DRBM-ClustNet: A Deep Restricted Boltzmann-Kohonen Architecture for Data Clustering.

A Bayesian deep restricted Boltzmann-Kohonen architecture for data clustering termed deep restricted Boltzmann machine (DRBM)-ClustNet is proposed. This core-clustering engine consists of a DRBM for processing unlabeled data by creating new features that are uncorrelated and have large variance with each other. Next, the number of clusters is predicted using the Bayesian information criterion (BIC), followed by a Kohonen network (KN)-based clustering layer. The processing of unlabeled data is done in three stages for efficient clustering of the nonlinearly separable datasets. In the first stage, DRBM performs nonlinear feature extraction by capturing the highly complex data representation by projecting the feature vectors of d dimensions into n dimensions. Most clustering algorithms require the number of clusters to be decided a priori; hence, here, to automate the number of clusters in the second stage, we use BIC. In the third stage, the number of clusters derived from BIC forms the input for the KN, which performs clustering of the feature-extracted data obtained from the DRBM. This method overcomes the general disadvantages of clustering algorithms, such as the prior specification of the number of clusters, convergence to local optima, and poor clustering accuracy on nonlinear datasets. In this research, we use two synthetic datasets, 15 benchmark datasets from the UCI Machine Learning repository, and four image datasets to analyze the DRBM-ClustNet. The proposed framework is evaluated based on clustering accuracy and ranked against other state-of-the-art clustering methods. The obtained results demonstrate that the DRBM-ClustNet outperforms state-of-the-art clustering algorithms.

Combined two-level damage identification strategy using ultrasonic guided waves and physical knowledge assisted machine learning.

Structural Health Monitoring of composite structures is one of the significant challenges faced by the aerospace industry. A combined two-level damage identification viz damage detection and localization is performed in this paper for a composite panel using ultrasonic guided waves. A novel physical knowledge-assisted machine learning technique is proposed in which domain knowledge and expert supervision is utilized to assist the learning process. Two supervised learning-based convolutional neural networks are trained for damage detection (binary classification) and localization (multi-class classification) on an experimental benchmark dataset. The performance of the trained models is evaluated using loss curve, accuracy, confusion matrix, and receiver-operating characteristics curve. It is observed that incorporating physical knowledge helps networks perform better than a direct deep learning approach. In this work, a combined damage identification strategy is proposed for a real-time application. In this strategy, the damage detection model works in an outer-loop and predicts the state of the structure (undamaged or damaged), whereas an inner-loop predicts the location of the damage only if the outer-loop detects damage. It is seen that the proposed technique offers advantages in terms of accuracy (above 99% for both detection and localization), computational time (prediction time per signal in milliseconds), sensor optimization, in-situ monitoring, and robustness towards the noise.

Scaling Effects on Chlorophyll Content Estimations with RGB Camera Mounted on a UAV Platform Using Machine-Learning Methods.

Timely monitoring and precise estimation of the leaf chlorophyll contents of maize are crucial for agricultural practices. The scale effects are very important as the calculated vegetation index (VI) were crucial for the quantitative remote sensing. In this study, the scale effects were investigated by analyzing the linear relationships between VI calculated from red-green-blue (RGB) images from unmanned aerial vehicles (UAV) and ground leaf chlorophyll contents of maize measured using SPAD-502. The scale impacts were assessed by applying different flight altitudes and the highest coefficient of determination (R2) can reach 0.85. We found that the VI from images acquired from flight altitude of 50 m was better to estimate the leaf chlorophyll contents using the DJI UAV platform with this specific camera (5472 × 3648 pixels). Moreover, three machine-learning (ML) methods including backpropagation neural network (BP), support vector machine (SVM), and random forest (RF) were applied for the grid-based chlorophyll content estimation based on the common VI. The average values of the root mean square error (RMSE) of chlorophyll content estimations using ML methods were 3.85, 3.11, and 2.90 for BP, SVM, and RF, respectively. Similarly, the mean absolute error (MAE) were 2.947, 2.460, and 2.389, for BP, SVM, and RF, respectively. Thus, the ML methods had relative high precision in chlorophyll content estimations using VI; in particular, the RF performed better than BP and SVM. Our findings suggest that the integrated ML methods with RGB images of this camera acquired at a flight altitude of 50 m (spatial resolution 0.018 m) can be perfectly applied for estimations of leaf chlorophyll content in agriculture.

Modified Red Blue Vegetation Index for Chlorophyll Estimation and Yield Prediction of Maize from Visible Images Captured by UAV.

