如何将图像或视频放在轮廓线的相同位置，python + opencv答案

【问题标题】：How to put an image or video in the same position of a contour line, python + opencv如何将图像或视频放在轮廓线的相同位置，python + opencv
【发布时间】：2017-09-07 20:33:35
【问题描述】：

我有轮廓的位置，我如何将图像或视频放在轮廓的相同位置，另一方面，我可以减小两者的大小。我的代码是

import cv2
import numpy as np

#Iniciar camara
captura = cv2.VideoCapture(0)

while(1):

#Caputrar una imagen y convertirla a hsv
_, imagen = captura.read()
hsv = cv2.cvtColor(imagen, cv2.COLOR_BGR2HSV)
img=cv2.imread('calibresult.png')
#Guardamos el rango de colores hsv (azules)
bajos = np.array([100,65,75], dtype=np.uint8)
altos = np.array([130, 255, 255], dtype=np.uint8)

#Crear una mascara que detecte los colores
mask = cv2.inRange(hsv, bajos, altos)

#Filtrar el ruido con un CLOSE seguido de un OPEN
kernel = np.ones((6,6),np.uint8)
mask = cv2.morphologyEx(mask, cv2.MORPH_CLOSE, kernel)
mask = cv2.morphologyEx(mask, cv2.MORPH_OPEN, kernel)

#Difuminamos la mascara para suavizar los contornos y aplicamos filtro canny
blur = cv2.GaussianBlur(mask, (5, 5), 0)
edges = cv2.Canny(mask,1,2)

#Si el area blanca de la mascara es superior a 500px, no se trata de ruido
contours, hier = cv2.findContours(edges,cv2.RETR_TREE,cv2.CHAIN_APPROX_SIMPLE)
areas = [cv2.contourArea(c) for c in contours]
i = 0
for extension in areas:
    if extension > 600:
        actual = contours[i]#position of countour
        approx = cv2.approxPolyDP(actual,0.05*cv2.arcLength(actual,True),True)
        if len(approx)==3:
            cv2.drawContours(imagen,[actual],0,(0,0,255),2)
            cv2.drawContours(mask,[actual],0,(0,0,255),2)

        i = i+1

cv2.imshow('mask', mask)
cv2.imshow('Camara', imagen)
tecla = cv2.waitKey(5) & 0xFF
if tecla == 27:
    break

cv2.destroyAllWindows()

“实际”是轮廓的位置例如，我希望代码注册一个矩形的轮廓，并在轮廓的相同位置放置一个图像，所有这些都是实时的，位置将相对于轮廓的中心

【问题讨论】：

我仍然不明白您所说的“如何将图像或视频放在轮廓的同一位置”是什么意思。你能进一步澄清这一点吗？你有你想要的输出的可视化表示吗？
例如，我希望代码注册一个矩形的轮廓，并在轮廓的相同位置放置一个图像，所有这些都是实时的，位置将相对于中心轮廓

标签： python opencv numpy

【解决方案1】：

来自here的出色回答：

dst = cv2.imread("destination.jpg", -1)
src = cv2.imread("source.jpg", -1)

gray = cv2.cvtColor(dst, cv2.COLOR_BGR2GRAY)
ret, binary = cv2.threshold(gray, 130, 255, cv2.THRESH_BINARY)

im2,contours,hierarchy = cv2.findContours(binary, cv2.RETR_LIST, cv2.CHAIN_APPROX_NONE)

for contour in contours:
    cv2.drawContours(dst, contour, -1, (0,0,255), thickness = 2)

locs = np.where(binary != 0) # Get the non-zero mask locations

# Case #1 - Other image is grayscale and source image is colour
if len(dst.shape) == 3 and len(src.shape) != 3:
    dst[locs[0], locs[1]] = src[locs[0], locs[1], None]
# Case #2 - Both images are colour or grayscale
elif (len(dst.shape) == 3 and len(src.shape) == 3) or (len(dst.shape) == 1 and len(src.shape) == 1):
    dst[locs[0], locs[1]] = src[locs[0], locs[1]]
# Otherwise, we can't do this
else:
    raise Exception("Incompatible input and output dimensions")

【讨论】：