Automatic building detection based on CIE LUV color space using very high resolution pleiades images

Building detection from high resolution images has become one of the important research topics in remote sensing considering that determination of geographic location of buildings and their geometries provide very useful information for various applications such as change detection, 3D city models, urban information systems and practices in municipalities. This paper proposes a very simple but robust method for automatic detection of buildings. Our approach consists of two steps, where in the first step, multispectral image's color space is converted to 'LUV' space (where L stands for luminance, whereas U and V represent chromaticity values of color images) using visible channels of Pleiades satellite imagery namely Red, Green and Blue 'RGB' bands. Then, in the second step, Otsu's automatic gray-level thresholding, which is efficient in allocation of the bimodal histogram distribution, is applied to the components of the new 'LUV' image. The proposed method has been successfully applied to very high resolution Pleiades images to accurately identify boundaries in the selected test regions. Results of this study indicated boundaries of buildings could be determined 80% or better accuracy. Accuracy analysis was applied to all individual pixels within the test images but not only the selected pixels; therefore, precision and recall values represent almost true values but not estimation.