如何从边界框获取图像

问题描述

如何从这张图片中提取边框？

这是我的代码

tf.disable_v2_behavior()
import cv2
import pathlib

model_path = "/Users/ebayb/Desktop/ndir/Model"

# path to the folder containing images
image_path = "/Users/ebayb/Desktop/top-cube/testing/2"

session = tf.Session(graph=tf.Graph())

tf.saved_model.loader.load(session,['serve'],model_path)

# extract the coordinates of the rectange that's to be drawn on the image
def draw_Boxes(height,width,Box,img):
  # starting coordinates of the Box
  ymin = int(max(1,(Box[0] * height)))
  xmin = int(max(1,(Box[1] * width)))
  # last coordinates of the Box
  ymax = int(min(height,(Box[2] * height)))
  xmax = int(min(width,(Box[3] * width)))
  coordinates_list=[ymin,ymax,xmin,xmax]
  print(f"cor: {coordinates_list}")
  # draw a rectange using the coordinates
  cv2.rectangle(img,(xmin,ymin),(xmax,ymax),(10,255,0),2)
  

for file in pathlib.Path(image_path).iterdir():
  # get the current image path
  current_image_path = r"{}".format(file.resolve())
  
  
  img_bytes = open(current_image_path,'rb').read()
  
  
  result = session.run(['detection_Boxes:0','detection_scores:0'],Feed_dict={'encoded_image_string_tensor:0': [img_bytes]})
  
  Boxes = result[0][0]
  scores = result[1][0]
  
  print("For file {}".format(file.stem))

 
  img = cv2.imread(current_image_path)
  
  imH,imW,_ = img.shape
               
  for i in range(len(scores)):
    # only consider a detected object if it's probability is above 25%
    if scores[i] > 0.258:
      print("The Box {} has probability {}".format(Boxes[i],scores[i]))
      draw_Boxes(imH,Boxes[i],img)
      
      
      
  
  

  cv2.imwrite("bound.jpg",img)

我想知道如何从中获得边界框？我尝试从其他线程的其他答案中使用ROI，但这似乎不起作用，它只会获取其他图像

（投资回报率示例） https://gyazo.com/5478ff4715311c0422d1dfbad7f03d18

以下是ROI的代码：

import numpy as np

# Load image,grayscale,Otsu's threshold 
image = cv2.imread('test.jpg')
original = image.copy()
gray = cv2.cvtColor(image,cv2.COLOR_BGR2GRAY)
thresh = cv2.threshold(gray,100,cv2.THRESH_BINARY_INV + cv2.THRESH_OTSU)[1]

# Find contours,obtain bounding Box,extract and save ROI
ROI_number = 0
cnts = cv2.findContours(thresh,cv2.RETR_EXTERNAL,cv2.CHAIN_APPROX_SIMPLE)
cnts = cnts[0] if len(cnts) == 2 else cnts[1]
for c in cnts:
    x,y,w,h = cv2.boundingRect(c)
    cv2.rectangle(image,(x,y),(x + w,y + h),(36,12),2)
    ROI = original[y:y+h,x:x+w]
    cv2.imwrite('ROI_{}.png'.format(ROI_number),ROI)
    ROI_number += 1

cv2.imshow('image',image)
cv2.waitKey()

这是我正在使用的测试图像

解决方法

您已经在第一个代码中绘制了矩形。如果您想要带有矩形的区域，请修改第一个代码以使用xmin,ymin,xmax,ymax值裁剪原始图像。

对于绘制的每个满足scores[i] > 0.258的矩形，请使用以上值在draw_boxes函数中裁剪并保存原始图像的一部分。那就不用轮廓了。

如果有重叠的边界框，请使用Non-max Suppression。