Features Of Hyperspectral Image Research Articles

Analysis of the spectral and textural features of hyperspectral images for the nondestructive prediction of amylopectin and amylose contents of sorghum

The contents of amylopectin and amylose are important indicators to evaluate the quality of sorghum. Hyperspectral imaging (HSI) is a technology based on machine vision and spectroscopy that has been widely used in the quality evaluation of various foods. Three combined algorithms were used to extract the characteristic wavelengths. The gray-level co-occurrence matrix (GLCM) was used to extract the textural features. Based on the spectra, textures, and fusion data, genetic algorithm-optimized back propagation neural network (BPNN-GA), cascade forest (CF), and partial least square regression (PLSR) models were established to predict the amylopectin and amylose contents. The BPNN-GA and CF models established using the characteristic wavelengths of visible-light were determined to be the best model, with RPD values of 8.2352 and 26.7308, respectively. These results demonstrated that HSI combined with machine learning algorithms could be used to enable the rapid and nondestructive prediction of amylopectin and amylose contents in sorghum.

An Enhanced Spectral Fusion 3D CNN Model for Hyperspectral Image Classification

With the continuous development of hyperspectral image technology and deep learning methods in recent years, an increasing number of hyperspectral image classification models have been proposed. However, due to the numerous spectral dimensions of hyperspectral images, most classification models suffer from issues such as breaking spectral continuity and poor learning of spectral information. In this paper, we propose a new classification model called the enhanced spectral fusion network (ESFNet), which contains two parts: an optimized multi-scale fused spectral attention module (FsSE) and a 3D convolutional neural network (3D CNN) based on the fusion of different spectral strides (SSFCNN). Specifically, after sampling the hyperspectral images, our model first implements the weighting of the spectral information through the FsSE module to obtain spectral data with a higher degree of information richness. Then, the weighted spectral data are fed into the SSFCNN to realize the effective learning of spectral features. The new model can maximize the retention of spectral continuity and enhance the spectral information while being able to better utilize the enhanced information to improve the model’s ability to learn hyperspectral image features, thus improving the classification accuracy of the model. Experiment results on the Indian Pines and Pavia University datasets demonstrated that our method outperforms other relevant baselines in terms of classification accuracy and generalization performance.

A novel strategy for classifying spectral-spatial shallow and deep hyperspectral image features using 1D-EWT and 3D-CNN

Hyperspectral images (HIs) are used in diverse disciplines, such as resource handling, land cover analysis, food science, anomaly detection, and precision agriculture. Researchers have been working on a number of visual processing and machine intelligence algorithms to handle this type of data as efficiently as feasible. Deep learning approaches have advanced significantly in the field of machine vision, which is also having a big impact on the analysis of hyperspectral data. To increase its discriminative potential for HI classification, this work suggests a powerful 3D-CNNs (Convolutional Neural Networks) architecture, where the shallow features extracted using 1D-EWT (Empirical Wavelet Transform) are served as input, and the ultimate output of the CNN are projected class-related outcomes. The framework is known as PEC, where P stands for PCA, E for 1D-EWT, and C for 3D-CNN model. Prior to features extraction, the HI undergoes spectral dimension reduction via Principal Component Anaalysis (PCA). To forecast segmentation for a volumetric area of a 3D HI sample, 3D CNNs use 3D convolutional kernels. The usage of more parameters by these CNNs means that the ability to employ interslice context that can boost performance. The CNN model’s parameters are optimised using a limited training set. The newly proposed PEC framework achieves a considerable overall accuracy of 99.58% and 99.94% (percentage increase of 0.08% and 0.54%) and Kappa score of 99.51% and 99.92 (percentage increase of 1.5% and 0.8%) compared to similar approaches for two real-world data sets IP and PU respectively with 30% training samples.

Multilayer graph spectral analysis for hyperspectral images

Hyperspectral imaging has broad applications and impacts in areas including environmental science, weather, and geo/space exploration. The intrinsic spectral–spatial structures and potential multi-level features in different frequency bands make multilayer graph an intuitive model for hyperspectral images (HSI). To study the underlying characteristics of HSI and to take the advantage of graph signal processing (GSP) tools, this work proposes a multilayer graph spectral analysis for hyperspectral images based on multilayer graph signal processing (M-GSP). More specifically, we present multilayer graph (MLG) models and tensor representations for HSI. By exploring multilayer graph spectral space, we develop MLG-based methods for HSI applications, including unsupervised segmentation and supervised classification. Our experimental results demonstrate the strength of M-GSP in HSI processing and spectral–spatial information extraction.

