Designment and Assessment of Far-Red Solar-Induced Chlorophyll Fluorescence Retrieval Method for the Terrestrial Ecosystem Carbon Inventory Satellite

Solar-Induced Chlorophyll Fluorescence (SIF),which is emitted by photosystem during photosynthesis under natural illumination,carries important information of actual photosynthesis of plants.Spaceborne remote sensing of SIF provides an unprecedented opportunity for monitoring global photosynthesis at regional to global scales.Up to date,in-orbit operational spaceborne sensors that are available for SIF retrieval are originally designed for atmosphere monitoring.The hyperspectral sensor onboard Chinese Terrestrial Ecosystem Carbon Inventory Satellite (CTECS) is expected to be the first operational spaceborne sensor that is specifically designed for sensing SIF from space (scheduled to be launched around 2020,2 years before the Fluorescence Explorer (FLEX) Mission).Data-driven approach has been selected as the main algorithm for far-red SIF retrieval for CTECS,but is to be refined and assessed according to sensor specifications (e.g.spectral resolution and signal-to-noise ratio).In this context,this study aims to improve the designment of far-red SIF retrieval method for CTECS.based on end-to-end simulation,we evaluate the precision and accuracy of SIF retrieval in several potential windows.We then analyze the sensitivity of SIF retrieval to number of features (nv) and fluorescence spectral shape function (hF) in the forward model in different windows.Results show that a broader fitting window increases retrieval precision,but is accompanied with lower accuracy and stronger sensitivity to nv and hF.Considering both retrieval precision and accuracy,the window of 735~758 nm with nv set to 6 and hFset as single peak Gaussian function (μ=740 nm and σ=30 nm) is selected as optimal fitting window for CTECS.SIF retrieval results based on proximal and airborne remote sensing data demonstrate the feasibility and reasonability of the designed method.Our results provide an important reference for far-red SIF retrieval for CTECS.     

Hyperspectral remote sensing IQA via learning multiple kernels from mid-level features

Hyperspectral image quality assessment (HIQA) is an indispensable technique in both academic and industry domain However, HIQA is still a challenging task since those fine-grained and quality-aware visual details are difficult to be captured. Compared with the conventional low-level features, mid-level features usually contain more semantic and quality clues and exhibit higher discriminant ability. Thus, we aim to leverage the mid-level features for HIQA. More specifically, three-scale superpixel mosaics are generated from the input image pre-processed by PCA. Each superpixel scale corresponds to various homogeneousobject parts. Subsequently, three mid-level visual features (fisher vector, combined mean features, reconstructed image matrix) as well as deep features of hyperspectral images are calculated with three-scale superpixel images to constitute multiple kernels. Afterwards, we integrate these kernels into a multimodal one, which is further integrated into a feature vector by row-wise stacking. The image quality evaluation can be calculated based on the designed similarity metric. Comprehensive experiments have demonstrated the effectiveness of our proposed HIQA algorithm.     

Hyperspectral image super-resolution combining with deep learning and spectral unmixing

In recent years, hyperspectral image super-resolution has attracted the attention of many researchers and has become a hot topic in the field of computer vision. However, it is difficult to obtain high-resolution images due to imaging hardware devices. At present, many existing hyperspectral image super-resolution methods have not achieved good results. In this paper, we propose a hyperspectral image super-resolution method combining with deep residual convolutional neural network (DRCNN) and spectral unmixing. Firstly, the spatial resolution of the image is enhanced by learning a priori knowledge of natural images. The DRCNN reconstructs high spatial resolution hyperspectral images by concatenating multiple residual blocks, each containing two convolutional layers. Secondly, the spectral features of low-resolution and high-resolution hyperspectral images are linked by spectral unmixing. This approach aims to obtain the endmember matrix and the abundance matrix. The final reconstruction result is obtained by multiplying the endmember matrix and the abundance matrix. In addition, in order to improve the visual effect of the reconstructed image, the total variation regularity is used to impose constraints on the abundance matrix to enhance the relationship between the pixels. The experimental results of remote sensing data based on ground facts show that the proposed method has good performance and preserves spatial information and spectral information without the need for auxiliary images.     

Hyperspectral Multiple Features Optimization Using Improved Firefly Algorithm

The utilization of hyperspectral remote sensing image is mainly based on the spectral information,and the spatial information is always be ignored.To solve this problem,a novel hyperspectral multiple features optimization approach based on improved firefly algorithm is presented.Firstly,four spatial features,the local statistical features,gray level co-occurrence matrix features,Gabor filtering features and morphological features of hyperspectral remote sensing image are extracted,and some spectral bands are selected and then combined with these spatial features,and the feature set is constructed.Then,the firefly algorithm is used to optimize the extracted features.In view of the slow convergence speed of firefly algorithm,we use the random inertia weight from particle swarm optimization algorithm to modifiy the location update formula of firefly algorithm,and JM(Jeffreys-Matusita)distance and Fisher Ratio are used as the objective function.Two urban hyperspectral datasets are used for performance evaluation,and the classification results derived from spectral information and spectral-spatial information are compared.The experiments show that random inertia weight can improve the speed of FA-based feature selection algorithm,the performance with multiple features is better than that of spectral information for urban land cover classification,The statistical results of the two sets of experimental data indicate that the selected number of morphological features are the most in the four spatial features.The local statistical features and morphological features are more helpful to the classification of hyperspectral remote sensing images than GLCM and Gabor features.     

