Effect of pan-sharpening multi-temporal Landsat 8 imagery for crop type differentiation using different classification techniques

Abstract

Effect of pan-sharpening multi-temporal Landsat 8 imagery for crop type differentiation using different classification techniques This study evaluates the potential of pan-sharpening multi-temporal Landsat 8 imagery for the differentiation of crops in a Mediterranean climate. Five Landsat 8 images covering the phenological stages of seven major crops types in the study area (Cape Winelands, South Africa) were acquired. A statistical pansharpening algorithm was used to increase the spatial resolution of the 30m multispectral bands to 15m. The pan-sharpened images and original multispectral bands were used to generate two sets of input features at 30 and 15 metre resolutions respectively. The two sets of spatial variables were separately used as input to decision trees (DTs), k-nearest neighbour (k-NN), support vector machine (SVM), and random forests (RT) machine learning classifiers. The analyses were carried out in both the object-based image analysis (OBIA) and pixel-based image analysis (PBIA) paradigms. For the OBIA experiments, three image segmentation scenarios were tested (good, over and under segmentation). The PBIA experiments were carried out at 30m and 15m resolutions. The results show that pan-sharpening led to dramatic (~15%) improvements in classification accuracies in both the PBIA and OBIA approaches. Compared to the other classifiers, SVM consistently produced superior results. When applied to the pansharpened imagery SVM produced an overall accuracy of nearly 96% using OBIA, while PBIA's overall accuracy was 1.63% lower. We conclude that pansharpening Landsat 8 imagery is highly beneficial for classifying agricultural fields whether an objector pixel-based approach is used.

Pan-Sharpening of Landsat-8 Images and Its Application in Calculating Vegetation Greenness and Canopy Water Contents

“…However, as we resampled the NIR band at 15 m, such limitation would not prevail in this study. Additionally, Gilbertson et al [11] applied a pansharpening algorithm such as the synthetic variable ratio (SVR) method [1,11]. However, it was unclear whether they evaluated relationships between PAN and NIR bands of Landsat-8 OLI images, which was in fact the prerequisite of such actions.…”

“…Firstly, we evaluated relationships between PAN band (0.503-0.676 µm) with a spatial resolution of 15 m and individual MS bands of Landsat-8 from blue (i.e., acquired in the range 0.452-0.512 µm), green (i.e., 0.533-0.590 µm), red (i.e., 0.636-0.673 µm), NIR (i.e., 0.851-0.879 µm), shortwave infrared-I (SWIR-I: 1.566-1.651 µm), and SWIR-II (2.107-2.294 µm) bands with a spatial resolution of 30 m. Secondly, we used the obtained relationships from the first step to determine the suitable individual MS bands to be enhanced into the spatial resolution of the PAN band. Finally, we calculated several vegetation greenness and canopy moisture indices (i.e., NDVI, EVI, NDWI-I, and NDWI-II) at 15 m spatial resolution and validated this using their equivalent-values at a spatial resolution of 30 m. Note that researchers performed pan-sharpening in order to enhance the spatial resolution of both red and NIR spectral bands and subsequently to calculate NDVI-values [10,11]. In fact, the major issue was not having an appropriate level of evaluation of the suitability of the NIR spectral band to be PAN sharpened, which was one of the considerations in the scope of this paper (i.e., first step as mentioned earlier).…”

“…In wavelet fusion, a high spatial resolution Pan image is first decomposed into a set of low resolution Pan images with corresponding wavelet coefficients (spatial details) for each level [11,28]. Individual bands for the MS image then replace the low spatial resolution Pan at the resolution level of the original MS image.…”

Developments in Landsat Land Cover Classification Methods: A Review

“…OBIA has also been applied on Landsat MSS, TM, ETM+ to classify land cover by using new machine learning techniques such as Random Forests, Nearest Neighbours and SVM [88,89]. The new Landsat images, Landsat OLI, produced good results (overall accuracy above 90%) when used with OBIA to map different land cover types such as urban areas [29] and agricultural areas [90]. While object-based land cover classification has been effective on different Landsat images, not much has been reported on the performance of OBIA on the earlier version of Landsat images (Landsat MSS) perhaps because of their lower spatial resolution, or the availability of newer imagery [91,92].…”

Performance Evaluation of Downscaling Sentinel-2 Imagery for Land Use and Land Cover Classification by Spectral-Spatial Features

“…The granules, also called tiles, are the minimum indivisible partition of a product in 100 × 100 km 2 with UTM/WGS84 projection [7]. The whole watershed is over the coverage of one tile, and two tiles acquired on 28 Figure 2 illustrates a flowchart of the methodology used in this study. In the first stage, the Sentinel-2 L1C product was geometrically unified into seven datasets at two different spatial resolutions.…”