CRNN_Chinese_Characters_Rec/lib/dataset/image_process_test.py

# from __future__ import print_function, absolute_import
import torch.utils.data as data
import os
import numpy as np
import cv2,math


def t1():
    img = cv2.imread("C:\\Users\\Administrator\\Desktop\\OCR_pytorch\\CRNN_Chinese_Characters_Rec\\images\\Snipaste_2023-06-12_16-25-38.jpg")
    img = cv2.imread("C:\\Users\\Administrator\\Desktop\\OCR_pytorch\\Attention_ocr.pytorch\\test_img\\20436312_1683447152.jpg")
    # img = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
    # img = cv2.cvtColor(img, cv2.COLOR_GRAY2BGR)
    print('img',img)
    cv2.imshow('1',img)
    cv2.waitKey(2000)
    # 数据增强
    # img = img_process(img)

    img_h, img_w = img.shape[:2]

    inp_h = 32
    inp_w = 960

    # img = cv2.resize(img, (0,0), fx=self.inp_w / img_w, fy=self.inp_h / img_h, interpolation=cv2.INTER_CUBIC)
    # img = np.reshape(img, (self.inp_h, self.inp_w, 1))

    # 图片真实长宽比
    ratio = img_w / float(img_h)

    # 按比例缩放
    if math.ceil(inp_h * ratio) > inp_w:
        # 如大于默认宽度，则宽度为imgW
        resized_w = inp_w
    else:
        # 如小于默认宽度则以图片真实宽为准
        resized_w = int(math.ceil(inp_h * ratio))
    # 缩放
    img = cv2.resize(img, (resized_w, inp_h))
    print('img2',img)
    cv2.imshow('2',img)
    cv2.waitKey(3000)

    img = img.astype(np.float32)
    # 标准化
    # img = (img/255. - 0.5) / 0.5
    img = (img/255. - 0.588) / 0.193

    cv2.imshow('33', img)
    cv2.waitKey(3000)

    img = img.transpose([2, 0, 1])
    print('img3',img)
    # 对宽度不足的位置，补0
    padding_im = np.full((3, inp_h, inp_w),255, dtype=np.float32)
    padding_im[:, :, 0:resized_w] = img
    print('img4',padding_im)
    cv2.imshow('3',padding_im.transpose((1,2,0)))
    cv2.waitKey(3000)

def t2():
    new = []
    with open("C:\\Users\\Administrator\\Desktop\\OCR_pytorch\\CRNN_Chinese_Characters_Rec\\lib\\dataset\\txt\\train.txt", "r", encoding='utf-8') as f:
        for n in f.readlines():
            new.append(n)
        f.close()
    with open("C:\\Users\\Administrator\\Desktop\\OCR_pytorch\\CRNN_Chinese_Characters_Rec\\lib\\dataset\\txt\\train2.txt", "w", encoding='utf-8') as f:
        for n in new[:30000]:
            f.write(n)
            # f.write("\n")
        f.close()


if __name__ == '__main__':
    t1()
    # t2()
    pass