Optimized intelligent algorithm for classifying cloud particles recorded by a Cloud Particle Imager

2021 ◽  
Author(s):  
Zepei Wu ◽  
Shuo Liu ◽  
Delong Zhao ◽  
Ling Yang ◽  
Zixin Xu ◽  
...  
Keyword(s):  
Processing Speed ◽  
Cloud Microphysics ◽  
Parameter Analysis ◽  
Ice Crystals ◽  
Support Vector ◽  
Svm Classifier ◽  
Ice Crystal ◽  
Cloud Particle ◽  
Average Accuracy ◽  
Cloud Particles

AbstractCloud particles have different shapes in the atmosphere. Research on cloud particle shapes plays an important role in analyzing the growth of ice crystals and the cloud microphysics. To achieve an accurate and efficient classification algorithm on ice crystal images, this study uses image-based morphological processing and principal component analysis, to extract features of images and apply intelligent classification algorithms for the Cloud Particle Imager (CPI). Currently, there are mainly two types of ice-crystal classification methods: one is the mode parameterization scheme, and the other is the artificial intelligence model. Combined with data feature extraction, the dataset was tested on ten types of classifiers, and the highest average accuracy was 99.07%. The fastest processing speed of the real-time data processing test was 2,000 images/s. In actual application, the algorithm should consider the processing speed, because the images are in the order of millions. Therefore, a support vector machine (SVM) classifier was used in this study. The SVM-based optimization algorithm can classify ice crystals into nine classes with an average accuracy of 95%, blurred frame accuracy of 100%, with a processing speed of 2,000 images/s. This method has a relatively high accuracy and faster classification processing speed than the classic neural network model. The new method could be also applied in physical parameter analysis of cloud microphysics.

scholarly journals Improving Machine Learning Identification of Unsafe Driver Behavior by Means of Sensor Fusion

2020 ◽  
Vol 10 (18) ◽  
pp. 6417 ◽  
Author(s):  
Emanuele Lattanzi ◽  
Giacomo Castellucci ◽  
Valerio Freschi
Keyword(s):  
Neural Network ◽  
Machine Learning ◽  
Driver Behavior ◽  
Ground Truth ◽  
Support Vector ◽  
Svm Classifier ◽  
Learning Technology ◽  
Average Accuracy ◽  
Unsafe Behaviors ◽  
Vehicle Sensors

Most road accidents occur due to human fatigue, inattention, or drowsiness. Recently, machine learning technology has been successfully applied to identifying driving styles and recognizing unsafe behaviors starting from in-vehicle sensors signals such as vehicle and engine speed, throttle position, and engine load. In this work, we investigated the fusion of different external sensors, such as a gyroscope and a magnetometer, with in-vehicle sensors, to increase machine learning identification of unsafe driver behavior. Starting from those signals, we computed a set of features capable to accurately describe the behavior of the driver. A support vector machine and an artificial neural network were then trained and tested using several features calculated over more than 200 km of travel. The ground truth used to evaluate classification performances was obtained by means of an objective methodology based on the relationship between speed, and lateral and longitudinal acceleration of the vehicle. The classification results showed an average accuracy of about 88% using the SVM classifier and of about 90% using the neural network demonstrating the potential capability of the proposed methodology to identify unsafe driver behaviors.

An Enhanced Facial Expression Recognition Model Using Local Feature Fusion of Gabor Wavelets and Local Directionality Patterns

2020 ◽  
Vol 11 (1) ◽  
pp. 48-70 ◽  
Author(s):  
Sivaiah Bellamkonda ◽  
Gopalan N.P
Keyword(s):  
Facial Expression ◽  
Feature Fusion ◽  
Support Vector ◽  
Svm Classifier ◽  
Source Image ◽  
Expression Recognition ◽  
Fusion Model ◽  
Average Accuracy ◽  
Benchmark Datasets ◽  
Operator Fatigue

Facial expression analysis and recognition has gained popularity in the last few years for its challenging nature and broad area of applications like HCI, pain detection, operator fatigue detection, surveillance, etc. The key of real-time FER system is exploiting its variety of features extracted from the source image. In this article, three different features viz. local binary pattern, Gabor, and local directionality pattern were exploited to perform feature fusion and two classification algorithms viz. support vector machines and artificial neural networks were used to validate the proposed model on benchmark datasets. The classification accuracy has been improved in the proposed feature fusion of Gabor and LDP features with SVM classifier, recorded an average accuracy of 93.83% on JAFFE, 95.83% on CK and 96.50% on MMI. The recognition rates were compared with the existing studies in the literature and found that the proposed feature fusion model has improved the performance.

scholarly journals Fungal Disease in Cotton Leaf Detection and Classification using Neural Networks and Support Vector Machine

