Mapping of Coastal Cities Using Optimized Spectral–Spatial Features Based Multi-Scale Superpixel Classification

Abstract

The high interior heterogeneity of land surface covers in high-resolution image of coastal cities makes classification challenging. To meet this challenge, a Multi-Scale Superpixels-based Classification method using Optimized Spectral–Spatial features, denoted as OSS-MSSC, is proposed in this paper. In the proposed method, the multi-scale superpixels are firstly generated to capture the local spatial structures of the ground objects with various sizes. Then, the normalized difference vegetation index and extend multi-attribute profiles are introduced to extract the spectral–spatial features from the multi-spectral bands of the image. To reduce the redundancy of the spectral–spatial features, the crossover-based search algorithm is utilized for feature optimization. The pre-classification results at each single scale are, therefore, obtained based on the optimized spectral–spatial features and random forest classifier. Finally, the ultimate classification is generated via the majority voting of those pre-classification results in each scale. Experimental results on the Gaofen-2 image of Qingdao and WorldView-2 image of Hong Kong, China confirmed the effectiveness of the proposed method. The experiments verify that the OSS-MSSC method not only works effectively on the homogeneous regions, but also is able to preserve the small local spatial structures in the high-resolution remote sensing images of coastal cities.

Synthesis of Biochar-Based Composites to Evaluate Morphology of Wheat Seedling

“…Pixel-based classification methods are hardly applicable for high-resolution remote sensing images due to the high interior heterogeneity of land surface covers. The separation between spectral signatures of different land surface covers is more difficult due to the abundant details in pixel-based classification [93]. To deal with this challenge, we are using superpixel-based classification which reduces the redundancy of the spatial features of different ground objects.…”

Bayesian Gravitation-Based Classification for Hyperspectral Images

“…To obtain high quality classification results, complete utilization of Spectral and Spatial (SS) information is extremely important [6,7]. In the past decades, researchers have developed numerous methods to synthetically utilize the SS information of HSIs [8][9][10][11][12][13][14][15][16][17][18][19][20][21][22][23][24]. Some studies first extract spatial features using texture filters on each spectral feature, such as gray-level co-occurrence matrix (GLCM) [8], Gabor [9], local binary pattern (LBP) [10], and Markov model [11].…”

“…In addition to the aforementioned methods, some other studies combine the SS information by integrating different classifiers [12,[19][20][21][22][23][24]. One typical kind of method is the fusion of pixel-based classification and superpixel-based segmentation results by the majority voting [12,32].…”

“…In addition to the aforementioned methods, some other studies combine the SS information by integrating different classifiers [12,[19][20][21][22][23][24]. One typical kind of method is the fusion of pixel-based classification and superpixel-based segmentation results by the majority voting [12,32]. In these researches, the spectral information contributes mainly to the classification while the spatial adjacent relationships between pixels are considered to refine the results.…”