Hyperspectral Image Classification with IFormer Network Feature Extraction

Convolutional neural networks (CNNs) are widely used for hyperspectral image (HSI) classification due to their better ability to model the local details of HSI. However, CNNs tends to ignore the global information of HSI, and thus lack the ability to establish remote dependencies, which leads to computational cost consumption and remains challenging. To address this problem, we propose an end-to-end Inception Transformer network (IFormer) that can efficiently generate rich feature maps from HSI data and extract high- and low-frequency information from the feature maps. First, spectral features are extracted using batch normalization (BN) and 1D-CNN, while the Ghost Module generates more feature maps via low-cost operations to fully exploit the intrinsic information in HSI features, thus improving the computational speed. Second, the feature maps are transferred to Inception Transformer through a channel splitting mechanism, which effectively learns the combined features of high- and low-frequency information in the feature maps and allows for the flexible modeling of discriminative information scattered in different frequency ranges. Finally, the HSI features are classified via pooling and linear layers. The IFormer algorithm is compared with other mainstream algorithms in experiments on four publicly available hyperspectral datasets, and the results demonstrate that the proposed method algorithm is significantly competitive among the HSI classification algorithms.

Rapeseed Variety Recognition Based on Hyperspectral Feature Fusion

As an important oil crop, rapeseed contributes to the food security of the world. In recent years, agronomists have cultivated many new varieties, which has increased human nutritional needs. Variety recognition is of great importance for yield improvement and quality breeding. In view of the low efficiency and damage of traditional methods, in this paper, we develop a noninvasive model for the recognition of rapeseed varieties based on hyperspectral feature fusion. Three types of hyperspectral image features, namely, the multifractal feature, color characteristics, and trilateral parameters, are fused together to identify 11 rapeseed species. An optimal feature is selected using a simple rule, and then the three kinds of features are fused. The support vector machine kernel method is employed as a classifier. The average recognition rate reaches 96.35% and 93.71% for distinguishing two species and 11 species, respectively. The abundance test model demonstrates that our model possesses robustness. The high recognition rate is almost independent of the number of modeling samples and classifiers. This result can provide some practical experience and method guidance for the rapid recognition of rapeseed varieties.

Quantitative study on impact damage of yellow peach based on hyperspectral image information combined with spectral information

Yellow peaches are graded based on their degree of impact damage to reduce economic losses, and the values of mechanical parameters can be used to characterize the degree of impact damage of them. In this paper, the combined hyperspectral image features and spectral features method was proposed to predict the values of mechanical parameters of yellow peaches. A collision device based on the pendulum principle was built to acquire the mechanical parameters which were maximum force, absorbed energy and average pressure. A hyperspectral imaging system was used to acquire the images of the bruised yellow peaches and the spectral information and color characteristics of images were extracted, respectively. The damage areas of yellow peaches were obtained based on the combined principal component analysis (PCA) and threshold segmentation method. Texture features of the characteristic wavelength images were extracted based on the grey-level co-generation matrix (GLCM). The PLSR models of mechanical parameters were built based on spectral information, image information, and spectral information combined with image information, respectively. The results show that the PLSR models based on spectral information combined with image information are best in prediction the values of damage area and maximum force. However, the PLSR models have not achieved satisfactory performance in prediction the values of absorbed energy and average pressure. To improve the performance of PLSR models of absorbed energy and average pressure, the feature wavelengths were selected by CARS, and the correlation analysis was used between image information and mechanical parameters. The PLSR models were built based on feature spectral information, image feature information, and feature spectral information combined with image feature information, respectively. The results show that the accuracy of all mechanical parameters is significantly improved by the PLSR models based on the feature spectral information combined with the image feature information. The prediction set correlation coefficient (RP) and the root mean square error (RMSEP) of the damaged area, absorbed energy, maximum force and average pressure are 0.928 and 86.632 mm2, 0.826 and 1.469 J, 0.924 and 61.765 N, 0.815 and 0.050 MPa, respectively. The results of the study not only provide a theoretical basis for postharvest grading of fruit, but also provide a reference for analyzing the mechanical properties of fruit.

