基于Hough Transform的虹膜定位代码运行报错求助

问题背景 #

此前可正常实现虹膜定位与裁剪的代码，再次运行时触发OpenCV参数错误，代码及报错信息如下：

原代码 #

# circles = cv2.HoughCircles(canny,cv2.HOUGH_GRADIENT,1,10000,param1=50,param2=30,minRadius=50,maxRadius=100)
circles = cv2.HoughCircles(canny,cv2.HOUGH_GRADIENT,1,10000,param1=50,param2=30,minRadius=0,maxRadius=1000)

height,width = gray.shape
r = 0
mask = np.zeros((height,width), np.uint8)
for i in circles[0,:]:
    cv2.circle(bgr,(i[0],i[1]),i[2],(0,255,0),3)
    cv2.circle(mask,(i[0],i[1]),i[2],(255,255,255),thickness=-1)
    blank_image = bgr[:int(i[1]),:int(i[1])]

    masked_data = cv2.bitwise_and(gray, gray, mask=mask)
    _,thresh = cv2.threshold(mask,1,255,cv2.THRESH_BINARY)
    contours = cv2.findContours(thresh,cv2.RETR_EXTERNAL,cv2.CHAIN_APPROX_SIMPLE)
    x,y,w,h = cv2.boundingRect(contours[0][0])
    crop = masked_data[y:y+h,x:x+w]
    r = i[2]
    crop_dim =cv2.cvtColor(crop, cv2.COLOR_RGB2BGR)

plt.imshow(bgr)
plt.show()
plt.imshow(crop_dim)
plt.savefig('Hough Transform')
plt.show()

报错信息 #

error                                     Traceback (most recent call last)
<ipython-input-10-74f81f85c0bc> in <module>()
      6 mask = np.zeros((height,width), np.uint8)
      7 for i in circles[0,:]:
----> 8     cv2.circle(bgr,(i[0],i[1]),i[2],(0,255,0),3)
      9     cv2.circle(mask,(i[0],i[1]),i[2],(255,255,255),thickness=-1)
     10     blank_image = bgr[:int(i[1]),:int(i[1])]

error: OpenCV(4.6.0) :-1: error: (-5:Bad argument) in function 'circle'
> Overload resolution failed:
>  - Can't parse 'center'. Sequence item with index 0 has a wrong type
>  - Can't parse 'center'. Sequence item with index 0 has a wrong type

错误原因 #

1. 参数类型不匹配：cv2.HoughCircles返回的圆坐标（圆心x、y）和半径是浮点数类型，但cv2.circle要求圆心坐标必须是整数类型的元组，直接传入浮点数会触发类型错误。
2. 未处理空检测结果：如果当前输入图像未检测到任何圆，circles会返回None，此时直接访问circles[0,:]会导致额外错误，这也是代码不稳定的潜在原因。

解决方法 #

步骤1：添加空值判断 #

先确认circles是否有效，避免空指针异常。

步骤2：转换参数类型 #

将检测到的圆心坐标和半径强制转换为整数类型，符合cv2.circle的参数要求。

修改后的代码 #

# circles = cv2.HoughCircles(canny,cv2.HOUGH_GRADIENT,1,10000,param1=50,param2=30,minRadius=50,maxRadius=100)
circles = cv2.HoughCircles(canny,cv2.HOUGH_GRADIENT,1,10000,param1=50,param2=30,minRadius=0,maxRadius=1000)

height,width = gray.shape
r = 0
mask = np.zeros((height,width), np.uint8)

# 先判断是否检测到圆
if circles is not None:
    # 将浮点数转换为整数
    circles = np.uint16(np.around(circles))
    for i in circles[0,:]:
        # 圆心坐标和半径都用整数
        cv2.circle(bgr,(i[0],i[1]),i[2],(0,255,0),3)
        cv2.circle(mask,(i[0],i[1]),i[2],(255,255,255),thickness=-1)
        blank_image = bgr[:i[1],:i[1]]

        masked_data = cv2.bitwise_and(gray, gray, mask=mask)
        _,thresh = cv2.threshold(mask,1,255,cv2.THRESH_BINARY)
        contours = cv2.findContours(thresh,cv2.RETR_EXTERNAL,cv2.CHAIN_APPROX_SIMPLE)
        x,y,w,h = cv2.boundingRect(contours[0][0])
        crop = masked_data[y:y+h,x:x+w]
        r = i[2]
        # 修正灰度图转BGR的逻辑
        if len(crop.shape) == 2:
            crop_dim = cv2.cvtColor(crop, cv2.COLOR_GRAY2BGR)
        else:
            crop_dim = cv2.cvtColor(crop, cv2.COLOR_RGB2BGR)

    plt.imshow(bgr)
    plt.show()
    plt.imshow(crop_dim)
    plt.savefig('Hough Transform')
    plt.show()
else:
    print("未检测到虹膜圆，请检查输入图像或HoughCircles参数")

额外说明 #

• 原代码中crop_dim = cv2.cvtColor(crop, cv2.COLOR_RGB2BGR)存在隐患：crop是灰度图（单通道），直接用COLOR_RGB2BGR会报错，修正为判断通道数后选择对应转换规则。
• 如果仍无法检测到圆，需调整cv2.HoughCircles的参数（如param1、param2、半径范围），适配当前输入图像的特征。

内容的提问来源于stack exchange，提问作者koboikampus