Research on oilcan target segmentation in remote sensing image based on improved Otsu algorithm

In order to improve the speed of oilcan target segmentation in optical remote sensing image, a fast approach is presented based on the previous improved algorithm. Gray distribution of optical remote sensing image is analyzed. The pixels which meet the priori knowledge are stored in an array. The missing gray values no longer participate in calculation so that the redundant computing is reduced greatly. The replacement function of the between-class variance function (the two functions both maximize at the same point) is transformed into a piecewise function to improve operation speed. It is proved in the experiment that this approach can be used in the segmentation of oilcans in optical remote sensing images, and the effect is better than that of Otsu. It is more than 5 times faster than previous improved algorithm for the images used in the experiment.