2019 ◽  
Vol 8 (10) ◽  
pp. 3664-3667 ◽  
Keyword(s):  
Support Vector Machine ◽  
Production Rate ◽  
Clustering Algorithm ◽  
Economic Status ◽  
Fungal Disease ◽  
Support Vector ◽  
Svm Classifier ◽  
Disease Detection ◽  
Cotton Leaf ◽  
Average Accuracy

Agriculture productivity is the main factor for improving economic status of India. Reduction in production rate is mainly due to various diseases in plants. Identification of plant disease in early stage is the main challenge for improving the production rate as well as economic status. This paper presents automatic disease detection in cotton crop for three types of diseases Alternaria Leaf Spot Fungal Disease (ALSFD), Grey Mildew Cotton Disease (GMCD), and Rust Foliar Fungal Disease (RFFD). The K-means clustering algorithm is used for disease segmentation for cotton leaf. The diseased cluster is segmented into three clusters. From cluster 2 the features Mean , Contrast, Energy, Correlation, Standard Deviation, Variance , Entropy, and Kurtosis are extracted. The extracted features for 30 samples are given to Artificial Neural Network (ANN) and Support Vector Machine (SVM) classifiers for disease classification. The performance of these classifiers are compared. The ALSF disease is classified 77.4% for ANN and 84.3% for SVM, GMC disease is 87.8% for ANN and 98.7% in SVM, RFF disease is 90.1%for ANN and 93.2% for SVM. The overall average accuracy of ANN classifier is 85.1% for three diseases and overall average accuracy for SVM is 92.06% for three diseases. It is clearly observed from the analysis SVM classifier gives accurate disease detection compared to ANN.

scholarly journals Detection and Classification of Ring, Rust and Yellow Sugarcane Leaf Diseases

2019 ◽  
Vol 8 (6) ◽  
pp. 3310-3315
Keyword(s):  
Social Life ◽  
Color Image ◽  
Region Of Interest ◽  
Support Vector ◽  
Svm Classifier ◽  
Morphological Operations ◽  
Color Processing ◽  
Grey Scale ◽  
Average Accuracy ◽  
Crop Disease

Agriculture is an important sector in Economic and Social life. Crop disease detection is an emerging field in India. We can minimize the diseases infection on sugarcane leaf by detecting and grading the leaf disease in early stages. In this paper, we are detecting and recognize Sugar cane leaf diseases by using grey scale and color image processing and analyze the efficacy by comparing both. In grey scale processing, we presented Gradient Magnitude, Otsu method, Morphological Operations and Normalization to extract the Region of interest (ROI) i.e., disease part. In color processing initially converted RGB to L*a*b format, later K-means clustering and edge detection operations are applied on L*a*b image format. The features of Grey scale & color processed image are extracted and feed to Support Vector Machine (SVM) classifier which classifies ring, rust & yellow spot sugarcane leaf diseases. The Sugarcane leaf diseases are classified successfully with an average accuracy of 84% & 92% for grey scale & color features respectively.

scholarly journals A convolutional neural network for classifying cloud particles recorded by imaging probes

2020 ◽  
Vol 13 (5) ◽  
pp. 2219-2239 ◽  
Author(s):  
Georgios Touloupas ◽  
Annika Lauber ◽  
Jan Henneberger ◽  
Alexander Beck ◽  
Aurélien Lucchi
Keyword(s):  
Neural Network ◽  
Convolutional Neural Network ◽  
Nonlinear Models ◽  
Fine Tuning ◽  
Support Vector ◽  
Automated Classification ◽  
Cloud Particle ◽  
Imaging Probes ◽  
Regularization Techniques ◽  
Cloud Particles

Abstract. During typical field campaigns, millions of cloud particle images are captured with imaging probes. Our interest lies in classifying these particles in order to compute the statistics needed for understanding clouds. Given the large volume of collected data, this raises the need for an automated classification approach. Traditional classification methods that require extracting features manually (e.g., decision trees and support vector machines) show reasonable performance when trained and tested on data coming from a unique dataset. However, they often have difficulties in generalizing to test sets coming from other datasets where the distribution of the features might be significantly different. In practice, we found that for holographic imagers each new dataset requires labeling a huge amount of data by hand using those methods. Convolutional neural networks have the potential to overcome this problem due to their ability to learn complex nonlinear models directly from the images instead of pre-engineered features, as well as by relying on powerful regularization techniques. We show empirically that a convolutional neural network trained on cloud particles from holographic imagers generalizes well to unseen datasets. Moreover, fine tuning the same network with a small number (256) of training images improves the classification accuracy. Thus, the automated classification with a convolutional neural network not only reduces the hand-labeling effort for new datasets but is also no longer the main error source for the classification of small particles.

