018.7.27)] IGARSS 2018 - 2018 IEEE International Geoscience and Remote Sensing Symposium - Validation of Fine Resolution Land-Surface Energy Fluxes Derived with Combined Sentinel-2 and Sentinel-3 Observations
摘要： A methodology for deriving land-surface energy fluxes estimated with the use of Sentinel-2 and Sentinel-3 observations is validated in a savannah landscape in central Spain. The fluxes are derived at two spatial resolution: fine (20m) and coarse (around 1km). At both resolutions the thermal observations from Sea and Land Surface Temperature Radiometer (SLSTR) on Sentinel-3 and optical observations from Multi-Spectral Instrument (MSI) on Sentinel-2 are used within a Two-Source Energy Balance (TSEB) modelling scheme. For the fine resolution estimates, the thermal observations acquired by SLSTR at around 1km resolution are sharpened using high-resolution (20m) optical observations taken by MSI and a machine learning algorithm. The results indicate that it is possible to derive fluxes with similar accuracy at both spatial scales, while obtaining more detailed separation of fluxes originating from individual landscape features at the fine scale.
关键词： machine learning,Sentinel-2,land-surface energy fluxes,Sentinel-3,thermal sharpening
更新于2025-09-10 09:29:36
-
[IEEE IGARSS 2018 - 2018 IEEE International Geoscience and Remote Sensing Symposium - Valencia, Spain (2018.7.22-2018.7.27)] IGARSS 2018 - 2018 IEEE International Geoscience and Remote Sensing Symposium - Target Aspect Identification in SAR Image: A Machine Learning Approach
摘要： Identifying the aspect for a given target is an important issue in synthetic aperture radar (SAR) image interpretation. A new SAR target aspect identification method based on machine learning theory is proposed in this paper. First, the aspect angles of the SAR target are discretized, and the spatial relationships of the neighborhoods of the SAR target samples are established. Then an optimal linear mapping is solved based on the proposed subspace aspect discriminant analysis. The samples will be projected into a low-dimensional space and be of a better aspect identifiability than in their original space. Finally, the projected samples are fed into a multi-layer neural network, and the aspects of the SAR targets will be indicated. Experimental results have shown the superiority of the proposed method based on the moving and stationary target acquisition and recognition (MSTAR) dataset.
关键词： machine learning,Synthetic aperture radar,multi-layer neural network,target aspect identification
更新于2025-09-10 09:29:36
-
[IEEE 2017 International Renewable and Sustainable Energy Conference (IRSEC) - Tangier (2017.12.4-2017.12.7)] 2017 International Renewable and Sustainable Energy Conference (IRSEC) - Data Driven Model for Short Term PV Power Forecasting using Least Square Support Vector Regression
摘要： This paper presents an off-line model for forecasting photovoltaic power. This model is suitable to provide short-term forecasts without the need of Numerical Weather predictions data. This is interesting especially for power system operators as well as for individuals who do not have access to weather data and forecasts. In this paper we investigate the influence of an additional input parameter to the accuracy of an already tested and validated offline model. To rectify the performances of our models we will compare their performances with a usual persistent model. The results of simulation shows the benefits of adding this input to improve the accuracy of our PV forecasting model.
关键词： Photovoltaic Power,Forecasting,Least Square Support Vector Regression,Smart Grid,Grid Management,Machine Learning
更新于2025-09-10 09:29:36
-
Development and validation of an algorithm to predict the treatment modality of burn wounds using thermographic scans: Prospective cohort study
摘要： Background The clinical evaluation of a burn wound alone may not be adequate to predict the severity of the injury nor to guide clinical decision making. Infrared thermography provides information about soft tissue viability and has previously been used to assess burn depth. The objective of this study was to determine if temperature differences in burns assessed by infrared thermography could be used predict the treatment modality of either healing by re-epithelization, requiring skin grafts, or requiring amputations, and to validate the clinical predication algorithm in an independent cohort. Methods and findings Temperature difference (ΔT) between injured and healthy skin were recorded within the first three days after injury in previously healthy burn patients. After discharge, the treatment modality was categorized as re-epithelization, skin graft or amputation. Potential confounding factors were assessed through multiple linear regression models, and a prediction algorithm based on the ΔT was developed using a predictive model using a recursive partitioning Random Forest machine learning algorithm. Finally, the prediction accuracy of the algorithm was compared in the development cohort and an independent validation cohort. Significant differences were found in the ΔT between treatment modality groups. The developed algorithm correctly predicts into which treatment category the patient will fall with 85.35% accuracy. Agreement between predicted and actual treatment for both cohorts was weighted kappa 90%. Conclusion Infrared thermograms obtained at first contact with a wounded patient can be used to accurately predict the definitive treatment modality for burn patients. This method can be used to rationalize treatment and streamline early wound closure.
关键词： treatment modality,prediction algorithm,burn wounds,Random Forest machine learning,infrared thermography
更新于2025-09-10 09:29:36