This study focused on the a 20-year evaluation of the dynamism of critical thermal anomalies in Busan metropolitan area prompted by unusual infrastructural development and demographic growth rate. Archived Landsat thermal data derived-LST was the major input for UTFVI and hot spot analysis (Getis-Ord Gi). Results revealed that the surface urban heat island-affected area has gradually expanded overtime from 23.32% to 32.36%; while the critical positive thermal anomalies (level-3 hotspots) have also spatially increased from 19.88% in 2000 to 23.56% in 2020, recording a net LST difference of > 5°C between the maximum level-3 hotspot and minimum level-3 coldspot each year. It is been observed that thermal conditions of Busan have gradually deteriorated with time, which is potentially inherent in the rate of urban expansion. Thus, this work serves as an eye-opener to powers that be, to think and act constructively towards a sustainable thermal conform for city dwellers.

We present quantitative estimations of changes in the areal extent of coral reef habitats at Weno Island, Micronesia, using high-spatial-resolution remote sensing images and field observations. Coral reef habitat maps were generated from Kompsat-2 satellite images for September 2008 and September 2010, yielding classifications with 78.6% and 72.4% accuracy, respectively, which is a relatively high level of agreement. The difference between the number of pixels occupied by each seabed type was calculated, revealing that the areal extent of living corals decreased by 8.2 percentage points between 2008 and 2010. This result is consistent with a comparison of the seabed types determined by field observations. This study can be used as a basis for remediation planning to diminish the impact of changes in coral reefs.

South Korea is pushing for the advancement of crop production technology to achieve food self-sufficiency and meet the demand for safe food. A medium-sized satellite for agriculture is being launched in 2023 with the aim of collecting and providing information on agriculture, not only in Korea but also in neighboring countries. The satellite is to be equipped with various sensors, though reference data for ground information are lacking. Hyperspectral remote sensing combined with 1st derivative is an efficient tool for the identification of agricultural crops. In our study, we develop a system for hyperspectral analysis of the ground-based reflectance spectrum, which is monitored seven times during the cultivation period of three soybean crops using a PSR-2500 hyperspectral sensor. In the reflection spectrum of soybean canopy, wavelength variations correspond with stages of soybean growths. The spectral reflection characteristics of soybeans can be divided according to growth into the vegetative (V) stage and the reproductive (R) stage. As a result of the first derivative analysis of the spectral reflection characteristics, it is possible to identify the characteristics of each wavelength band. Using our developed monitoring system, we observed that the near-infrared (NIR) variation was largest during the vegetative (V1-V3) stage, followed by a similar variation pattern in the order of red-edge and visible. In the reproductive stage (R1-R8), the effect of the shape and color of the soybean leaf was reflected, and the pattern is different from that in the vegetative (V) stage. At the R1 to R6 stages, the variation in NIR was the largest, and red-edge and green showed similar variation patterns, but red showed little change. In particular, the reflectance characteristics of the R1 stage provides information that could help us distinguish between the three varieties of soybean that were studied. In the R7-R8 stage, close to the harvest period, the red-edge and NIR variation patterns and the visible variation patterns changed. These results are interpreted as a result of the large effects of pigments such as chlorophyll for each of the three soybean varieties, as well as from the formation and color of the leaf and stem. The results obtained in this study provide useful information that helps us to determine the wavelength width and range of the optimal band for monitoring and acquiring vegetation information on crops using satellites and unmanned aerial vehicles (UAVs)