Detection of Internally Bruised Blueberries Using Hyperspectral Transmittance Imaging

2017 ◽  
Vol 60 (5) ◽  
pp. 1489-1502 ◽  
Author(s):  
Mengyun Zhang ◽  
Changying Li ◽  
Fumiomi Takeda ◽  
Fuzeng Yang
Keyword(s):  
Near Infrared ◽  
Imaging System ◽  
Mechanical Damage ◽  
Color Images ◽  
Economic Losses ◽  
Support Vector ◽  
Svm Classifier ◽  
Average Accuracy ◽  
Fruit Samples ◽  
And Control

Abstract. Internal bruise damage that occurs in blueberry fruit during harvest operations and postharvest handling lowers the overall quality and causes significant economic losses. The main goal of this study was to nondestructively detect internal bruises in blueberries after mechanical damage using hyperspectral transmittance imaging. A total of 600 hand-harvested blueberries were divided into 20 groups of four storage times (30 min, 3 h, 12 h, and 24 h), two storage temperatures (22°C and 4°C), and three treatments (stem bruise, equator bruise, and control). A near-infrared hyperspectral imaging system was used to acquire transmittance images from 970 to 1400 nm with 5 nm bandwidth. Images were acquired from three orientations (calyx-up, stem-up, and equator-up) for fruit in the control and stem bruise groups and from four orientations (calyx-up, stem-up, equator-up, and equator-down) in the equator bruise groups. Immediately after imaging, the fruit samples were sliced, and the sliced surfaces were photographed. The color images of sliced fruit were used as references. By comparing with the reference color images, the profiles of spatial and spectral intensities were evaluated to observe the effect of orientation and help extract regions of interest (ROIs) of bruised and healthy tissues. A support vector machine (SVM) classifier was trained and tested to classify pixels of bruised and healthy tissues. Classification maps were produced, and the bruise ratio was calculated to identify bruised blueberries (bruise ratio >25%). The average accuracy of blueberry identification was 94.5% with the stem-up orientation. The results indicate that detecting bruised blueberries as soon as 30 min after mechanical damage is feasible using hyperspectral transmittance imaging. Keywords: Blueberry, Bruise detection, Classification, Hyperspectral imagery, Transmittance mode.

scholarly journals Link prediction in multiplex online social networks

2017 ◽  
Vol 4 (2) ◽  
pp. 160863 ◽  
Author(s):  
Mahdi Jalili ◽  
Yasin Orouskhani ◽  
Milad Asgari ◽  
Nazanin Alipourfard ◽  
Matjaž Perc
Keyword(s):  
Social Networks ◽  
Online Social Networks ◽  
Link Prediction ◽  
Support Vector ◽  
Svm Classifier ◽  
Multiplex Network ◽  
Average Accuracy ◽  
Potential Applications ◽  
Meta Path ◽  
Highly Correlated

Online social networks play a major role in modern societies, and they have shaped the way social relationships evolve. Link prediction in social networks has many potential applications such as recommending new items to users, friendship suggestion and discovering spurious connections. Many real social networks evolve the connections in multiple layers (e.g. multiple social networking platforms). In this article, we study the link prediction problem in multiplex networks. As an example, we consider a multiplex network of Twitter (as a microblogging service) and Foursquare (as a location-based social network). We consider social networks of the same users in these two platforms and develop a meta-path-based algorithm for predicting the links. The connectivity information of the two layers is used to predict the links in Foursquare network. Three classical classifiers (naive Bayes, support vector machines (SVM) and K-nearest neighbour) are used for the classification task. Although the networks are not highly correlated in the layers, our experiments show that including the cross-layer information significantly improves the prediction performance. The SVM classifier results in the best performance with an average accuracy of 89%.

scholarly journals Box model trajectory studies of contrail formation using a particle-based cloud microphysics scheme

2021 ◽  
Author(s):  
Andreas Bier ◽  
Simon Unterstrasser ◽  
Xavier Vancassel
Keyword(s):  
Relative Humidity ◽  
Geometric Mean ◽  
Cloud Microphysics ◽  
Ice Crystals ◽  
Box Model ◽  
Radial Dependence ◽  
Ice Crystal ◽  
Ambient Conditions ◽  
Microphysics Scheme ◽  
Soot Particles

