opencv画坐标系（便于与世界坐标转化）、检测色块（目标物体）

 
 

 
① 根据图像尺寸遍历出所有节点；
② 根据行列个数得出边界节点；
③ 画点、两点连线；
④ 将视野中心点特殊标记；
⑤ 实现当色块位于网格节点或者视野中心点时，显示坐标值在图像上；
⑥ 可调整网格间距，10、20或者40等可以被尺寸值整除的数

 
 

from collections import  deque  
import numpy as np   
import cv2  
import time  


class Painter(object):
    def __init__(self,img,start_point=(20,20)):
        self.horizons_center = (0,0)
        self.point_list = []
        self.line_list = []
        self.get_point_line(img,start_point)

    def add_text(self,img,text,point,size=0.5,color=(0,0,255)):
        cv2.putText(img,text,point,cv2.FONT_HERSHEY_COMPLEX,size,color,1)
        return img

    def draw_a_circle(self,img,point,radius=3,color=(255,0,0),thickness=2):
        img = self.draw_a_point(img,point,radius,color,thickness)
        return img

    def draw_a_point(self,img,point,radius=3,color=(255,0,0),thickness=-4):
        cv2.circle(img, point, radius, color, thickness)
        return img

    def draw_all_point(self,img):
        for point in self.point_list:
            if point == self.horizons_center:
                img = self.draw_a_point(img,self.horizons_center,6,(0,0,255))
                lable_point = (self.horizons_center[0]-5,self.horizons_center[1]-5)
                img = self.add_text(img,str(self.horizons_center),lable_point)
            else:
                img = self.draw_a_point(img,point)
        return img

    def draw_a_line(self,img,start_point,end_point,color=(0,255,0),thickness=1,lineType=4):
        cv2.line(img, start_point, end_point, color, thickness, lineType)
        return img

    def draw_all_line(self,img):
        for point in self.line_list:
            img = self.draw_a_line(img, point[0], point[1])
        return img

    def get_point_line(self,img,start_point):
        start_value = start_point[0]
        space = start_value
        img_x,img_y = img.shape[1],img.shape[0]
        self.horizons_center = (int(img_x/2),int(img_y/2))
        
        for x_point in range(start_value,img_x,space):
            for y_point in range(start_value,img_y,space):
                self.point_list.append((x_point,y_point))

        num_row = int((img_y-start_value)/space)
        num_column = int((img_x-start_value)/space)
        for j in range(1,num_column+1):
            self.line_list.append(((j*space,start_value),(j*space,(img_y-space))))
        for i in range(1,num_row+1):
            self.line_list.append(((start_value,i*space),((img_x-space),i*space)))


class FindTarget(object):
    def __init__(self):
        self.redLower = np.array([156, 43, 46]) #设定红色阈值，HSV空间 
        self.redUpper = np.array([180, 255, 255])
        self.target_position = (0,0)
        self.target_radius = 0
        cv2.namedWindow("camera",1)
        self.camera = cv2.VideoCapture(0)

    def get_one_flame(self):
        ret,img = self.camera.read()
        return ret,img

    def start_find(self,img):
        hsv = cv2.cvtColor(img, cv2.COLOR_BGR2HSV)   #转到HSV空间
        mask_red = cv2.inRange(hsv, self.redLower, self.redUpper)  #根据阈值构建掩膜
        mask_red = cv2.erode(mask_red, None, iterations=2)  #腐蚀操作
        mask_red = cv2.dilate(mask_red, None, iterations=2)  #膨胀操作，其实先腐蚀再膨胀的效果是开运算，去除噪点
        cnts_red = cv2.findContours(mask_red.copy(), cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE)[-2]  #轮廓检测
        if len(cnts_red) > 0:
            c = max(cnts_red, key=cv2.contourArea) #找到面积最大的轮廓
            (red_x,red_y), self.target_radius = cv2.minEnclosingCircle(c)
            self.target_position = (int(red_x),int(red_y))
            return True,img
        return False,img

    def is_horizons_center(self,center):
        x_low,x_high = center[0]-5,center[0]+5
        y_low,y_high = center[1]-5,center[1]+5
        x_target,y_target = self.target_position[0],self.target_position[1]
        if x_low<x_target<x_high and y_low<y_target<y_high:
            return True
        return False

    def is_target_on_node(self,point_list):
        x_target,y_target = self.target_position[0],self.target_position[1]
        for point in point_list:
            if (point[0]-3)<x_target<(point[0]+3) and (point[1]-3)<y_target<(point[1]+3):
                return True
        return False

    def show_result(self,img):
        cv2.imshow('Frame', img)

    def is_exit(self):
        k = cv2.waitKey(5)&0xFF 
        if k == 27:  
            return True
        return False

    def close(self):
        self.camera.release()  #摄像头释放 
        cv2.destroyAllWindows()#销毁所有窗口 

    
    
if __name__ == '__main__':

    toFind = FindTarget()
    ret,flame = toFind.get_one_flame()
    
    if ret:
        painter = Painter(flame,(40,40))
        while True:
            _,flame = toFind.get_one_flame()
            result,flame = toFind.start_find(flame)
            if result:
                flame = painter.draw_a_circle(flame,toFind.target_position,int(toFind.target_radius),(0,255,255))
                flame = painter.draw_a_point(flame,toFind.target_position,5,(0,255,0))
                flame = painter.add_text(flame,str(toFind.target_position),toFind.target_position)
                if toFind.is_horizons_center(painter.horizons_center):
                    flame = painter.add_text(flame,'arrive center point!',(30,30),1,(0,0,255))
                if toFind.is_target_on_node(painter.point_list):
                    flame = painter.add_text(flame,'YES!',(550,30),1,(0,0,255))                   
                print('目标物体坐标:  ',toFind.target_position)               
            flame = painter.draw_all_point(flame)
            flame = painter.draw_all_line(flame)

            toFind.show_result(flame)
            if toFind.is_exit():
                break
    else:
        print ('No Camera')
        
    toFind.close()