Python+OpenCV实现图片及视频中选定区域颜色识别

近期，需要实现检测摄像头中指定坐标区域内的主体颜色，通过查阅大量相关的内容，最终实现代码及效果如下，具体的实现步骤在代码中都详细注释，代码还可以进一步优化，但提升有限。

主要实现过程：按不同颜色的取值范围，对图像进行循环遍历，转换为灰度图，将本次遍历的颜色像素转换为白色，对白色部分进行膨胀处理，使其更加连续，计算白色部分外轮廓包围的面积累加求和，比较每种颜色围起来面积，保存最大值及其颜色，所有颜色遍历完后，返回最大值对应的颜色，显示在图像上

如果有类似的颜色识别的任务，可参考以下代码修改后实现具体需求

colorList.py

import numpy as np
import collections

# 将rgb图像转换为hsv图像后，确定不同颜色的取值范围
def getColorList():
   dict = collections.defaultdict(list)

# black
   lower_black = np.array([0, 0, 0])
   upper_black = np.array([180, 255, 46])
   color_list_black = []
   color_list_black.append(lower_black)
   color_list_black.append(upper_black)
   dict['black'] = color_list_black

# gray
   lower_gray = np.array([0, 0, 46])
   upper_gray = np.array([180, 43, 220])
   color_list_gray= []
   color_list_gray.append(lower_gray)
   color_list_gray.append(upper_gray)
   dict['gray'] = color_list_gray

# white
   lower_white = np.array([0, 0, 221])
   upper_white = np.array([180, 30, 255])
   color_list_white = []
   color_list_white.append(lower_white)
   color_list_white.append(upper_white)
   dict['white'] = color_list_white

# red
   lower_red = np.array([156, 43, 46])
   upper_red = np.array([180, 255, 255])
   color_list_red = []
   color_list_red.append(lower_red)
   color_list_red.append(upper_red)
   dict['red'] = color_list_red

# red2
   lower_red = np.array([0, 43, 46])
   upper_red = np.array([10, 255, 255])
   color_list_red2 = []
   color_list_red2.append(lower_red)
   color_list_red2.append(upper_red)
   dict['red2'] = color_list_red2

# orange
   lower_orange = np.array([11, 43, 46])
   upper_orange = np.array([25, 255, 255])
   color_list_orange = []
   color_list_orange.append(lower_orange)
   color_list_orange.append(upper_orange)
   dict['orange'] = color_list_orange

# yellow
   lower_yellow = np.array([26, 43, 46])
   upper_yellow = np.array([34, 255, 255])
   color_list_yellow = []
   color_list_yellow.append(lower_yellow)
   color_list_yellow.append(upper_yellow)
   dict['yellow'] = color_list_yellow

# green
   lower_green = np.array([35, 43, 46])
   upper_green = np.array([77, 255, 255])
   color_list_green = []
   color_list_green.append(lower_green)
   color_list_green.append(upper_green)
   dict['green'] = color_list_green

# cyan
   lower_cyan = np.array([78, 43, 46])
   upper_cyan = np.array([99, 255, 255])
   color_list_cyan = []
   color_list_cyan.append(lower_cyan)
   color_list_cyan.append(upper_cyan)
   dict['cyan'] = color_list_cyan

# blue
   lower_blue = np.array([100, 43, 46])
   upper_blue = np.array([124, 255, 255])
   color_list_blue = []
   color_list_blue.append(lower_blue)
   color_list_blue.append(upper_blue)
   dict['blue'] = color_list_blue

# purple
   lower_purple = np.array([125, 43, 46])
   upper_purple = np.array([155, 255, 255])
   color_list_purple = []
   color_list_purple.append(lower_purple)
   color_list_purple.append(upper_purple)
   dict['purple'] = color_list_purple

return dict

if __name__ == '__main__':
   color_dict = getColorList()
   print(color_dict)

num = len(color_dict)
   print('num=', num)

for d in color_dict:
       print('key=', d)
       print('value=', color_dict[d][1])

image_color_realize.py

import cv2
import colorList

# 实现对图片中目标区域颜色的识别
def get_color(frame):
   print('go in get_color')
   hsv = cv2.cvtColor(frame, cv2.COLOR_BGR2HSV)
   maxsum = 0
   color = None
   color_dict = colorList.getColorList()

