Object based segmentation method

摘要

This invention is related to a method that enables the object based segmentation of especially air/satellite images that are displayed in high resolution. The aim of the invention is to determine automatically the borders of objects by using statistical, spatial and structural relationships/characteristics and also by using high resolution air/satellite image data. Another aim of the invention is to develop a method that can operate by being minimally affected by limiting aspects such as ambient light, weather conditions or resolution and that can provide the determination of an object as a whole instead of sensing it in pixels.

主权项

1. An object based segmentation method for ensuring segmentation of objects in an image comprising;
carrying out morphological opening and closing operations until a first local maximum is obtained for each pixel by using an increasingly large disk type structuring elements; appointing the diameter of the first local maximum structuring element and the type of opening/closing operation as the attribute of that pixel; grouping all pixels with the same attributes together to obtain primitive structures in order to draw out primitive structures formed by pixel groups which have homogeneous brightness values within themselves following morphologic operations and to deduce the spatial size of the primitive structures; calculating the mean value and standard deviation of the primitive structures; matching each primitive structure with the neighboring primitive structure which has the closest standard deviation and mean value to itself, in order to establish primitive objects by measuring statistically the similarities of the primitive structures with their neighboring primitive structures and then regrouping the similar structures; determining the range value at a pixel as the standard deviation of the primitive object that contains the pixel; determining the spatial size at a pixel as the size (pixel count) of the primitive object that contains the pixel, in order to determine the range that shows change in terms of spatial size and brightness values of the primitive objects; calculating the vectoral difference in relation to the pixel of the average of the points inside the bandwidths around the pixel at the mean shift test point and the segmentation resolution determined in proportion to the size of the bandwidths used with the mean shift method which is based on estimating the density gradient; assuming that the zero convergence point corresponds to a meaningful mode inside the tested picture or an object inside the picture, grouping and determining the segments of the pixels that establish connected components which converge to the same point and which are within the range value of that said point, in order to filter the image by accepting that the spatial and range sizes of the primitive objects provide an estimate of the mean shift technique's bandwidths (range and spatial bandwidths) and to determine the preferred objects by grouping (segmenting) the pixels that belong to the same mode.