09-09 03:40 阅读 134

OpenCV简单标准数字识别的完整实例

这篇文章主要给大家介绍了关于OpenCV简单标准数字识别的相关资料,需要的朋友可以参考下

在学习openCV时，看到一个问答做数字识别，里面配有代码，应用到了openCV里面的ml包，很有学习价值。

https://stackoverflow.com/questions/9413216/simple-digit-recognition-ocr-in-opencv-python#

import sys

import numpy as np

import cv2

im = cv2.imread('t.png')

im3 = im.copy()

gray = cv2.cvtColor(im,cv2.COLOR_BGR2GRAY)   #先转换为灰度图才能够使用图像阈值化

thresh = cv2.adaptiveThreshold(gray,255,cv2.ADAPTIVE_THRESH_GAUSSIAN_C,cv2.THRESH_BINARY,11,2)  #自适应阈值化

##################      Now finding Contours         ###################

#

image,contours,hierarchy = cv2.findContours(thresh,cv2.RETR_LIST,cv2.CHAIN_APPROX_SIMPLE)

#边缘查找，找到数字框，但存在误判

samples =  np.empty((0,900))    #将每一个识别到的数字所有像素点作为特征，储存到一个30*30的矩阵内

responses = []                  #label

keys = [i for i in range(48,58)]    #48-58为ASCII码

count =0

for cnt in contours:

    if cv2.contourArea(cnt)>80:     #使用边缘面积过滤较小边缘框

        [x,y,w,h] = cv2.boundingRect(cnt)   

        if  h>25 and h < 30:        #使用高过滤小框和大框

            count+=1

            cv2.rectangle(im,(x,y),(x+w,y+h),(0,0,255),2)

            roi = thresh[y:y+h,x:x+w]

            roismall = cv2.resize(roi,(30,30))

            cv2.imshow('norm',im)

            key = cv2.waitKey(0)

            if key == 27:  # (escape to quit)

                sys.exit()

            elif key in keys:

                responses.append(int(chr(key)))

                sample = roismall.reshape((1,900))

                samples = np.append(samples,sample,0)

            if count == 100:        #过滤一下过多边缘框，后期可能会尝试极大抑制

                break

responses = np.array(responses,np.float32)

responses = responses.reshape((responses.size,1))

print ("training complete")

np.savetxt('generalsamples.data',samples)

np.savetxt('generalresponses.data',responses)

#

cv2.waitKey()

cv2.destroyAllWindows()

训练数据为：

测试数据为：

使用openCV自带的ML包，KNearest算法

import sys

import cv2

import numpy as np

 #######   training part    ############### 

samples = np.loadtxt('generalsamples.data',np.float32)

responses = np.loadtxt('generalresponses.data',np.float32)

responses = responses.reshape((responses.size,1))

model = cv2.ml.KNearest_create()

model.train(samples,cv2.ml.ROW_SAMPLE,responses)

def getNum(path):

    im = cv2.imread(path)

    out = np.zeros(im.shape,np.uint8)

    gray = cv2.cvtColor(im,cv2.COLOR_BGR2GRAY)

    #预处理一下

    for i in range(gray.__len__()):

        for j in range(gray[0].__len__()):

            if gray[i][j] == 0:

                gray[i][j] == 255

            else:

                gray[i][j] == 0

    thresh = cv2.adaptiveThreshold(gray,255,1,1,11,2)

    image,contours,hierarchy = cv2.findContours(thresh,cv2.RETR_LIST,cv2.CHAIN_APPROX_SIMPLE)

    count = 0

    numbers = []

    for cnt in contours:

        if cv2.contourArea(cnt)>80:

            [x,y,w,h] = cv2.boundingRect(cnt)

            if  h>25:

                cv2.rectangle(im,(x,y),(x+w,y+h),(0,255,0),2)

                roi = thresh[y:y+h,x:x+w]

                roismall = cv2.resize(roi,(30,30))

                roismall = roismall.reshape((1,900))

                roismall = np.float32(roismall)

                retval, results, neigh_resp, dists = model.findNearest(roismall, k = 1)

                string = str(int((results[0][0])))

                numbers.append(int((results[0][0])))

                cv2.putText(out,string,(x,y+h),0,1,(0,255,0))

                count += 1

        if count == 10:

            break

    return numbers

numbers = getNum('1.png')

总结

到此这篇关于OpenCV简单标准数字识别的文章就介绍到这了

原文链接：https://blog.csdn.net/huang_nansen/article/details/83241143