查看: 4360|回复: 5

【行空板】实时背景替换

本帖最后由云天于 2022-5-11 09:26 编辑

由于许多原因，视频的背景需要修改，如背景中有很多其他中断或背景颜色不适合该人。因此，我们使用实时背景替换技术来替换背景并添加替换为所需内容。
流行的背景去除技术
图像剪切路径 - 如果图像的主题具有锐利的边缘，则使用此技术。所有落在路径之外的元素都将被消除。

图像剪切 – 在这里，我们剪切帧中所需的区域或主题并删除背景。

图像遮罩 – 如果图像有褶边或细边缘，我们可以使用图像遮罩技术。

擦除背景 – 使用任何不同的工具擦除图像的背景

许多著名的应用程序使用背景去除技术并用自定义技术替换它。在这里，我们将实现类似的东西，使用 OpenCV 和 MediaPipe 。

MediaPipe 库可以让我们轻松运行像人脸检测、手部跟踪、姿势估计等，以及图像处理和其他 AI 功能。
【安装MediaPipe 库】
安装Mediapipe库与OpenCV库，参考AI人工智能应用：https://wiki.unihiker.com/ai_project

【单背景测试】


import cv2
import mediapipe as mp
import numpy as np
mp_drawing = mp.solutions.drawing_utils
mp_selfie_segmentation = mp.solutions.selfie_segmentation

# For webcam input:
BG_COLOR = (192, 192, 192) # gray
cap = cv2.VideoCapture(0)

cap.set(cv2.CAP_PROP_FRAME_WIDTH, 320)
cap.set(cv2.CAP_PROP_FRAME_HEIGHT, 240)
cap.set(cv2.CAP_PROP_BUFFERSIZE, 1)

cv2.namedWindow('MediaPipe Selfie Segmentation',cv2.WND_PROP_FULLSCREEN)    #Set the windows to be full screen.
cv2.setWindowProperty('MediaPipe Selfie Segmentation', cv2.WND_PROP_FULLSCREEN, cv2.WINDOW_FULLSCREEN)    #Set the windows to be full screen.

with mp_selfie_segmentation.SelfieSegmentation(
    model_selection=1) as selfie_segmentation:
  bg_image = cv2.imread("back.png")
  while cap.isOpened():
    success, image = cap.read()
    if not success:
      print("Ignoring empty camera frame.")
      # If loading a video, use 'break' instead of 'continue'.
      continue

    # Flip the image horizontally for a later selfie-view display, and convert
    # the BGR image to RGB.
    image = cv2.cvtColor(cv2.flip(image, 1), cv2.COLOR_BGR2RGB)
    # To improve performance, optionally mark the image as not writeable to
    # pass by reference.
    image.flags.writeable = False
    results = selfie_segmentation.process(image)

    image.flags.writeable = True
    image = cv2.cvtColor(image, cv2.COLOR_RGB2BGR)

    # Draw selfie segmentation on the background image.
    # To improve segmentation around boundaries, consider applying a joint
    # bilateral filter to "results.segmentation_mask" with "image".
    condition = np.stack(
      (results.segmentation_mask,) * 3, axis=-1) > 0.1
    # The background can be customized.
    #   a) Load an image (with the same width and height of the input image) to
    #      be the background, e.g., bg_image = cv2.imread('/path/to/image/file')
    #   b) Blur the input image by applying image filtering, e.g.,
    #      bg_image = cv2.GaussianBlur(image,(55,55),0)
    if bg_image is None:
      bg_image = np.zeros(image.shape, dtype=np.uint8)
      bg_image[:] = BG_COLOR
    
    output_image = np.where(condition, image, bg_image)
    output_image = cv2.rotate(output_image, cv2.ROTATE_90_COUNTERCLOCKWISE)
    cv2.imshow('MediaPipe Selfie Segmentation', output_image)
    if cv2.waitKey(5) & 0xFF == 27:
      break
cap.release()
复制代码

