提高opencv找圆的准确率

问题描述

所以我有一些图像需要从中提取白色圆圈的数量（准确度为 +90%），到目前为止我已经想出了这个

original image(has 183 circles)

and here's the result I get with my code

如你所见，它可以检测到大约一半，但它不是那么准确

目前的代码：

def check_for_cells(img):
    found = 0
    found_list = []

    img = cv2.imread(img,cv2.IMREAD_COLOR)
    width,height = img.shape[:2]
    img = cv2.resize(img,(int(width/2),int(height/3)))
    gray = cv2.cvtColor(img,cv2.COLOR_BGR2GRAY)


    blurred = cv2.GaussianBlur(gray,(15,15),0)
    histo_norm = cv2.equalizeHist(blurred) 
    _,th = cv2.threshold(blurred,121,255,cv2.THRESH_BINARY + cv2.THRESH_OTSU) 

    min_dist = 20
    param1 = 95
    param2 = 18
    minRadius = 3
    maxRadius = 14


    circles = cv2.HoughCircles(histo_norm,cv2.HOUGH_GRADIENT,1,min_dist,param1=param1,param2=param2,minRadius=minRadius,maxRadius=maxRadius)
    circles_th = cv2.HoughCircles(th,param1=79,param2=12,maxRadius=maxRadius)

    if circles is not None:
        circles = np.uint16(np.around(circles))
        for i in circles[0,:]:
            cv2.circle(img,(i[0],i[1]),i[2],(0,0),2)
            found_list.append([(i[0],i[2]])
            found += 1


    if circles_th is not None:
        circles_th = np.uint16(np.around(circles_th))
        for i in circles_th[0,:]:
            if [(i[0],i[2]] not in found_list:
                cv2.circle(img,2)
                found_list.append([(i[0],i[2]])
                found += 1


    font = cv2.FONT_HERShey_SIMPLEX
    cv2.putText(img,str(found),100),font,4,10)
    # print(found)

    cv2.namedWindow('img',cv2.WINDOW_norMAL)
    cv2.imshow('img',img)
    cv2.waitKey(0)
    cv2.destroyAllWindows()

对改进代码的任何帮助将不胜感激！

解决方法

暂无找到可以解决该程序问题的有效方法，小编努力寻找整理中！

如果你已经找到好的解决方法，欢迎将解决方案带上本链接一起发送给小编。

小编邮箱:dio#foxmail.com (将#修改为@）