This study proposes a two-stage hybrid classification model for crop classification using multi-temporal remote sensing images; the model combines feature embedding by using an autoencoder (AE) with a convolutional neural network (CNN) classifier to fully utilize features including informative temporal and spatial signatures. Long short-term memory (LSTM)-based AE (LAE) is fine-tuned using class label information to extract latent features that contain less noise and useful temporal signatures. The CNN classifier is then applied to effectively account for the spatial characteristics of the extracted latent features. A crop classification experiment with multi-temporal unmanned aerial vehicle images is conducted to illustrate the potential application of the proposed hybrid model. The classification performance of the proposed model is compared with various combinations of conventional deep learning models (CNN, LSTM, and convolutional LSTM) and different inputs (original multi-temporal images and features from stacked AE). From the crop classification experiment, the best classification accuracy was achieved by the proposed model that utilized the latent features by fine-tuned LAE as input for the CNN classifier. The latent features that contain useful temporal signatures and are less noisy could increase the class separability between crops with similar spectral signatures, thereby leading to superior classification accuracy. The experimental results demonstrate the importance of effective feature extraction and the potential of the proposed classification model for crop classification using multi-temporal remote sensing images.

Although waxy corn varieties developed after the 1980s show differences depending on development stages and conditions, studies on the characteristics of waxy corn during the growth stage are rare. The subject of this study was a field survey and unmanned aerial vehicle (UAV) image acquisition of four waxy corn varieties cultivated in Idam-ri, Gammul-myeon, Goesan-gun, Korea. The study was conducted in four stages at intervals of two weeks after planting in 2019. The growth characteristics of each of the four varieties were analyzed using growth curves obtained based on field survey and UAV imagery data. The characteristics of each growth stage of the four varieties of corn, as assessed using normalized difference vegetation index (NDVI) and plant height (P.H.) values, were as follows. The growth model was identified as a model in which three-parameter logistic (3PL) curves reflect the growth characteristics of corn well. In particular, it was found that the variations in growth rate shown by P.H. and NDVI values clearly explain the differences between corn varieties. Among the four cultivars, growth and development first occurred at the early vegetative stage in Daehakchal, followed by Mibaek 2, Miheukchal, and finally Hwanggeummatchal. The variations in P.H. and NDVI were achieved quickly and earlier in Daehakchal, followed by Mibaek 2, Hwanggeummatchal, and Miheukchal. It was confirmed that these results reflected the characteristics of the fast white-type varieties, while the black-type varieties were delayed, as in a previous study. These results reflect the resistance to lodging that affects the cultivation environment and the response characteristics to nutrients and moisture. It was confirmed that UAV accurately provides growth information that is very useful for analyzing the growth characteristics of each corn variety.

With the increasing application of UAVs in intelligent transportation systems, vehicle detection for aerial images has become an essential engineering technology and has academic research significance. In this paper, a vehicle detection method for aerial images based on the YOLOv4 deep learning algorithm is presented. At present, the most known datasets are VOC (The PASCAL Visual Object Classes Challenge), ImageNet, and COCO (Microsoft Common Objects in Context), which comply with the vehicle detection from UAV. An integrated dataset not only reflects its quantity and photo quality but also its diversity which affects the detection accuracy.
The method integrates three public aerial image datasets VAID, UAVD, DOTA suitable for YOLOv4. The training model presents good test results especially for small objects, rotating objects, as well as compact and dense objects, and meets the real-time detection requirements. For future work, we will integrate one more aerial image dataset acquired by our lab to increase the number and diversity of training samples, at the same time, while meeting the real-time requirements.

This paper introduces a method of selecting priority update areas for subdivided land cover maps by training orthoimages and serial cadastral maps in a deep learning model. For the experiment, orthoimages and serial cadastral maps were obtained from the National Spatial Data Infrastructure Portal. Based on the VGG-16 model, 51,470 images were trained on 33 subdivided classifications within the experimental area and an accuracy evaluation was conducted. The overall accuracy was 61.42%. In addition, using the differences in the classification prediction probability of the misclassified polygon and the cosine similarity that numerically expresses the similarity of the land category features with the original subdivided land cover class, the cases were classified and the areas in which the boundary setting was incorrect and in which the image itself was determined to have a problem were identified as the priority update polygons that should be checked by operators.

