Band Selection Research Articles

A hyperspectral band selection algorithm for identifying high oleic acid peanuts

High oleic acid peanuts have higher oleic acid content and stronger oxidation stability than common peanuts, but their appearances are similar, which imposes difficulties for classifying. Based on this, the study aims to classify high oleic acid peanut to ensure its purity by using hyperspectral imaging technology. However, classification accuracy and efficiency are limited given the large amount of redundant information of hyperspectral images. The band iteration algorithm (BIA) is proposed to select characteristic bands by reducing the redundant information between spectral bands for the peanut classification. Hyperspectral images with 616 bands (from 400 nm to 1100 nm) of 126 high oleic acid peanuts and 126 common peanuts were collected. Then, BIA selected optimal bands as characteristic bands from adjacent bands according to the classification accuracy of each band subsets. Thirdly, three classification models, namely linear discriminant analysis, support vector machine, and partial least squares-discriminant analysis (PLS-DA), were employed to compare the performance of BIA with successive projections algorithm and competitive adaptive reweighted sampling, respectively. The experimental results show that BIA can effectively improve the classification ability of spectral data. The BIA-PLS-DA model had the best classification efficiency, and the accuracy of the test set reached 93.26%. For peanut individuals, only one peanut sample was misclassified with a classification error rate of 1.43%.

A discriminatory groups-based supervised band selection technique for hyperspectral image classification

ABSTRACT Dimensionality reduction is a prevalent preprocessing step when analysing hyperspectral images (HSIs). It uses band selection (BS) and feature extraction (FE) techniques, but the former can retain the physical meaning and interpretability of the HSI datasets. Hence, these techniques are preferred to select a proper subset of bands without transformation. This letter introduces a straightforward supervised BS technique using discriminatory groups (DGs) that does not require complex computations. In this technique, total DGs are achieved equal to the spectral dimension of the HSI dataset, and each holds bands similar to the number of classes. The constituent bands of the DGs may present more than once, but the total number of bands in the group remains unchanged. The obtained DGs are ranked using high- to low-discrimination fractions, which indicate the uniqueness of the respective DG. The highest value that a DG can attain represents the group’s ability to distinguish among all classes (the maximum number of classes). Hence, the proposed technique has been termed ‘maximum class discriminatory group (MCDG)’. The discrimination fraction works as the control parameter for the selection of bands. Subsequently, the unique bands are selected from the ranked DGs and used for HSI classification. The proposed MCDG technique has an articulate and quick implementation without requiring any initialization of bands and outperforms various state-of-the-art methods while experimenting using real HSI datasets.

DTMSgram: a novel optimal demodulation frequency band selection method for wheelset bearings fault diagnosis under wheel-rail excitation

Due to the influence of the wheel-rail excitation and complex transmission path, the fault signature of wheelset bearings is often obscured by complex background noise, which brings great challenges to the adaptive determination of the informative frequency band (IFB) in envelope analysis. In this paper, the vibration response characteristics of the axle box under wheel–rail excitation are revealed through full-scale bench tests. The experimental results show that tread damage will provoke periodic transient impacts and it has an obvious sparsity in the frequency domain. Inspired by this feature, a DTMSgram method is proposed to enhance the fault feature components of vibration signals through time-domain and frequency-domain noise reduction technology, and improve the accuracy of demodulation frequency band selection. Firstly, the amplitude spectrum of different weight coefficients is used to preprocess the vibration signal and adjust the full band component of the vibration signal. Then, the autocorrelation process is performed on each layer’s narrowband filtered signal envelope in kurtogram to further reduce noise interference from a time-domain perspective. Moreover, a two-dimensional color map is constructed showing the normalized squared envelope spectrum of the IFBs determined simultaneously from the modified signal. Finally, the effectiveness of the proposed method is verified by simulated signal and experimental data of three sets of full-size wheelset bearing systems. The analysis results indicate this proposed method can effectively overcome the influence of complex wheel–rail excitation interference and can diagnose multi-source faults simultaneously.

