The way urban climates are classified affects both sustainable urban development and environmental planning. Local Climate Zone (LCZ) classification offers a comprehensive framework to classify different urban areas based on their climate-related characteristics. This paper investigates the application of deep learning techniques for LCZ categorization using multispectral Sentinel-2 satellite images. Sentinel-2's capacity to record optical data over a wide range of spectrum bands makes it an invaluable tool for understanding variations in urban climate. This study uses a deep learning model called convolutional neural networks (CNNs) to effectively extract and learn spatial attributes from the multispectral Sentinel images. The work uses a labeled dataset with Sentinel images for training the model and classifications of LCZ. During the training phase, the model parameters are tuned to enhance the interpretability of climate-related patterns in urban environments. Using a validation dataset, classification metrics such as accuracy, precision, recall, and F1 score are used to evaluate the model performance. These conclusions offer useful information to environmental scientists, urban planners, legislators, and those involved in climate-resilient urban design. This demonstrates the efficacy of using multispectral and SAR images for precise LCZ categorization, advancing our understanding of the variability of urban climate and assisting planners in making well-informed decisions regarding urban development strategies.
An increasing number of applications require land cover information from remote sensing images, thereby resulting in an urgent demand for automatic land use and land cover classification. Therefore, effectively improving the accuracy of land cover classification is a main objective in remote sensing image processing. We propose a land cover classification postprocessing framework based on iterative self-adaptive superpixel segmentation (LCPP-ISSS) for remote sensing image data. This framework can further optimize the land cover classification results obtained by neural networks without changing the network structure. First, we propose the iterative self-adaptive superpixel segmentation algorithm for high-resolution remote sensing images to extract the boundary information of different land cover classes. Then, we propose a land cover classification result optimization method based on patch complexity to optimize the classification result by combining the boundary information with the semantic information. In an experiment, we compare the classification accuracy before and after using LCPP-ISSS and with other common methods. The results show that LCPP-ISSS outperforms the dense conditional random field and provides a 4% increase in the mean intersection over union and a 10% increase in overall accuracy.
Dual-comb spectroscopy (DCS) is an emerging spectroscopic tool with the potential to simultaneously achieve a broad spectral coverage and ultrahigh spectral resolution with rapid data acquisition. However, the need for two independently stabilized ultrafast lasers significantly hampers the potential application of DCS. We demonstrate mode-resolved DCS in the THz region based on a free-running single-cavity dual-comb fiber laser with the adaptive sampling method. While the use of a free-running single-cavity dual-comb fiber laser eliminates the need for two mode-locked lasers and their frequency control, the adaptive sampling method strongly prevents the degradation of spectroscopic performance caused by the residual timing jitter in the free-running dual-comb laser. Doppler-limit-approaching absorption features with linewidths down to 25 MHz are investigated for low-pressure acetonitrile/air mixed gas by comb-mode-resolved THz spectroscopy. The successful demonstration clearly indicates its great potential for the realization of low-complexity, Doppler-limited THz spectroscopy instrumentation.
Terahertz dual-comb spectroscopy (THz-DCS) has the potential to be used as universal THz spectroscopy with high spectral resolution, high spectral accuracy, and broad spectral coverage; however, the requirement for dual stabilized femtosecond lasers hampers its versatility due to the bulky size, high complexity, and high cost. We here report the first demonstration of dual THz comb spectroscopy using a single free-running fiber laser. While greatly reducing the size, complexity, and cost of the laser source, THz-DCS maintains the spectroscopic performance comparable to a system equipped with dual stabilized fiber lasers, and can be effectively applied to gas spectroscopy.
Means of synthetic aperture radar (SAR) images represent the radiation densities of scenes, and the preservation of means is significant in speckle denoising for the application of SAR images. We provide an improved scheme of the minimum biased diffusion (MinBAD) algorithm for speckle denoising using partial differential equations. Considering the characteristics of SAR speckle and the radiation accuracy for postprocessing needs, several improvements such as normalization, homomorphic transformation, and average-preserving processing are introduced into the MinBAD algorithm. Besides the equivalent number of looks and edge preserving index, a new index, radiation accuracy error, is defined to evaluate the denoising effect. Experimental results for both artificial images and real SAR images are used to validate the performance of the proposed unbiased-average MinBAD speckle reducing approach.
