Search

In this letter, we investigate fine-grained classification of arbitrary-oriented ships in very high resolution optical imagery using Bag of Word model with Spatial Pyramid (SP-BoW). Given that base...

When floods hit populated areas, quick detection of flooded areas is crucial for initial response by local government, residents, and volunteers. Space-borne polarimetric synthetic aperture radar (PolSAR) is an authoritative data sources for flood mapping since it can be acquired immediately after a disaster even at night time or cloudy weather. Conventionally, a lot of domain-specific heuristic knowledge has been applied for PolSAR flood mapping, but their performance still suffers from confusing pixels caused by irregular reflections of radar waves. Optical images are another data source that can be used to detect flooded areas due to their high spectral correlation with the open water surface. However, they are often affected by day, night, or severe weather conditions (i.e., cloud). This paper presents a convolution neural network (CNN) based multi-modal approach utilizing the advantages of both PolSAR and optical images for flood mapping. First, reference training data is retrieved from optical images by manual annotation. Since clouds may appear in the optical image, only areas with a clear view of flooded or non-flooded are annotated. Then, a semi-supervised polarimetric-features-aided CNN is utilized for flood mapping using PolSAR data. The proposed model not only can handle the issue of learning with incomplete ground truth but also can leverage a large portion of unlabelled pixels for learning. Moreover, our model takes the advantages of expert knowledge on scattering interpretation to incorporate polarimetric-features as the input. Experiments results are given for the flood event that occurred in Sendai, Japan, on 12th March 2011. The experiments show that our framework can map flooded area with high accuracy (F1 = 96.12) and outperform conventional flood mapping methods.

Since 1972, LandSat program has experienced six successful missions that have contributed to nearly 40 years record of Earth Observations for monitoring the land cover and change dynamics. LandSat images have provided new dataset for field monitoring with geospatial information at high spatial resolution and become one of the most widely used sources of satellite images in various domains. The LandSat 8 generation launched in February 2013 continues the mission of collecting images of the Earth with an open and free data policy. However, clouds are often obscure the detection of land surface and restrict the analysis. The paper proposes improvements for an effective method of cloud detection that widely used in LandSat 5-7 in order to obtain better results for the new satellite generation LandSat 8’s images. The validation demonstrates that cloud and cloud contaminated pixels were almost detected with overall recall, precision, accuracy and error of over 98.59, 99.94, 99.89 and 0.11%, respectively on testing datasets. In comparison with the original method, our approach is able to detect 13.64% more cloud pixels with a lower error of 1.03% and 1.04 % higher accuracy rate.