import cv2
import numpy as np
def calculate_overlap_percentage(a_image_path, b_image_path):
# 读取A组和B组图像
a_image = cv2.imread(a_image_path)
b_image = cv2.imread(b_image_path)
# 将图像从BGR转为HSV色彩空间，便于颜色筛选
a_hsv = cv2.cvtColor(a_image, cv2.COLOR_BGR2HSV)
b_hsv = cv2.cvtColor(b_image, cv2.COLOR_BGR2HSV)
# 设置B组红色的HSV范围
lower_red = np.array([0, 100, 100])
upper_red = np.array([10, 255, 255])
# 提取B组红色区域的掩膜
b_red_mask = cv2.inRange(b_hsv, lower_red, upper_red)
# 设置A组颜色区域的HSV范围（棕色和绿色）
# 棕色的HSV范围
lower_brown = np.array([10, 30, 30])
upper_brown = np.array([25, 255, 200])
# 绿色的HSV范围
lower_green = np.array([35, 50, 50])
upper_green = np.array([85, 255, 255])
# 提取A组棕色和绿色区域的掩膜
a_brown_mask = cv2.inRange(a_hsv, lower_brown, upper_brown)
a_green_mask = cv2.inRange(a_hsv, lower_green, upper_green)
# 合并A组的棕色和绿色区域掩膜
a_color_mask = cv2.bitwise_or(a_brown_mask, a_green_mask)
# 计算B组红色区域的像素数量
b_red_area = cv2.countNonZero(b_red_mask)
# 在A组中找到与B组红色区域重合的区域
a_overlap = cv2.bitwise_and(a_color_mask, b_red_mask)
# 计算A组与B组红色区域重合的像素数量
a_overlap_area = cv2.countNonZero(a_overlap)
# 计算A组有颜色区域占B组红色区域的百分比
overlap_percentage = (a_overlap_area / b_red_area) * 100 if b_red_area != 0 else 0
return overlap_percentage
# 示例调用
a_image_path = "path_to_a_image.jpg"
b_image_path = "path_to_b_image.jpg"
percentage = calculate_overlap_percentage(a_image_path, b_image_path)
print(f"A组颜色区域占B组红色区域的百分比为: {percentage:.2f}%")
下面查看提取的mask是否正确
# 调整显示图像的大小（设置为目标宽度 600 像素）
height, width = b_red_mask.shape
target_width = 600
target_height = int(target_width * height / width)
# 使用 cv2.resize 调整图像大小
b_red_mask_resized = cv2.resize(b_red_mask, (target_width, target_height))
# 显示调整后的B组红色掩膜
cv2.imshow("B Red Mask", b_red_mask_resized)
cv2.waitKey(0) # 等待按键按下
cv2.destroyAllWindows() # 关闭所有窗口下面直接输出图片是户更方便
cv2.imwrite("a_color_mask.png", a_color_mask)
把a组与b组文件夹下的所有对应的图片进行计算，然后输出到一个txt文件
import cv2
import numpy as np
import os
def calculate_overlap_percentage(a_image_path, b_image_path):
# 读取A组和B组图像
a_image = cv2.imread(a_image_path)
b_image = cv2.imread(b_image_path)
# 将图像从BGR转为HSV色彩空间，便于颜色筛选
a_hsv = cv2.cvtColor(a_image, cv2.COLOR_BGR2HSV)
b_hsv = cv2.cvtColor(b_image, cv2.COLOR_BGR2HSV)
# 设置B组红色的HSV范围
lower_red = np.array([0, 100, 100])
upper_red = np.array([10, 255, 255])
# 提取B组红色区域的掩膜
b_red_mask = cv2.inRange(b_hsv, lower_red, upper_red)
# 设置A组颜色区域的HSV范围（棕色和绿色）
# 棕色的HSV范围
lower_brown = np.array([10, 30, 30])
upper_brown = np.array([25, 255, 200])
# 绿色的HSV范围
lower_green = np.array([35, 50, 50])
upper_green = np.array([85, 255, 255])
# 提取A组棕色和绿色区域的掩膜
a_brown_mask = cv2.inRange(a_hsv, lower_brown, upper_brown)
a_green_mask = cv2.inRange(a_hsv, lower_green, upper_green)
# 合并A组的棕色和绿色区域掩膜
a_color_mask = cv2.bitwise_or(a_brown_mask, a_green_mask)
# 计算B组红色区域的像素数量
b_red_area = cv2.countNonZero(b_red_mask)
# 在A组中找到与B组红色区域重合的区域
a_overlap = cv2.bitwise_and(a_color_mask, b_red_mask)
# 计算A组与B组红色区域重合的像素数量
a_overlap_area = cv2.countNonZero(a_overlap)
# 计算A组有颜色区域占B组红色区域的百分比
overlap_percentage = (a_overlap_area / b_red_area) * 100 if b_red_area != 0 else 0
return overlap_percentage
def process_images(a_folder, b_folder, output_file):
# 获取文件夹中的所有文件
a_files = sorted(os.listdir(a_folder))
b_files = sorted(os.listdir(b_folder))
# 打开输出文件
with open(output_file, 'w') as f:
# 遍历文件夹中的每一对图片
for a_file, b_file in zip(a_files, b_files):
# 拼接完整的文件路径
a_image_path = os.path.join(a_folder, a_file)
b_image_path = os.path.join(b_folder, b_file)
# 确保文件是图片文件
if not a_file.endswith(('.jpg', '.png', '.jpeg')) or not b_file.endswith(('.jpg', '.png', '.jpeg')):
continue
# 计算重叠百分比
percentage = calculate_overlap_percentage(a_image_path, b_image_path)
# 写入结果到txt文件
f.write(f'{a_file} vs {b_file}: {percentage:.2f}%\n')
print(f'{a_file} vs {b_file}: {percentage:.2f}%')
# 文件夹路径
a_folder = 'path_to_a_folder' # A组文件夹路径
b_folder = 'path_to_b_folder' # B组文件夹路径
output_file = 'overlap_results.txt' # 输出的txt文件路径
# 处理文件夹中的所有图片并输出到txt文件
process_images(a_folder, b_folder, output_file)