Lightweight-VGG: A Fast Deep Learning Architecture Based on Dimensionality Reduction and Nonlinear Enhancement for Hyperspectral Image Classification

In the past decade, deep learning methods have proven to be highly effective in the classification of hyperspectral images (HSI), consistently outperforming traditional approaches. However, the large number of spectral bands in HSI data can lead to interference during the learning process. To address this issue, dimensionality reduction techniques can be employed to minimize data redundancy and improve HSI classification performance. Hence, we have developed an efficient lightweight learning framework consisting of two main components. Firstly, we utilized band selection and principal component analysis to reduce the dimensionality of HSI data, thereby reducing redundancy while retaining essential features. Subsequently, the pre-processed data was input into a modified VGG-based learning network for HSI classification. This method incorporates an improved dynamic activation function for the multi-layer perceptron to enhance non-linearity, and reduces the number of nodes in the fully connected layers of the original VGG architecture to improve speed while maintaining accuracy. This modified network structure, referred to as lightweight-VGG (LVGG), was specifically designed for HSI classification. Comprehensive experiments conducted on three publicly available HSI datasets consistently demonstrated that the LVGG method exhibited similar or better performance compared to other typical methods in the field of HSI classification. Our approach not only addresses the challenge of interference in deep learning methods for HSI classification, but also offers a lightweight and efficient solution for achieving high classification accuracy.

Investigation into Impedance Measurements for Rapid Capacity Estimation of Lithium-ion Batteries in Electric Vehicles

With the dramatic increase in electric vehicles (EVs) globally, the demand for lithium-ion batteries has grown dramatically, resulting in many batteries being retired in the future. Developing a rapid and robust capacity estimation method is a challenging work to recognize the battery ageing level on service and provide regroup strategy of the retied batteries in secondary use. There are still limitations on the current rapid battery capacity estimation methods, such as direct current internal resistance (DCIR) and electrochemical impedance spectroscopy (EIS), in terms of efficiency and robustness. To address the challenges, this paper proposes an improved version of DCIR, named pulse impedance technique (PIT), for rapid battery capacity estimation with more robustness. First, PIT is carried out based on the transient current excitation and dynamic voltage measurement using the high sampling frequency, in which the coherence analysis is used to guide the selection of a reliable frequency band. The battery impedance can be extracted in a wide range of frequency bands compared to the traditional DCIR method, which obtains more information on the battery capacity evaluation. Second, various statistical variables are used to extract ageing features and Pearson correlation analysis is applied to determine the highly correlated features. Then a linear regression model is developed to map the relationship between extracted features and battery capacity. To validate the performance of the proposed method, the experimental system is designed to conduct comparative studies between PIT and EIS based on the two 18650 batteries connected in series. The results reveal that the proposed PIT can provide comparative indicators to EIS, which contributes higher estimation accuracy of the proposed PIT method than EIS technology with lower time and cost. Conflict of Interest Statement The authors declare no conflicts of interest.

New data-driven estimation of metal element in rocks using a hyperspectral data and geochemical data

The inversion of Cu contents in rocks based on hyperspectral remote sensing is indicative of geochemical exploration as well as mineral resource exploration. Due to the strong redundancy of band information in hyperspectral data, characteristic band selection is the key to improving the accuracy of metal element inversion models in hyperspectral data. In this study, we proposed a framework that innovatively focuses on secondary screening method (SSM) as the core, allowing users to add different data preprocessing methods and machine learning algorithms to the framework based on expert experience, to obtain high-precision metal element inversion models suitable for sampling data. we first preprocessed the measured spectral curves using the first order derivative, second order derivative, continuum removal, multiple scattering correction and variance analysis to highlight band absorption characteristics. In addition, we performed SSM to obtain Cu-related characteristic bands, which were further used as input data to train and compare the performance of Partial Least Square Regression (PLSR), Support Vector Regression (SVR), Multi-Layer Perceptron (MLP), Random Forest (RF) and Gradient Boosting Regression Tree (GBRT). The results show that both the spectral pretreatment and modeling method could potentially impact the inversion accuracy. The accuracy of the ensemble method applied in fresh faces data (i.e., a CR-SSM-RF with R2 values of 0.77 and CR-SSM-GBRT with R2 values of 0.81) performed better than other models with an accuracy improvement of up to 80%. Compared to the full band model, the accuracy of characteristic bands selection model increased 50%. This also illustrates that the CR-SSM-RF or CR-SSM-GBRT model significantly improves the accuracy of the inversion of Cu contents in rocks compared to the original spectral reflectance. Our study provides a new idea for selecting different spectral preprocessing methods as the input values for the future model and will also offer an executable framework for the inversion of Cu contents in rocks of different lithology on a large scale in the future.

