Search

High-quality optical data is difficult to acquire in southwest China, due to the cloudy and rainy weather. The synthetic aperture radar (SAR) can work for all-time and all-weather, but the revisit period is usually long. Therefore, it is a high demand for the synergy of optical and radar data rather than a single remote sensing data source, particularly for the survey and assessment on the land resource, as well as the global change. This study performed an object-oriented guided segmentation on the difference images generated from 2016 and 2019 quad-pol Radarsat-2 data using the 2019 Sentinel-2 optical data of Meishan region. A Fractal Net Evolution Approach (FNEA) was first utilized to segment the 2016 optical image for an initial segmentation. Then, the second segmentation was implemented using the same approach, where the features of the SAR image were provided for the purpose of guided segmentation. As such, the regions of growth were merged from small to large, in order to eventually form a complete geographical object. A final segmentation was obtained to select the changed and unchanged samples. More importantly, an object contained both changed and unchanged pixels, particularly with a single type of pixel inside the samples. Then, the object-level multidimensional features were extracted using the samples. The initial radar parameters included the backward scattering coefficient, Pauli and Freeman-Durden polarization decomposition parameters. The segmentation was utilized to generate the geographic objects. Meanwhile, some parameters were calculated using the initial parameters, including the mean, and standard deviation, whereas, the texture feature parameters using the Gray-level co-occurrence matrix (GLCM) included the mean, standard deviation, homogeneity, contrast, entropy, and the correlation for a total of 72-dimensional features. The redundant feature dataset was filtered to select the extracted feature set. Consequently, an FSO-RF change detection framework was proposed to achieve the feature optimization and the acquisition of the final change detection. The inter-sample distance metric learning was used to optimize the separability of feature space. The optimal features were used to calculate the distance between the changed and unchanged samples in different features. Finally, the change detection task was implemented using an object-oriented random forest classifier. The existing change detection pixel-based methods were selected to verify the improved model, including the change vector analysis (CVA), principal component analysis (PCA), multivariate alteration detection (MAD), Iteration multivariate alteration detection (IR-MAD), support vector machines (SVM), and random forest (RF) change detection using radar difference image segmentation. The change detection framework showed a remarkable improvement, with an overall accuracy of 92.90% in the test area. Two representative areas were selected in the study area for sampling and validation. The experiments demonstrated that better performance was achieved than that of the traditional, although the accuracy indexes were degraded in some areas with the complex features. The overall accuracies of the two sample areas reached 95.08 % and 88.16%. The algorithm in this paper can well meet the needs of detecting changes in different feature categories such as towns and farmlands, and has certain application value in the monitoring of national land resources. [ABSTRACT FROM AUTHOR]

Emergency flood relief needs fast and precise spatial information of flood-inundated area. Because of spectral similarity to some extent, inundated area was often viewed as water body and then derived from remote sensing classification result of the water areas before flood disaster minus that after flood disaster, which often caused difficulties in separating inundated area from true water body (e.g. river, lake or reservoir) and waterlogged cropland. This paper proposed an optimized method for identifying inundated area based on human perceptual visual attention and spatio-temporal variation characteristics of flood. Flood-inundated area is a temporary compound of water and other flooded objects (e.g., crops). It has specific spatio-temporal dynamic property, which differs from that of water body and could be used as a stable basis for remote sense identification of the inundated area. After the specific property was digitalized into numerical and visual data by generating a composite RGB color imagery of NDVI, MNDWI before the flood event and NDWI after the inundation, apparent coloring difference between water body and the inundated area in the imagery was reached. To gain reliable machine-vision detection based results of the flood-inundated area, an imagery of Munsell HLC color was transformed from the RGB imagery and then adopted in unsupervised classification based on coefficient of NBS color distance and K-means clustering algorithm. A case was studied on applying the proposed method to derive information of the inundated area in 2016 flood disaster in middle reaches of Yangtze River Basin and using Landsat OLI images. And an error confusion matrix method was adopted in the accuracy assessment on recognition results of the inundated area. The results showed that the proposed method gained excellent detection of the area with coefficients of Kappa and composite classification accuracy (CCA) equal to 93.4% and 88.5%, respectively. A controlled experiment on the credibility of different remote sensing classification methods for the same identification object (i.e. flooded-inundated area) was also made. It proved the proposed method for detection of flood-inundated area based on the selective visual attention mechanism could effectively improve the accuracy of remote sensing identification on the area with its CCA coefficient of 5% larger than that of the traditional method. The proposed method also had good performance on easily separating the flooded-inundated areas from water and waterlogged cropland, especially helped solving the misclassification between the flooded-inundated area and the water body. These benefited from three keys: 1) The inundated area was viewed as that not equal to water coverage in remote sensing classification, which could be proven by the former’s higher heterogeneity and its different spatio-temporal dynamic property. As a result, the identification uncertainty of the area was reduced. 2) Digital visualization of spatio-temporal variation features of the inundated area was completed through data fusion of NDWI after the inundation and NDVI and MNDWI before the flood event, which enhanced the area’s visual saliency on the remote sensing image and made it possible to identify them using machine-vision color clustering. 3) Difference of the flooded-inundated areas from the water and waterlogged cropland was precisely captured in the NBS image in HLC space that was transformed from the fusion image at an optimized RGB color matching scheme (i.e. red given to NDWI after the inundation, green and blue to NDVI and MNDWI before the inundation, respectively). During summer flooding of the study area in 2016, the waterlogged farmland with crops uncovered by water was about 19 143.35 hm² while the inundated acreage was up to 142 157.5 hm², among which 16.6% (i.e. 23 579 hm²) was being planted with rice paddy. Additionally, most of the inundated areas located in low-lying land along mainstream of the middle Yangtze River and its two branches, namely Hanjiang River and Fuhe River. The most severe flood disaster occurred in Wuhan City with 34 492 hm² of the inundated acreage. [ABSTRACT FROM AUTHOR]