python实现图像外边界跟踪

share一些python实现的code

#!/usr/bin/env python
#coding=utf-8

import cv2

img = cv2.imread("trace_border2.bmp")
[img_h, img_w, img_channel] = img.shape

trace = []
start_x = 0
start_y = 0

gray = img[:,:,1]
for h in range(img_h):
    for w in range(img_w):
        if (gray[h,w] > 128):
            gray[h,w] = 255
        else:
            gray[h,w] = 0

#python 跳出多重循环
#https://www.cnblogs.com/xiaojiayu/p/5195316.html
class getoutofloop(Exception): pass
try:
    for h in range(img_h - 2):
        for w in range(img_w - 2):
            if gray[h,w] == 0:
                start_x = w
                start_y = h
                raise getoutofloop
except getoutofloop:
    pass


print("Start Point (%d %d)"%(start_x, start_y))
trace.append([start_x, start_y])


# 8邻域 顺时针方向搜索
neighbor = [[-1,-1],[0,-1],[1,-1],[1,0],[1,1],[0,1],[-1,1],[-1,0]]
neighbor_len = len(neighbor)

#先从当前点的左上方开始，
# 如果左上方也是黑点(边界点)：
#         搜索方向逆时针旋转90 i-=2
# 否则：
#         搜索方向顺时针旋转45 i+=1
i = 0
cur_x = start_x + neighbor[i][0]
cur_y = start_y + neighbor[i][1]

is_contour_point = 0

try:
    while not ((cur_x == start_x) and (cur_y == start_y)):
        is_contour_point = 0
        while is_contour_point == 0:
            #neighbor_x = cur_x +
            if gray[cur_y, cur_x] == 0:
                is_contour_point = 1
                trace.append([cur_x, cur_y])
                i -= 2
                if i < 0:
                    i += neighbor_len
            else:
                i += 1
                if i >= neighbor_len:
                    i -= neighbor_len
            #print(i)
            cur_x = cur_x + neighbor[i][0]
            cur_y = cur_y + neighbor[i][1]
except:
    print("throw error")

for i in range(len(trace)-1):
    cv2.line(img,(trace[i][0],trace[i][1]), (trace[i+1][0], trace[i+1][1]),(0,0,255),3)
    cv2.imshow("img", img)
    cv2.waitKey(10)

cv2.rectangle(img,(start_x, start_y),(start_x + 20, start_y + 20),(255,0,0),2)
cv2.imshow("img", img)
cv2.waitKey(0)
cv2.destroyWindow("img")

搜索过程，红色标记线如下：