Due to the larger orbital arc and longer synthetic aperture time in medium Earth orbit (MEO) synthetic aperture radar (SAR), it is difficult for conventional SAR imaging algorithms to achieve a good imaging result. An improved higher order nonlinear chirp scaling (NLCS) algorithm is presented for MEO SAR imaging. First, the point target spectrum of the modified equivalent squint range model-based signal is derived, where a concise expression is obtained by the method of series reversion. Second, the well-known NLCS algorithm is modified according to the new spectrum and an improved algorithm is developed. The range dependence of the two-dimensional point target reference spectrum is removed by improved CS processing, and accurate focusing is realized through range-matched filter and range-dependent azimuth-matched filter. Simulations are performed to validate the presented algorithm.
Adaptive sliding receive-window (ASRW) technique was usually introduced in airborne squint synthetic aperture radar (SAR) systems. Airborne squint spotlight SAR varies its receive-window starting time pulse-by-pulse as a function of range-walk, namely, the linear term of range cell migration (RCM). As a result, a huge data volume of the highly squint spotlight SAR echo signal can be significantly reduced. Because the ASRW technique changes the echo-receive starting time and Doppler history, the conventional image algorithm cannot be employed to directly focus airborne squint spotlight ASRW-SAR data. Therefore, a fast image-formation algorithm, based on the principle of the wave number domain algorithm (WDA) and azimuth deramping processing, was proposed for accurately and efficiently focusing the squint spotlight ASRW-SAR data. Azimuth deramping preprocessing was implemented for eliminating azimuth spectrum aliasing. Moreover, bulk compression and modified Stolt mapping were utilized for high-precision focusing. Additionally, geometric correction was employed for compensating the image distortion resulting from the ASRW technique. The proposed algorithm was verified by evaluating the image performance of point targets in different squint angles. In addition, a detailed analysis of computation loads in the appendix indicates that the processing efficiency can be greatly improved, e.g., the processing efficiency could be improved by 17 times in the 70-deg squint angle by applying the proposed image algorithm to the squint spotlight ASRW-SAR data.
A general synthetic aperture radar (SAR) signal model based on the Maxwell's equations is derived, and three approximations are discussed for engineering applications. Based on this signal model, a novel operation of SAR, called outer circular synthetic aperture radar (Outer-CSAR), is investigated for wide observation. The Outer-CSAR works similarly to the general circular SAR, but the beam of the SAR antenna points at the outer of the circle instead of the inner. The signal model and imaging algorithm are presented for the Outer-CSAR, and furthermore, simulation is given to validate the signal model and imaging algorithm.
The user requirement for the rice satellite (RICESAT) is analyzed, basing on the radar backscattering coefficient characteristic of rice crops. The orbital elements of RICESAT are selected. The payload of RICESAT was designed as C-band synthetic aperture radar (SAR) system with HH and VV polarization, operating in Strip Map mode or Scan SAR mode. The orbital elements of RICESAT are selected, and the key parameters of the SAR payload are designed. The computer simulation results are presented in the paper, which demonstrates the preliminary feasibility of RICESAT.
The orbital elements determination method for 3-demensional Formation-flying SAR satellites system is developed, basing on the closed form solution to the free orbit movements acquired from Kepler's equation. The quantificational relationship between the interferometric baseline, the orbital elements and formation configuration parameters is induced. The general baseline design approach is proposed, basing on the analysis of critical space baseline and critical time baseline. Typical configurations are selected for the L-band FF-SAR satellite systems, including circular configuration, circular horizontal projection configuration and linear horizontal projection. The efficiency of the baseline design technique is validated by the computer simulation results with measurement accuracy of terrain height and velocity.
Interferometric synthetic aperture radar(InSAR) is usually employed to provide altitude information of terrain. There are many factors to cause interferogram noise and the noise decreases height measurement accuracy of InSAR. In this paper, Multilook processing technique, which is used to reduce interferogram noise in InSAR, is studied. And a new concept - critical number of looks for multilook processing is proposed. The number of looks must be smaller than critical number of looks when multilook processing is applied to the interferogram noise suppression. The new concept benefits the multilook processing for InSAR noise suppression. With the knowledge of critical number of looks, the proper number of looks can be chosen for InSAR multilook processing to reduce the interferogram noise. The equations of range and azimuth critical numbers of looks are presented for both single baseline spaceborne InSAR and multibaseline spaceborne InSAR. The critical number of looks can be obtained from these equations. In the end, multilook processing with different numbers of looks is applied to InSAR simulated data and the results show the validity of critical number of looks in InSAR.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.