SimpleBlobDetector-隔离与周围环境不同的Blob

时间:2019-05-09 19:21:01

标签: python opencv

我正在使用SimpleBlobDetector来定位小数点和其他类型的标点符号,如下图所示,有时检测器会从文本的实心区域中拾取斑点(底部中9),我正在寻找一种通过SimpleBlobDetector或后期处理来过滤掉这些检测结果的方法。

是否有一种方法可以指定斑点必须与背景色分开?也许是边缘检测方法?

感谢您的帮助。

检测器代码为:

    params = cv2.SimpleBlobDetector_Params()
    params.filterByArea = True
    params.minArea = 30
    params.minThreshold = 50
    params.maxThreshold = 200
    params.filterByConvexity = True
    params.minConvexity = 0.87
    params.filterByColor = True
    detector = cv2.SimpleBlobDetector_create(params)
    detections = detector.detect(img)

输出带有检测的图像 Output image with detections 原版的: Original

1 个答案:

答案 0 :(得分:3)

这不是使用SimpleBlobDetector,而是一种利用边缘/轮廓检测的解决方案,该方法可以进行更多的过滤控制。主要思想是

  • 将图像转换为灰度
  • 高斯模糊
  • 阈值图像将主要特征与背景分开
  • 执行Canny边缘检测
  • 扩大Canny图像以增强和封闭轮廓
  • 在图像中查找轮廓并使用最小/最大阈值区域进行过滤

阈值图像

enter image description here

Canny边缘检测

enter image description here

扩张以增强轮廓

enter image description here

根据面积检测和过滤轮廓

enter image description here

输出结果

  

检测到的轮廓:1

import numpy as np
import cv2

original_image = cv2.imread("1.jpg")
image = original_image.copy()
gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
blurred = cv2.GaussianBlur(gray, (3, 3), 0)
thresh = cv2.threshold(blurred, 110, 255,cv2.THRESH_BINARY)[1]
canny = cv2.Canny(thresh, 150, 255, 1)
kernel = cv2.getStructuringElement(cv2.MORPH_RECT, (5,5))
dilate = cv2.dilate(canny, kernel, iterations=1)

cv2.imshow("dilate", dilate)
cv2.imshow("thresh", thresh)
cv2.imshow("canny", canny)

# Find contours in the image
cnts = cv2.findContours(dilate.copy(), cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE)
cnts = cnts[0] if len(cnts) == 2 else cnts[1]

contours = []

threshold_min_area = 1100
threshold_max_area = 1200

for c in cnts:
    area = cv2.contourArea(c)
    if area > threshold_min_area and area < threshold_max_area:
        cv2.drawContours(original_image,[c], 0, (0,255,0), 3)
        contours.append(c)

cv2.imshow("detected", original_image) 
print('contours detected: {}'.format(len(contours)))
cv2.waitKey(0)