Abstract. We investigate the microphysics of contrail formation behind commercial aircraft by means of the particle-based LCM (Lagrangian Cloud Module) box model. We extend the original LCM to cover the basic pathway of contrail formation of soot particles being activated into liquid droplets that soon after freeze into ice crystals. In our particle-based microphysical approach, simulation particles are used to represent different particle types (soot, droplets, ice crystals) and properties (mass/radius, number). The box model is applied in two frameworks. In the classical framework, we prescribe the dilution along one average trajectory in a single box model run. In the second framework, we perform a large ensemble of box model runs using 25000 different trajectories inside an expanding exhaust jet as simulated by the LES (large-eddy simulation) model FLUDILES. In the ensemble runs, we see a strong radial dependence of the temperature and relative humidity evolution. Droplet formation on soot particles happens first near the plume edge and a few tenths of seconds later in the plume centre. Averaging over the ensemble runs, the number of formed droplets/ice crystals increases more smoothly over time than for the single box model run with the average dilution. Consistent with previous studies, contrail ice crystal number varies strongly with atmospheric parameters like temperature and relative humidity near the contrail formation threshold. Close to this threshold, the freezing fraction of soot particles depends strongly on the geometric-mean dry core radius and the hygroscopicity parameter of soot particles. This sensitivity is quite low at ambient conditions far away from the formation threshold. Absolute ice crystal numbers, on the other hand, are controlled by the soot number emission index for all atmospheric conditions. The comparison with a recent contrail formation study by Lewellen (2020) (using similar microphysics) shows a later onset of our contrail formation due to a weaker prescribed plume dilution. If we use the same dilution data, our and Lewellen's evolution in contrail ice nucleation show an excellent agreement cross-validating both microphysics implementations. This means that differences in contrail properties mainly result from different representations of the plume mixing and not from the microphysical modelling. The presented aerosol and microphysics scheme describing contrail formation is of intermediate complexity and thus suited to be incorporated in an LES model for 3D contrail formation studies explicitly simulating the jet expansion. The presented box model results will help interpreting the upcoming, more complex 3D results.

Online Direction Classification and Recognition Using an Intuitive Tactile Communication

2014 ◽  
Vol 543-547 ◽  
pp. 1542-1545
Author(s):  
Hao Ying Wu ◽  
Kai Fan
Keyword(s):  
Feature Extraction ◽  
Extraction Method ◽  
Support Vector ◽  
Svm Classifier ◽  
Feature Extraction Method ◽  
Average Accuracy ◽  
Vector Machines ◽  
Tactile Communication ◽  
Low Costs ◽  
Robot Cooperation

This paper proposed an online direction classifying method for constructing an intuitive tactile communication during human-robot cooperation. The proposed approach abstracts a suitable feature set from a tactile array sensor equipped on a hand-bar. This lower computation feature extraction method analyze the weighting values concerned with oriental information from principle component analysis (PCA), together with support vector machines (SVM) classifier for direction classification and recognition. Experimental results showed an average accuracy of 96.3% and a low costs of 512μs with respect to different handle gestures of the 6 touch directions, which is practicable utilized for human-robot cooperation based on tactile recognition.

scholarly journals Particle Habit Imaging Using Incoherent Light: A First Step toward a Novel Instrument for Cloud Microphysics

2011 ◽  
Vol 28 (4) ◽  
pp. 493-512 ◽  
Author(s):  
Roland Schön ◽  
Martin Schnaiter ◽  
Zbigniew Ulanowski ◽  
Carl Schmitt ◽  
Stefan Benz ◽  
...  
Keyword(s):  
Ccd Camera ◽  
Cloud Microphysics ◽  
Resolving Power ◽  
Ice Crystal ◽  
Ambient Conditions ◽  
Incoherent Light ◽  
Cloud Particle ◽  
Ice Particles ◽  
Institute Of Technology ◽  
A Charge

Abstract The imaging unit of the novel cloud particle instrument Particle Habit Imaging and Polar Scattering (PHIPS) probe has been developed to image individual ice particles produced inside a large cloud chamber. The PHIPS produces images of single airborne ice crystals, illuminated with white light of an ultrafast flashlamp, which are captured at a maximum frequency of ∼5 Hz by a charge-coupled device (CCD) camera with microscope optics. The imaging properties of the instrument were characterized by means of crystalline sodium hexafluorosilicate ice analogs, which are stable at room temperature. The optical resolving power of the system is ∼2 μm. By using dedicated algorithms for image processing and analysis, the ice crystal images can be analyzed automatically in terms of size and selected shape parameters. PHIPS has been operated at the cloud simulation chamber facility Aerosol Interaction and Dynamics in the Atmosphere (AIDA) of the Karlsruhe Institute of Technology at different temperatures between −17° and −4°C in order to study the influence of the ambient conditions, that is, temperature and ice saturation ratio, on ice crystal habits. The area-equivalent size distributions deduced from the PHIPS images are compared with the retrieval results from Fourier transform infrared (FTIR) extinction spectroscopy in case of small (<20 μm) and with single particle data from the cloud particle imager in case of larger (>20 μm) ice particles. Good agreement is found for both particle size regimes.

Sign in / Sign up

Export Citation Format

Share Document