8-7

遇到个要解决的问题，图片中有若干个矩形，需要将所有相交的矩形用一个大矩形将它们替代：

1、这种方法能快速找到相邻的两个矩形

            #判断patch中是否有重叠，有则合为一个，取两个的并集
            #算法思路：矩形中心坐标任一轴之差大于该轴方向的二分之一边之和； (A - 1)!
            #先进行排序，左右排序，A; 再进行相邻距离差计算，如果 ( x2 -x1 ) < w1 距离在之内，则比较上下，A*A
            i = -1
            patch = []
            extend_patch = []
            Centroid = []
            watches = sorted(watches, key=takeFrist)      #指定第一个元素进行顺序排序
            print watches
            print len(watches)
            for (x,y,w,h) in watches:
                cv2.rectangle(img, (x, y), (x + w+2, y + h+2), (0, 0, 255), 1)

            for count_ in  range(len(watches)):
                x,y,w,h = watches[count_]
                print  watches[count_]
                if count_ + 1 >= len(watches):
                    pass
                else:
                    if (watches[count_+1][0] - watches[count_][0]) > watches[count_][2]:#x方向
                        pass
                    else:
                        if abs(watches[count_+1][1] - watches[count_][1]) > min(watches[count_][3],watches[count_+1][3]):
                            pass
                        else:
                            x = x
                            w = watches[count_+1][2] + (watches[count_+1][0] - x) 
                            if y > watches[count_+1][1]:
                                h = h + (y - watches[count_+1][1])
                                y = watches[count_+1][1]
                            else:
                                y =  y
                                h = watches[count_+1][3] + (watches[count_+1][1] - y)
                            count_ += 1
                            if count_ >= len(watches) - 1:
                                count_ = len(watches) - 1
                patch.append(img_bak[y:y + h, x:x + w])
                cv2.rectangle(img, (x-2, y-2), (x + w, y + h), (0, 255, 255), 2)