Prediction Model for Tea Polyphenol Content with Deep Features Extracted Using 1D and 2D Convolutional Neural Network

The content of tea polyphenols (TP) is one of the important indicators for judging the quality of tea. Accurate and non-destructive estimation technology for tea polyphenol content has attracted more and more attention, which has become a key technology for tea production, quality identification, grading and so on. Hyperspectral imaging technology is a fusion of spectral analysis and image processing technology, which has been proven to be an efficient technology for predicting tea polyphenol content. To make full use of spectral and spatial features, a prediction model of tea polyphenols based on spectral-spatial deep features extracted using convolutional neural network (CNN) was proposed, which not only broke the limitations of traditional shallow features, but also innovated the technical path of integrated deep learning in non-destructive detection for tea. Firstly, one-dimensional convolutional neural network (1D-CNN) and two-dimensional convolutional neural network (2D-CNN) models were constructed to extract the spectral deep features and spatial deep features of tea hyperspectral images, respectively. Secondly, spectral deep features, spatial deep features, and spectral-spatial deep features are used as input variables of machine learning models, including Partial Least Squares Regression (PLSR), Support Vector Regression (SVR) and Random Forest (RF). Finally, the training, testing and evaluation were realized using the self-built hyperspectral dataset of green tea from different grades and different manufacturers. The results showed that the model based on spectral-spatial deep features had the best prediction performance among the three machine learning models (R2 = 0.949, MAE = 0.533 for training sets, R2 = 0.938, MAE = 0.799 for test sets). Moreover, the visualization of estimation results of tea polyphenol content further demonstrated that the model proposed in this study had strong estimation ability. Therefore, the deep features extracted using CNN can provide new ideas for estimation of the main components of tea, which will provide technical support for the estimation tea quality estimation.

FusionNet: A Convolution–Transformer Fusion Network for Hyperspectral Image Classification

In recent years, deep-learning-based hyperspectral image (HSI) classification networks have become one of the most dominant implementations in HSI classification tasks. Among these networks, convolutional neural networks (CNNs) and attention-based networks have prevailed over other HSI classification networks. While convolutional neural networks with perceptual fields can effectively extract local features in the spatial dimension of HSI, they are poor at capturing the global and sequential features of spectral–spatial information; networks based on attention mechanisms, for example, Transformer, usually have better ability to capture global features, but are relatively weak in discriminating local features. This paper proposes a fusion network of convolution and Transformer for HSI classification, known as FusionNet, in which convolution and Transformer are fused in both serial and parallel mechanisms to achieve the full utilization of HSI features. Experimental results demonstrate that the proposed network has superior classification results compared to previous similar networks, and performs relatively well even on a small amount of training data.

Small Sample Hyperspectral Image Classification Method Based on Dual-Channel Spectral Enhancement Network

Deep learning has achieved significant success in the field of hyperspectral image (HSI) classification, but challenges are still faced when the number of training samples is small. Feature fusing approaches based on multi-channel and multi-scale feature extractions are attractive for HSI classification where few samples are available. In this paper, based on feature fusion, we proposed a simple yet effective CNN-based Dual-channel Spectral Enhancement Network (DSEN) to fully exploit the features of the small labeled HSI samples for HSI classification. We worked with the observation that, in many HSI classification models, most of the incorrectly classified pixels of HSI are at the border of different classes, which is caused by feature obfuscation. Hence, in DSEN, we specially designed a spectral feature extraction channel to enhance the spectral feature representation of the specific pixel. Moreover, a spatial–spectral channel was designed using small convolution kernels to extract the spatial–spectral features of HSI. By adjusting the fusion proportion of the features extracted from the two channels, the expression of spectral features was enhanced in terms of the fused features for better HSI classification. The experimental results demonstrated that the overall accuracy (OA) of HSI classification using the proposed DSEN reached 69.47%, 80.54%, and 93.24% when only five training samples for each class were selected from the Indian Pines (IP), University of Pavia (UP), and Salinas Scene (SA) datasets, respectively. The performance improved when the number of training samples increased. Compared with several related methods, DSEN demonstrated superior performance in HSI classification.