Land Cover Extraction in the Ejina Oasis by Hyperspectral Remote Sensing

The extraction of land surface coverage is the basis of ecological environment evaluation,vegetation change analysis and regional ecological and hydrological processes.Aerial hyperspectral remote sensing has great advantage in land surface coverage extraction,such as flexible,wide coverage,high spatial resolution and high spectral resolution.Research area has landscape characteristics of vegetation,landscape fragmentation and heterogeneity in Ejina Poplar Forest National Nature Reserve.Comparison and analysis of two methods of dimension reduction based on minimum noise transform and principal component analysis,three supervised classification methods based on maximum likelihood method,support vector machine and object\|oriented classification.Land surface coverage is extracted by NDVI threshold segmentation,minimum noise transform dimensionality reduction method and maximum likelihood classification method according to the characteristics of landscape fragmentation,heterogeneity and high redundancy of hyperspectral data based on the Airborne Hyperspectral Data of Ejina oasis in the lower reaches of Heihe.The land surface coverage results overall accuracy and Kappa coefficient are 87.95% and 0.885 by random sampling based on airborne remote sensing data.The results show that the classification results of high accuracy can provide effective parameters for ecological research.     

A bottom-up and top-down human visual attention approach for hyperspectral anomaly detection

Hyperspectral anomaly detection (HAD) is a branch of target detection which tries to locate pixels that are spectrally or spatially different from their background. In this paper, a visual attention approach is developed to leverage HAD. Traditional HAD methods often try to locate anomalous pixels based on spectral information. However, the spatial features of hyperspectral datasets provide valuable information. Here, we aim to fuse spatial and spectral anomaly features based on bottom-up (BU) and top-down (TD) visual attention mechanisms. Owe to the BU attention, spatial features are extracted by mimicking the primary visual cortex neurons functionality. Also, spectral information is obtained throughout a deep neural network that imitating the TD visual attention. The BU and TD approaches’ results are then integrated to provide both spectral and spatial information. The key findings of our results demonstrate the proposed method outperforms the six state-of-the-art AD methods based on different evaluation metrics.     

Dual link distributed source coding scheme for the transmission of satellite hyperspectral imagery

Traditional lossless compression methods for satellite hyperspectral imagery focus on exploiting spatial and/or spectral redundancy. Those methods do not consider the temporal redundancy between images of the same area that are captured at different times. To exploit the temporal redundancy between hyperspectral images and reduce the amount of information to be transmitted from the space-satellite to the ground station via the downlink, this paper introduces a dual link distributed source coding (DLDSC) scheme for hyperspectral space-satellite communication. The proposed scheme employs the space-satellite dual link (i.e., the downlink and the uplink). The satellite onboard uses some side information from the ground station to calculate the hyperspectral image band coset values, and then, without syndrome coding, transmits to the ground station via the downlink. Coset coding is a typical technique used in distributed source coding (DSC), and here the coset values represent the timely hyperspectral image details. Typically, the coset values have lower entropy than that of the original source values. To exploit the temporal redundancy, the side information is computed in the ground station using the image captured at the previous time for the same area and transmitted to the space-satellite via the uplink. Hyperspectral images from the Hyperion satellite are used for the validation of the proposed scheme. The experimental results indicate that the proposed DLDSC scheme can reduce the original signal entropy by approximately 3.2 bits per sample (bps) and can achieve up to 1.0 bps and 1.6 bps gains over the lossless JPEG2000 standard and the state-of-art predictive CCSDS-123 method, respectively.     

PoNet: A universal physical optimization-based spectral super-resolution network for arbitrary multispectral images

Spectral super-resolution is a very important technique to obtain hyperspectral images from only multispectral images, which can effectively solve the high acquisition cost and low spatial resolution of hyperspectral images. However, in practice, multispectral channels or images captured by the same sensor are often with different spatial resolutions, which brings a severe challenge to spectral super-resolution. This paper proposed a universal spectral super-resolution network based on physical optimization unfolding for arbitrary multispectral images, including single-resolution and cross-scale multispectral images. Furthermore, two new strategies are proposed to make full use of the spectral information, namely, cross-dimensional channel attention and cross-depth feature fusion. Experimental results on five data sets show superiority and stability of PoNet addressing any spectral super-resolution situations.     

Real-time pixel based early apple bruise detection using short wave infrared hyperspectral imaging in combination with calibration and glare correction techniques

High speed data processing for online food quality inspection using hyperspectral imaging (HSI) is challenging as over hundred spectral images have to be analyzed simultaneously. In this study, a real-time pixel based early apple bruise detection system based on HSI in the shortwave infrared (SWIR) range has been developed. This systems consists of a novel, homogeneous SWIR illumination unit and a line scan camera. The system performance was tested on Jonagold apples bruised less than two hours before scanning. Partial least squares-discriminant analysis was used to discriminate bruised pixel spectra from sound pixel spectra. As the glossiness of many fruit and vegetables limits the accuracy in the detection of defects, several reflectance calibrations and pre-processing techniques were compared for glare correction and maximizing the signal to noise ratio. With the best combination of first derivative and mean centering, followed by image post-processing, this system was able to detect fresh bruises in thirty apples with 98% accuracy at the pixel level with a processing time per apple below 200 ms.