Refined fire detection and band selection method in hyperspectral remote sensing imagery based on sparse-VIT

Hyperspectral remote sensing has powerful ability to perceive information of land surface, and can be applied to the field of fire monitoring to achieve refined fire analysis. In this paper, we study the fire detection method based on hyperspectral remote sensing, design an end to end fire detection model based on sparse visual transformer, and propose a sensitive band selection method under the transformer framework. In order to suppress the interference of invalid bands in hyperspectral data, the sparse attention mechanism and the top-k selection mechanism are introduced in the model to zeroing the attention score of invalid bands. To achieve dimension reduction, a non-maximum attention suppression algorithm under the transformer framework is designed and band pruning is further introduced to eliminate invalid bands and redundant bands. The model adopts input mode of band-exclusive-token, which makes the pruning operation equivalent to band selection. We publish a hyperspectral fire detection dataset and the performance of our proposed model is verified on the dataset.

Self-Supervised Deep Multi-Level Representation Learning Fusion-Based Maximum Entropy Subspace Clustering for Hyperspectral Band Selection

As one of the most important techniques for hyperspectral image dimensionality reduction, band selection has received considerable attention, whereas self-representation subspace clustering-based band selection algorithms have received quite a lot of attention with good effect. However, many of them lack the self-supervision of representations and ignore the multi-level spectral–spatial information of HSI and the connectivity of subspaces. To this end, this paper proposes a novel self-supervised multi-level representation learning fusion-based maximum entropy subspace clustering (MLRLFMESC) method for hyperspectral band selection. Firstly, to learn multi-level spectral–spatial information, self-representation subspace clustering is embedded between the encoder layers of the deep-stacked convolutional autoencoder and its corresponding decoder layers, respectively, as multiple fully connected layers to achieve multi-level representation learning (MLRL). A new auxiliary task is constructed for multi-level representation learning and multi-level self-supervised training to improve its capability of representation. Then, a fusion model is designed to fuse the multi-level spectral–spatial information to obtain a more distinctive coefficient matrix for self-expression, where the maximum entropy regularization (MER) method is employed to promote connectivity and the uniform dense distribution of band elements in each subspace. Finally, subspace clustering is conducted to obtain the final band subset. Experiments have been conducted on three hyperspectral datasets, and the corresponding results show that the proposed MLRLFMESC algorithm significantly outperforms several other band selection methods in classification performance.

Deep learning-based spectral reconstruction in camouflaged target detection

Camouflaged target detection aims to detect targets that blend into their surroundings, but RGB has difficulty distinguishing between targets and backgrounds. While methods using multispectral image (MSI) can distinguish targets from background via spectral information, they are limited by imaging speed, resolution, and high cost for camouflaged target detection. Here, we propose a novel camouflaged target detection workflow based on reconstructed MSI from RGB image. Specifically, we propose a spectral reconstruction model, S2HFormer, which utilizes the deep neural network to fit the mapping of RGB image to MSI without additional information. And the reconstructed MSI based on S2HFormer achieves higher accuracy in both reconstruction and target detection, outperforming existing methods. Furthermore, we integrate a spectral band selection algorithm to optimize the number of bands used for improving detection efficiency. Experimental results show that the proposed method acquires MSI at 55 frames per second (FPS) and achieves an F-score of 0.925, achieving real-time (24 FPS) MSI acquisition. The evaluation indicates the effectiveness and efficiency of our method for camouflaged target detection.

Contrasting Effects of Structural Similarity and Entropic Metrics on Band Selection

Contrasting Effects of Structural Similarity and Entropic Metrics on Band Selection

Potential to use free satellite imagery to retrieve the past bathymetry of large rivers