SPAD monitoring of saline vegetation based on Gaussian mixture model and UAV hyperspectral image feature classification

The chlorophyll content of saline vegetation can indirectly reflect salinization. Rapid and non-destructive capture of chlorophyll content of saline vegetation at a regional scale is essential for saline soil improvement and sustainability of saline agriculture. However, traditional soil–plant analyzer development (SPAD) monitoring based on SPAD-502 is carried out at the leaf scale, which does not allow rapid access to SPAD information for the whole region. In this study, we proposed that post-hyperspectral classification based on spectral differences that could contribute to enhanced estimation of SPAD in saline vegetation. To test this proposal, we partitioned the hyperspectral images using a Gaussian mixture model. Then, estimation models based on spectral partition and full-sample SPAD were developed using in situ observation data and a random forest model, respectively. Finally, the unmanned aerial vehicles (UAV) hyperspectral images were used as the input data source to digitally map the SPAD of saline vegetation in the region, using the prediction model. The results indicated that there were significant intensity and shape differences in the spectral reflectance characteristics of saline vegetation under different clusters. The SPAD prediction model, based on spectral feature partition, performed significantly better than the full sample. The SPAD maps of saline vegetation before and after clustering displayed similar spatial distribution models, but the prediction uncertainty of the models, based on spectral feature partition, was relatively low. Our results confirm the effectiveness and stability of UAV hyperspectral and spectral partition-based modeling in developing SPAD spatial distribution estimation models for saline vegetation.

Self-Supervised Assisted Semi-Supervised Residual Network for Hyperspectral Image Classification

Due to the scarcity and high cost of labeled hyperspectral image (HSI) samples, many deep learning methods driven by massive data cannot achieve the intended expectations. Semi-supervised and self-supervised algorithms have advantages in coping with this phenomenon. This paper primarily concentrates on applying self-supervised strategies to make strides in semi-supervised HSI classification. Notably, we design an effective and a unified self-supervised assisted semi-supervised residual network (SSRNet) framework for HSI classification. The SSRNet contains two branches, i.e., a semi-supervised and a self-supervised branch. The semi-supervised branch improves performance by introducing HSI data perturbation via a spectral feature shift. The self-supervised branch characterizes two auxiliary tasks, including masked bands reconstruction and spectral order forecast, to memorize the discriminative features of HSI. SSRNet can better explore unlabeled HSI samples and improve classification performance. Extensive experiments on four benchmarks datasets, including Indian Pines, Pavia University, Salinas, and Houston2013, yield an average overall classification accuracy of 81.65%, 89.38%, 93.47% and 83.93%, which sufficiently demonstrate that SSRNet can exceed expectations compared to state-of-the-art methods.

Analysis of protein denaturation, and chemical visualisation, of frozen grass carp surimi containing soluble soybean polysaccharides

SummaryThe mechanism of protein denaturation of frozen surimi enriched with soluble soybean polysaccharides (SSPS) was investigated. Near‐infrared (NIR) and hyperspectral imaging (HSI) technology were used to predict protein denaturation. The fresh grass carp surimi was divided into four groups with SSPS additions of 0%, 1%, 3% and 5%, respectively, which were frozen and stored at −18 °C. Samples were examined after 0, 1, 2, 4 and 8 weeks, for salt‐soluble protein content, total SH, NIR hyperspectral image and Raman spectrum features. The results showed that addition of SSPS decelerated protein degradation. After 8 weeks of storage, the 5% SSPS addition maintained the highest salt‐soluble protein content, while for total SH, 3% and 5% SSPS addition had similar effect. Raman spectra illustrated that SSPS had the ability of maintaining an α‐helix content, as well as decreasing the exposure of polar groups, and reducing the oxidation of SH group into disulphide bonds. NIR spectra showed that the overall reflectance of frozen surimi increased with the increase in SSPS. Two partial least squares regression (PLSR) models were established after pretreatment of HSI spectral statistics and selection of characteristic wavelengths. The distribution maps were finally generated based on the simplified PLSR models.

Spectral-Spatial Attention Transformer with Dense Connection for Hyperspectral Image Classification.

In recent years, deep learning has been widely used in hyperspectral image (HSI) classification and has shown good capabilities. Particularly, the use of convolutional neural network (CNN) in HSI classification has achieved attractive performance. However, HSI contains a lot of redundant information, and the CNN-based model is limited by the receptive field of CNN and cannot balance the performance and depth of the model. Furthermore, considering that HSI can be regarded as sequence data, CNN-based models cannot mine sequence features well. In this paper, we propose a model named SSA-Transformer to address the above problems and extract spectral-spatial features of HSI more efficiently. The SSA-Transformer model combines a modified CNN-based spectral-spatial attention mechanism and a self-attention-based transformer with dense connection. The SSA-Transformer model can combine the local and global features of HSI to improve the performance of the model. A series of experiments showed that the SSA-Transformer achieved competitive classification accuracy compared with other CNN-based classification methods using three HSI datasets: University of Pavia (PU), Salinas (SA), and Kennedy Space Center (KSC).