Since aerosol has a relatively short duration and significant spatial variation, satellite observations become more important for the spatially and temporally continuous quantification of aerosol. However, optical remote sensing has the disadvantage that it cannot detect AOD (Aerosol Optical Depth) for the regions covered by clouds or the regions with extremely high concentrations. Such missing values can increase the data uncertainty in the analyses of the Earth's environment. This paper presents a spatial gap-filling framework using a univariate statistical method such as DCT-PLS (Discrete Cosine Transform-based Penalized Least Square Regression) and FMM (Fast Matching Method) inpainting. We conducted a feasibility test for the hourly AOD product from AHI (Advanced Himawari Imager) between January 1 and December 31, 2019, and compared the accuracy statistics of the two spatial gap-filling methods. When the null-pixel area is not very large (null-pixel ratio < 0.6), the validation statistics of DCT-PLS and FMM techniques showed high accuracy of CC=0.988 (MAE=0.020) and CC=0.980 (MAE=0.028), respectively. Together with the AI-based gap-filling method using extra explanatory variables, the DCT-PLS and FMM techniques can be tested for the low-resolution images from the AMI (Advanced Meteorological Imager) of GK2A (Geostationary Korea Multi-purpose Satellite 2A), GEMS (Geostationary Environment Monitoring Spectrometer) and GOCI2 (Geostationary Ocean Color Imager) of GK2B (Geostationary Korea Multi-purpose Satellite 2B) and the high-resolution images from the CAS500 (Compact Advanced Satellite) series soon.

Synthetic Aperture Radar (SAR) at C-band is an ideal remote sensing system for crop monitoring owing to its short wavelength, which interacts with the upper parts of the crop canopy. This study evaluated the potential of dual polarimetric Sentinel-1 at C-band for monitoring rice phenology. Rice phenological variations occur in a short period. Hence, the short revisit time of Sentinel-1 SAR system can facilitate the tracking of short-term temporal morphological variations in rice crop growth. The sensitivity of SAR backscattering coefficients, backscattering ratio, and polarimetric decomposition parameters on rice phenological stages were investigated through a time-series analysis of 33 Sentinel-1 Single Look Complex images collected from 10th April to 25th October 2020 in Gimhae, South Korea. Based on the observed temporal variations in SAR parameters, we could identify and distinguish the phenological stages of the Gimhae rice growth cycle. The backscattering coefficient in VH polarisation and polarimetric decomposition parameters showed high sensitivity to rice growth. However, amongst SAR parameters estimated in this study, the VH backscattering coefficient realistically identifies all phenological stages, and its temporal variation patterns are preserved in both Sentinel-1A (S1A) and Sentinel-1B (S1B). Polarimetric decomposition parameters exhibited some offsets in successive acquisitions from S1A and S1B. Further studies with data collected from various incidence angles are crucial to determine the impact of different incidence angles on polarimetric decomposition parameters in rice paddy fields.

Coastal erosion has been a threat to coastal communities and emerged as an urgent problem. Among the coastal communities that are under perceived threat, Cotonou located in Benin, West Africa, is considered as one of the most dangerous area due to its high vulnerability. To address this problem, in 2013, the Benin authorities established seven groynes at east of Cotonou port, and two additional intermediate groynes have recently been integrated in April 2018. However, there is no quantitative analysis of groynes so far, so it is hard to know how effective they have been. To analyze effectiveness, we used optical satellite images from different time periods, especially 2004 and 2020, and then compared changes in length, width and area of shoreline in Cotonou. The study area is divided into two sectors based on the location of Cotonou port. The difference of two areas is that Sector 2 has groynes installed while Sector 1 hasn't. As result of this study, shoreline in Sector 1 showed accretion by recovering 1.20 ㎢ of area. In contrast, 3.67 ㎢ of Sector 2 disappeared due to coastal erosion, although it has groynes. This may imply that groynes helped to lessen the rate of average erosion, however, still could not perfectly stop the coastal erosion in the area. Therefore, for the next step, we assume it is recommended to study how to maximize effectiveness of groynes.