# count = 0

for d in color_dict:
       mask = cv2.inRange(hsv, color_dict[d][0], color_dict[d][1])  # 在后两个参数范围内的值变成255
       binary = cv2.threshold(mask, 127, 255, cv2.THRESH_BINARY)[1]  # 在灰度图片中，像素值大于127的都变成255，[1]表示调用图像，也就是该函数第二个返回值

# cv2.imshow("0",binary)
       # cv2.waitKey(0)
       # count+=1

binary = cv2.dilate(binary, None, iterations=2)  # 使用默认内核进行膨胀操作，操作两次，使缝隙变小，图像更连续
       cnts = cv2.findContours(binary, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE)[-2]  # 获取该函数倒数第二个返回值轮廓
       sum = 0
       for c in cnts:
           sum += cv2.contourArea(c)  # 获取该颜色所有轮廓围成的面积的和
       # print("％s  , ％d" ％(d, sum ))
       if sum > maxsum:
           maxsum = sum
           color = d
           if color == 'red2':
               color = 'red'
           elif color == 'orange':
               color = 'yellow'
           elif color == 'purple' or color == 'blue' or color == 'cyan' or color == 'white' or color == 'green':
               color = 'normal'
   return color

if __name__ == '__main__':
   filename = "C:/Users/admin/Desktop/water_samples/live01.jpg"
   frame = cv2.imread(filename)
   # frame = frame[180:280, 180:380]  # [y:y+h, x:x+w] 注意x,y顺序
   color = get_color(frame)

# 绘制文本
   cv2.putText(img=frame,text=color,org=(20,50),fontFace=cv2.FONT_HERSHEY_SIMPLEX,
               fontScale=1.0,color=(0,255,0),thickness=2)

# cv2.namedWindow('frame',cv2.WINDOW_NORMAL)  # 设置显示窗口可调节
   cv2.imshow('frame',frame)
   cv2.waitKey(0)

video_color_realize.py

import cv2
import xf_color

# 对视频或摄像头获取的影像目标区域颜色进行识别

cap = cv2.VideoCapture("C:/Users/admin/Desktop/water_samples/01.mp4")
# cap = cv2.VideoCapture(0)
cap.set(cv2.CAP_PROP_FRAME_WIDTH, 1100)  # 这里窗口大小调节只对摄像头有效
cap.set(cv2.CAP_PROP_FRAME_HEIGHT, 750)

while cap.isOpened():
   ret, frame0 = cap.read()
   # 对图像帧进行翻转（因为opencv图像和我们正常是反着的） 视频是正常的，摄像头是反转的
   # frame0 = cv2.flip(src=frame0, flipCode=2)

# frame = frame[180:280, 180:380]  # [y:y+h, x:x+w]
   # frame = frame0[200:400, 100:300]  # 设置检测颜色的区域，四个顶点坐标
   frame = frame0

# frame=cv2.resize(src=frame,dsize=(750,600))
   hsv_frame = cv2.cvtColor(src=frame, code=cv2.COLOR_BGR2HSV)
   # 获取读取的帧的高宽
   height, width, channel = frame.shape
   color = xf_color.get_color(hsv_frame)
   # 绘制文本
   cv2.putText(img=frame0, text=color, org=(20, 50), fontFace=cv2.FONT_HERSHEY_SIMPLEX,
               fontScale=1.0, color=(0, 255, 0), thickness=2)
   cv2.imshow('frame', frame0)
   key = cv2.waitKey(1)
   if key == 27:
       break

cap.release()
cv2.destroyAllWindows()

if __name__ == '__main__':
   print('Pycharm')

效果如下：

示例图片1

示例图片2

示例图片3

来源：https://www.cnblogs.com/wydilearn/p/16505887.html