Information on bathymetric changes in large rivers is useful for flood planning, navigation, and predicting channel change. Recent studies have demonstrated the potential for measuring bathymetry of rivers from satellite images, but such studies have used high-resolution, expensive data applied to ideal river conditions, such as low turbidity. Many of the large rivers of the world are located in countries that cannot afford the carefully planned field surveys and commercial satellite images used in these studies. In this study we investigate whether past bathymetry of large rivers could be retrieved by using readily collated free satellite images, combined with the limited cross-sectional surveys that are typically available in less wealthy countries. We tested the method in the heavily regulated Han River of central China. The Landsat-8 and Sentinel-2 image-derived depth could quantitatively match in situ depths along cross-sections up to depths of 13.9 m, where the suspended sediment concentration was low due to dam regulation. However, in most of the river where turbidity was high, the method was less successful. The image-to-depth quantile transformation (IDQT) approach could improve passive bathymetry retrieval in cases like the Han River, where there was some spatial and temporal misalignment between in situ data and remote-sensing data. The lateral depth pattern up to the 2.5 m depth contour could also be qualitatively and visually interpreted over an extended length (15 km). These results suggest that there is potential to use free satellite imagery to derive bathymetry in many large rivers, but this potential can only be achieved in very limited (low turbidity) conditions, and with image to depth correlation analysis and band selection.

Research on the Protection and Dissemination of Cultural Heritage in Rural Landscapes Based on Image Processing

Abstract The countryside is an important part of the social development process, but with the acceleration of urbanization, the protection of rural landscapes as cultural heritage is facing an increasingly severe situation. In this study, image radiation correction, image fusion, cropping and mosaicing, geometric correction, band selection, and image enhancement are applied to rural landscapes using remote sensing image processing technology. A digital protection system for rural landscape cultural heritage is created using the processed rural landscapes. By comparing the accuracy of this paper’s method to other classification methods, we can explore the classification performance of this paper’s PCA method. By comparing the changes in rural landscape types before and after landscape protection, the effect of landscape protection is explored. Finally, the communication effect of rural landscapes on social media is explored by utilizing the evaluation index system. The PCA method employed in this paper achieves a landscape classification accuracy of 83%, which is significantly superior to the IHS transformation method (73.5%) and the Brovey transformation method (76%). After the rural landscape protection, the degree of landscape fragmentation in Village A was significantly improved compared to before the protection, and the effect of protection was remarkable. The scores of social media users on each dimension of the rural landscape communication effect in Village A were greater than 4, and rural landscape communication achieved a positive effect.

End-to-End Hyperspectral Image Change Detection Based on Band Selection

End-to-End Hyperspectral Image Change Detection Based on Band Selection

Joint Spatial and Spectral Graph-Based Consistent Self-Representation for Unsupervised Hyperspectral Band Selection

Joint Spatial and Spectral Graph-Based Consistent Self-Representation for Unsupervised Hyperspectral Band Selection

Integration of Spatial Spectral Structure with Improved Local Density for Hyperspectral Image Band Selection Method

Integration of Spatial Spectral Structure with Improved Local Density for Hyperspectral Image Band Selection Method

A Cross-Scene Self-Representative Network for Hyperspectral Band Selection

A Cross-Scene Self-Representative Network for Hyperspectral Band Selection

MOBS-TD: Multiobjective Band Selection With Ideal Solution Optimization Strategy for Hyperspectral Target Detection

MOBS-TD: Multiobjective Band Selection With Ideal Solution Optimization Strategy for Hyperspectral Target Detection

Design of Information Management System Based on Random Leapfrog Band Selection Algorithm

With the rapid development of information technology, the traditional information management system faces the challenge of efficiency and accuracy when dealing with complex data and decision-making problems. In order to solve these problems, this paper integrates the random leapfrog band selection algorithm into the information management system. This algorithm shows excellent performance in dealing with large-scale, nonlinear and multi-variable optimization problems, resource allocation, scheduling optimization and data mining. This paper mainly explains the design framework of information management system, the optimization of information management by random leap-band selection algorithm and the design of information management system. Finally, two groups of simulation experiment results are as follows: the accuracy of the optimized information management system is 96%, the running time is 7.1min, the algorithm performance is better than the traditional system, and the response time of the system is reduced by 195ms. This research provides an efficient, flexible and scalable solution for dealing with various complex information management system optimization problems.

A Real-Time Unsupervised Hyperspectral Band Selection via Spatial-Spectral Information Fusion-Based Downscaled Region

A Real-Time Unsupervised Hyperspectral Band Selection via Spatial-Spectral Information Fusion-Based Downscaled Region

PU-KBS: A Robust Positive and Unlabeled Learning Framework With Key Band Selection for One-Class Hyperspectral Image Classification

PU-KBS: A Robust Positive and Unlabeled Learning Framework With Key Band Selection for One-Class Hyperspectral Image Classification