窗口的集合是一种逆四叉树分割的过程。在四叉树分割的过程中,分区实现当像素显示不同的属性。区域尺度的减小的过程。视窗化聚合开始所需的最小规模的四叉树分割。分离的区域合并如果研究区域内的最大区别是小于阈值集。过程描述如下。(我)标签的所有地区连续数。设置窗口的规模(W x W W×2)初始化。(2)扫描整个图像的窗口左上角来者。最大的区别在W x W窗口计算建筑面积。五种方式计算覆盖区域如图1所示。 The shadow area in figure 1 shows the possible homogenous regions. If the difference in the observed regions is less than the threshold value, combine the observed regions. Label the combined the region with minimum number of the pre-merger regions. (iii) Set W as the length of step, scan the whole image and merge similar regions. (iv) Modify the scale of window, W = W x 2, after scanning the whole image. (v) Repeat steps (ii) to (iv), until the size of window is equal to the original image size.Plz help to do this.......