【按键A、B测试】


from pinpong.extension.unihiker import *
from pinpong.board import Board,Pin


Board().begin()

while True:
  if (button_a.is_pressed()==True):
    buzzer.play(buzzer.DADADADUM,buzzer.Once)
  if (button_b.is_pressed()==True):
    buzzer.play(buzzer.ENTERTAINER,buzzer.Once)
复制代码

【准备多背景】现在在这里创建项目目录中的文件夹，我正在创建一个名为*'BackgroundImages'*的文件夹。你可以下载任何图像或任意数量的图像并将它们放在此目录中。

【完整程序】

这里我们将获得背景替换图像或帧的输出。然后使用一个简单的 if 语句，分配键来更改背景。

例如，如果我们有 5 张背景图像，根据上面的代码，我们可以使用键“A”或键“B”来更改帧的背景。


import cv2
import mediapipe as mp
import numpy as np
import os
from pinpong.extension.unihiker import *
from pinpong.board import Board,Pin

Board().begin()
mp_drawing = mp.solutions.drawing_utils
mp_selfie_segmentation = mp.solutions.selfie_segmentation

# For webcam input:
BG_COLOR = (192, 192, 192) # gray
cap = cv2.VideoCapture(0)

cap.set(cv2.CAP_PROP_FRAME_WIDTH, 320)
cap.set(cv2.CAP_PROP_FRAME_HEIGHT, 240)
cap.set(cv2.CAP_PROP_BUFFERSIZE, 1)

cv2.namedWindow('MediaPipe Selfie Segmentation',cv2.WND_PROP_FULLSCREEN)    #Set the windows to be full screen.
cv2.setWindowProperty('MediaPipe Selfie Segmentation', cv2.WND_PROP_FULLSCREEN, cv2.WINDOW_FULLSCREEN)    #Set the windows to be full screen.
listImg = os.listdir("./BackgroundImages")
imgList = []
for imgPath in listImg:
    img = cv2.imread(f'BackgroundImages/{imgPath}')
    imgList.append(img)
 
indexImg = 0

with mp_selfie_segmentation.SelfieSegmentation(
    model_selection=1) as selfie_segmentation:
  
  while cap.isOpened():
    success, image = cap.read()
    if not success:
      print("Ignoring empty camera frame.")
      # If loading a video, use 'break' instead of 'continue'.
      continue

    # Flip the image horizontally for a later selfie-view display, and convert
    # the BGR image to RGB.
    image = cv2.cvtColor(cv2.flip(image, 1), cv2.COLOR_BGR2RGB)
    # To improve performance, optionally mark the image as not writeable to
    # pass by reference.
    image.flags.writeable = False
    results = selfie_segmentation.process(image)

    image.flags.writeable = True
    image = cv2.cvtColor(image, cv2.COLOR_RGB2BGR)

    # Draw selfie segmentation on the background image.
    # To improve segmentation around boundaries, consider applying a joint
    # bilateral filter to "results.segmentation_mask" with "image".
    condition = np.stack(
      (results.segmentation_mask,) * 3, axis=-1) > 0.1
    # The background can be customized.
    #   a) Load an image (with the same width and height of the input image) to
    #      be the background, e.g., bg_image = cv2.imread('/path/to/image/file')
    #   b) Blur the input image by applying image filtering, e.g.,
    #      bg_image = cv2.GaussianBlur(image,(55,55),0)
    if (button_a.is_pressed()==True):
      if indexImg>0:
            indexImg -=1

    if (button_b.is_pressed()==True):
      if indexImg<len(imgList)-1:
            indexImg +=1

    
    output_image = np.where(condition, image, imgList[indexImg])
    output_image = cv2.rotate(output_image, cv2.ROTATE_90_COUNTERCLOCKWISE)
    cv2.imshow('MediaPipe Selfie Segmentation', output_image)
    if cv2.waitKey(5) & 0xFF == 27:
      break
cap.release()
复制代码