补充：将OCR识别的结果与数据集正确结果合在一起进行比对

链接

[link]补充内容为利用文字定位模块与文字识别模块的接口批量处理图片生成OCR结果（加上标记结果进行比对）。https://blog.****.net/weixin_42343812/article/details/85100044

生成后的结果如下所示。

代码片

下面展示一些 内联代码片。

// A code block
var foo = 'bar';

s = ''
j = 2000
k = 3
if os.path.exists('./img10000_ocr/'+'pict_0.xlsx'):
    os.remove('./img10000_ocr/'+'pict_0.xlsx')
book = xlsxwriter.Workbook('./img10000_ocr/pict_0.xlsx')
sheet = book.add_worksheet('demo')
i = 1
flag_flag=0
file_json = open('/home/xudong/ocr/dataset.txt', 'r')
line=file_json.readline()
json_engchi=[]
#tuple存储所有的读取后的标记数据
tuplel=[] 
#读取所有的json文件——标记数据
while(line!=''):
    eng_j=line.split(',')[1]
    chi_j = line.split(',')[2]
    tuplel=[]
    tuplel.append(eng_j)
    tuplel.append(chi_j)
    json_engchi.append(tuplel)
    line = file_json.readline()


for i in range(1,len(b)):
    shenqing = b_img_name[i]

    ocr = b[i].split('，')[0]
    chi = b[i].split('，')[1]
    eng = b[i].split('，')[2]
    if (i != 0):
        if i % 1000 == 0:
            if flag_flag==0:
                flag_flag=1
            else:
                #book.close()
                if os.path.exists('./img10000_ocr/'+'pict_'+str(i/1000)+'.xlsx'):
                    os.remove('./img10000_ocr/'+'pict_'+str(i/1000)+'.xlsx')
                book = xlsxwriter.Workbook('./img10000_ocr/pict_'+str(int(i/1000))+'.xlsx')
                sheet = book.add_worksheet('demo')
                k=3
        sheet.insert_image('A' + str(k), '/home/xudong/ocr/img10000/' + shenqing + '.jpg')
        sheet.insert_image('B' + str(k), '/home/xudong/ocr/img10000/' + shenqing + '_ctpn.jpg')
        sheet.write(k-1 , 2, ocr)
        sheet.write(k-1 , 4, chi)
        sheet.write(k-1 , 7, eng)
        sheet.write(k-1, 8, json_engchi[int(shenqing)][0])
        sheet.write(k-1, 5, json_engchi[int(shenqing)][1])

        i += 1
        print(i)
        k += 1
print(i)

book.close()