The vegetation index (VI) has been successfully used to monitor the growth and to predict the yield of agricultural crops. In this paper, a long-term observation was conducted for the yield prediction of maize using an unmanned aerial vehicle (UAV) and estimations of chlorophyll contents using SPAD-502. A new vegetation index termed as modified red blue VI (MRBVI) was developed to monitor the growth and to predict the yields of maize by establishing relationships between MRBVI- and SPAD-502-based chlorophyll contents. The coefficients of determination (R2s) were 0.462 and 0.570 in chlorophyll contents' estimations and yield predictions using MRBVI, and the results were relatively better than the results from the seven other commonly used VI approaches. All VIs during the different growth stages of maize were calculated and compared with the measured values of chlorophyll contents directly, and the relative error (RE) of MRBVI is the lowest at 0.355. Further, machine learning (ML) methods such as the backpropagation neural network model (BP), support vector machine (SVM), random forest (RF), and extreme learning machine (ELM) were adopted for predicting the yields of maize. All VIs calculated for each image captured during important phenological stages of maize were set as independent variables and the corresponding yields of each plot were defined as dependent variables. The ML models used the leave one out method (LOO), where the root mean square errors (RMSEs) were 2.157, 1.099, 1.146, and 1.698 (g/hundred grain weight) for BP, SVM, RF, and ELM. The mean absolute errors (MAEs) were 1.739, 0.886, 0.925, and 1.356 (g/hundred grain weight) for BP, SVM, RF, and ELM, respectively. Thus, the SVM method performed better in predicting the yields of maize than the other ML methods. Therefore, it is strongly suggested that the MRBVI calculated from images acquired at different growth stages integrated with advanced ML methods should be used for agricultural- and ecological-related chlorophyll estimation and yield predictions.

Correction to: Multitemporal time series analysis using machine learning models for ground deformation in the Erhai Region, China.

Unfortunately, the name of the corresponding author (Wenxiang Wu) was missing in the author group section of the published paper.

Multitemporal time series analysis using machine learning models for ground deformation in the Erhai region, China.

Ground deformation (GD) has been widely reported as a global issue and is now an ongoing problem that will profoundly endanger the public safety. GD is a complex and dynamic problem with many contributing factors that occur over time. In the literature, there are only a few methods that can effectively monitor GD. Microwave remote sensing data such as interferometric synthetic aperture radar (InSAR) are mostly adopted to assess GD. These data can reveal the surface deforming areas with great precision, mapping GD results at a large scale. In this study, the effects of GD and the influencing factors, such as the building area, the water level, the cumulative precipitation, and the cumulative temperature, are modeled in the Erhai region with small baseline subset interferometric SAR (SBAS-InSAR) data that are applied using machine learning (ML) methods. The ML methods, namely, multiple linear regression (MLR), multilayer perceptron backpropagation (MLP-BP), least squares support vector machine (LSSVM), and particle swarm optimization (PSO)-LSSVM, are used to predict GD, and the results are compared. Particularly, the PSO-LSSVM method has obtained the least root mean square error (RMSE) and mean relative error (MRE) of 11.448 and 0.112, respectively. Therefore, the results have proven that the proposed PSO-LSSVM is very efficient in analyzing GD.

Mission Aware Motion Planning (MAP) Framework With Physical and Geographical Constraints for a Swarm of Mobile Stations.

In this paper, we propose a mission aware motion planning (MAP) framework for a swarm of autonomous unmanned ground vehicles (UGVs) or mobile stations in an uncertain environment for efficient supply of resources/services to unmanned aerial vehicles (UAVs) performing a specific mission. The MAP framework consists of two levels, namely, centralized mission planning and decentralized motion planning. On the first level, the centralized mission planning algorithm estimates the density of UAV in a given environment for determining the number of UGVs and their initial operating location. In the subsequent level, a decentralized motion planning algorithm which provides a closed-form expression for velocity command using adaptive density estimation has been proposed. Further, the physical and geographical constraints are integrated into motion planning. A Monte-Carlo simulation is performed to evaluate the advantages of the MAP over distributed stationary stations (DSSs) often used in the literature. The obtained results clearly indicate that in comparison with DSS, MAP reduces the average distance traveled by UAVs about 20%, reduces the loss of mission time by 90 s per interruption and power loss by 3 dB.

Detection of the power lines in UAV remote sensed images using spectral-spatial methods.

In this paper, detection of the power lines on images acquired by Unmanned Aerial Vehicle (UAV) based remote sensing is carried out using spectral-spatial methods. Spectral clustering was performed using Kmeans and Expectation Maximization (EM) algorithm to classify the pixels into the power lines and non-power lines. The spectral clustering methods used in this study are parametric in nature, to automate the number of clusters Davies-Bouldin index (DBI) is used. The UAV remote sensed image is clustered into the number of clusters determined by DBI. The k clustered image is merged into 2 clusters (power lines and non-power lines). Further, spatial segmentation was performed using morphological and geometric operations, to eliminate the non-power line regions. In this study, UAV images acquired at different altitudes and angles were analyzed to validate the robustness of the proposed method. It was observed that the EM with spatial segmentation (EM-Seg) performed better than the Kmeans with spatial segmentation (Kmeans-Seg) on most of the UAV images.