Fusion network for local and global features extraction for hyperspectral image classification

ABSTRACT Hyperspectral image (HSI) contains hundreds of contiguous spectral bands compared with red green blue (RGB) image, making the precise identification of materials possible by capturing subtle spectral and spatial features. Owing to the special advantage in image processing, convolutional neural networks (CNNs) have been proven to be a successful architecture in HSI classification. However, due to the limitation of receptive field of fixed convolution kernel, CNNs can only extract local features of hyperspectral image. Besides, CNNs fail to mine and represent the sequence attributes of spectral bands because of the limitations of its inherent network backbone. With the emergence of vision transformer, the network can break through the limitation of receptive field and obtain the global correlation of the whole image, but it only focuses on the global information of the object and ignores the local information existing in the sequence and images. Moreover, the correlation of spectral information will be destroyed if traditional methods are directly used to convert hyperspectral images into sequence. To solve this issue, a novel convolution and vision transformer fusion network called CAVFN is devised which contains a new cube-embedding module that can reduce the loss of spectral information effectively by dividing the large HSI cube into several small cubes and encoding them into sequences. More significantly, this paper also combines 1D-CNN and 2D-CNN with vision transformer to extract the local features of sequences and patches, and combines them with global features to obtain better classification results. Finally, this paper evaluates the classification results of the proposed network on three HSI datasets by conducting extensive experiments, showing that our network outperforms other state-of-the-art methods.

DEEP FEATURE EXTRACTION BASED ON DYNAMIC GRAPH CONVOLUTIONAL NETWORKS FOR ACCELERATED HYPERSPECTRAL IMAGE CLASSIFICATION

Abstract. Deep learning has achieved impressive results on hyperspectral images (HSIs) classification. Among them, supervised learning convolutional neural networks (CNNs) and semi-supervised learning graph neural networks (GNNs) are the two main network frameworks. However, 1) the supervised learning CNN faces the problem of high model time complexity as the number of network layers deepens; 2) the semi-supervised learning GNN faces the problem of high spatial complexity due to the computation of adjacency relations. In this paper, a novel dynamic graph convolutional HSI classification method is proposed, which is called dynamic graph convolutional networks (DGCNet). We first obtain two classification features by implementing flattening and global average pooling operation on the results of the convolution layer, which fully exploits the spatial-spectral information contained in the hyperspectral data. Then the dynamic graph convolution module is applied to extract the intrinsic structural information of each patch. Finally, HSI is classified based on spatial, spectral and structural features. DGCNet uses three branches to process multiple features of HSI in parallel and is trained in a supervised learning manner. In addition, DropBlock and label smoothing regularization techniques are applied to further improve the generalization capability of the model. Comparative experiments show that our proposed algorithm is comparable with the state-of-the art supervised learning models in terms of accuracy while also significantly outperforming in terms of time.

Integrating Hybrid Pyramid Feature Fusion and Coordinate Attention for Effective Small Sample Hyperspectral Image Classification

In recent years, hyperspectral image (HSI) classification (HSIC) methods that use deep learning have proved to be effective. In particular, the utilization of convolutional neural networks (CNNs) has proved to be highly effective. However, some key issues need to be addressed when classifying hyperspectral images (HSIs), such as small samples, which can influence the generalization ability of the CNNs and the HSIC results. To address this problem, we present a new network that integrates hybrid pyramid feature fusion and coordinate attention for enhancing small sample HSI classification results. The innovative nature of this paper lies in three main areas. Firstly, a baseline network is designed. This is a simple hybrid 3D-2D CNN. Using this baseline network, more robust spectral-spatial feature information can be obtained from the HSI. Secondly, a hybrid pyramid feature fusion mechanism is used, meaning that the feature maps of different levels and scales can be effectively fused to enhance the feature extracted by the model. Finally, coordinate attention mechanisms are utilized in the network, which can not only adaptively capture the information of the spectral dimension, but also include the direction-aware and position sensitive information. By doing this, the proposed CNN structure can extract more useful HSI features and effectively be generalized to test samples. The proposed method was shown to obtain better results than several existing methods by experimenting on three public HSI datasets.

One-Shot Dense Network with Polarized Attention for Hyperspectral Image Classification

In recent years, hyperspectral image (HSI) classification has become a hot research direction in remote sensing image processing. Benefiting from the development of deep learning, convolutional neural networks (CNNs) have shown extraordinary achievements in HSI classification. Numerous methods combining CNNs and attention mechanisms (AMs) have been proposed for HSI classification. However, to fully mine the features of HSI, some of the previous methods apply dense connections to enhance the feature transfer between each convolution layer. Although dense connections allow these methods to fully extract features in a few training samples, it decreases the model efficiency and increases the computational cost. Furthermore, to balance model performance against complexity, the AMs in these methods compress a large number of channels or spatial resolutions during the training process, which results in a large amount of useful information being discarded. To tackle these issues, in this article, a novel one-shot dense network with polarized attention, namely, OSDN, was proposed for HSI classification. More precisely, since HSI contains rich spectral and spatial information, the OSDN has two independent branches to extract spectral and spatial features, respectively. Similarly, the polarized AMs contain two components: channel-only AMs and spatial-only AMs. Both polarized AMs can use a specially designed filtering method to reduce the complexity of the model while maintaining high internal resolution in both the channel and spatial dimensions. To verify the effectiveness and lightness of OSDN, extensive experiments were carried out on five benchmark HSI datasets, namely, Pavia University (PU), Kennedy Space Center (KSC), Botswana (BS), Houston 2013 (HS), and Salinas Valley (SV). Experimental results consistently showed that the OSDN can greatly reduce computational cost and parameters while maintaining high accuracy in a few training samples.

Hyperspectral image feature region of solution composition analysis method based on multidimensional spectra

Redundant information is inevitable when acquisition of high-dimension spectra, which contribute to the overfitting.In order to effectively reduce the dimensions of multi-dimensional spectral data and improve the accuracy of spectral detection of complex solution components, a method based on the gradient distribution of the multi-dimensional spectral image is proposed in this paper, aiming to extract the feature region, which is highly correlated with the solution concentration. In this research, firstly, photon propagation simulations in the wedge sample were performed using Monte Carlo, and the image of multi-dimensional spectra under different optical parameters was obtained. Furthermore, a method of extracting the position of the feature region based on the gradient distribution is proposed, according to the distribution characteristics of the image. Subsequently, a series of experiments were carried out of the visible-near-infrared spectra of 39 groups of mixed solutions, which composed of India Ink and Intralipid of sequential changed concentrations, to reveal the effectiveness of the feature region. The regression models for prediction of Intralipid concentration were developed by partial least square regression (PLSR). In contrast to the traditional spectral analysis method, the dimensionality of data was reduced to 1/9, by extracting the feature region of the multi-dimensional spectral data, while the average correlation coefficient (Rp) of prediction set reached 98.87%, and the average mean square error (RMSEP) of prediction set was 0.1502, showing an increase of 0.12% in Rp and a decrease of 6.12% in RMSEP, respectively. The results showed that models incorporating the feature region, improved the accuracy of component analysis of complex solutions with high-concentration and also the analysis speed and robustness of the model. A rapid, non-invasive, low-cost method is proposed in this paper, using visible-near-infrared spectroscopy with the wedge sample to analyze the components of complex solutions.

Classification of hyperspectral image based on dual-branch feature interaction network

ABSTRACT Hyperspectral image (HSI) classification method based on convolutional neural network (CNN) has been widely used and has achieved remarkable results. However, the convolution operation is good at extracting local features, it is difficult to capture the global representation. Recently, Vision Transformer (ViT) shines with cascading self-attention modules that can aggregate global representation between patch embeddings. The self-attention module cascaded in ViT can capture long-distance feature dependencies, but it is prone to ignore the local feature details. In order to address those series of problems, this article proposes a dual-branch feature interaction (DBFI) network based on CNN and ViT, which allows global information and local feature to fully interact to enhance feature representation capabilities. To resolve the problem of the semantic gap between the two branches, a feature interaction unit (FIU) based on the attention mechanism is designed as an information interactive bridge to connect the two branches. The features are branched from ViT to CNN through the upsampling stream to enhance the global perception of the network. Similarly, the down-sampling stream is used as a coupling to connect the CNN branch to the Transformer branch to supplement the local detailed features of the network. In addition, the channel attention and spectral attention in FIU can further emphasize the distinguishability of HSI features and improve the performance of network. Experiments are conducted on three benchmark HSI datasets demonstrate that the DBFI network is better than other classification methods.