Batch Image Translation Demo Using Youdao OCR API with Python

The author describes a personal need to translate cosmetic product labels for a girlfriend and decides to develop a batch image translation demo using Youdao's AI‑powered OCR translation service.

Effect Demonstration

Several screenshots show successful translation results for different product images, including mixed Korean/English text.

Preparing API Call – Generating App ID and Secret

Instructions explain how to create an instance and application on the Youdao platform to obtain the required appKey and appSecret .

API Interface Introduction

The OCR translation endpoint is https://openapi.youdao.com/ocrtransapi , accessed via POST with form data and JSON response.

API Call Parameters

Field Name

Type

Meaning

Required

Notes

type

text

File upload type

True

Set to 1 for Base64

from

text

Source language

True

Can be "auto"

to

text

Target language

True

Can be "auto"

appKey

text

Application ID

True

Find in application management

salt

text

UUID

True

e.g., 1995882C5064805BC30A39829B779D7B

sign

text

Signature

True

MD5(appKey+q+salt+appSecret)

ext

text

Audio format for result

False

mp3

q

text

Image to recognize

True

Base64 of image when type=1

docType

text

Response type

False

json

render

text

Return rendered image

False

0 (no) or 1 (yes)

nullIsError

text

Return error if no text detected

False

"false" or "true"

Signature generation steps: concatenate appKey , q , salt , and appSecret in that order, then compute the MD5 hash to obtain sign .

Development Process

1. API Interface Overview

The core of the project is calling the Youdao OCR translation API.

2. Detailed Development

The demo consists of three Python files: maindow.py (Tkinter UI), transclass.py (image handling), and pictranslate.py (API wrapper).

UI Code (maindow.py)

<code>root=tk.Tk()
root.title("netease youdao translation test")
frm = tk.Frame(root)
frm.grid(padx='50', pady='50')
btn_get_file = tk.Button(frm, text='选择待翻译图片', command=get_files)
btn_get_file.grid(row=0, column=0, ipadx='3', ipady='3', padx='10', pady='20')
text1 = tk.Text(frm, width='40', height='10')
text1.grid(row=0, column=1)
btn_get_result_path=tk.Button(frm,text='选择翻译结果路径',command=set_result_path)
btn_get_result_path.grid(row=1,column=0)
text2=tk.Text(frm,width='40', height='2')
text2.grid(row=1,column=1)
btn_sure=tk.Button(frm,text="翻译",command=translate_files)
btn_sure.grid(row=2,column=1)
root.mainloop()</code>

File selection, result path selection, and translation trigger are implemented with Tkinter dialogs.

<code>def get_files():
    files = filedialog.askopenfilenames(filetypes=[('text files', '.jpg')])
    translate.file_paths = files
    if files:
        for file in files:
            text1.insert(tk.END, file + '\n')
            text1.update()
    else:
        print('你没有选择任何文件')
</code>

<code>def set_result_path():
    result_path = filedialog.askdirectory()
    translate.result_root_path = result_path
    text2.insert(tk.END, result_path)
</code>

<code>def translate_files():
    if translate.file_paths:
        translate.translate_files()
        tk.messagebox.showinfo("提示", "搞定")
    else:
        tk.messagebox.showinfo("提示", "无文件")
</code>

Batch Image Processing (transclass.py)

<code>class Translate():
    def __init__(self, name, file_paths, result_root_path, trans_type):
        self.name = name
        self.file_paths = file_paths
        self.result_root_path = result_root_path
        self.trans_type = trans_type
    def translate_files(self):
        for file_path in self.file_paths:
            file_name = os.path.basename(file_path)
            print('===========' + file_path + '===========' )
            trans_reult = self.translate_use_netease(file_path)
            open(self.result_root_path + '/result_' + file_name.split('.')[0] + '.txt','w').write(trans_reult)
    def translate_use_netease(self, file_content):
        result = connect(file_content)
        return result
</code>

Calling Youdao API (pictranslate.py)

<code>def connect(file_content, fromLan, toLan):
    f = open(file_content, 'rb')
    q = base64.b64encode(f.read()).decode('utf-8')
    f.close()
    data = {}
    data['from'] = 'auto'
    data['to'] = 'auto'
    data['type'] = '1'
    data['q'] = q
    salt = str(uuid.uuid1())
    signStr = APP_KEY + q + salt + APP_SECRET
    sign = encrypt(signStr)
    data['appKey'] = APP_KEY
    data['salt'] = salt
    data['sign'] = sign
    response = do_request(data)
    result = json.loads(str(response.content, encoding="utf-8"))
    translateResults = result['resRegions']
    pictransresult = ""
    for i in translateResults:
        pictransresult = pictransresult + i['tranContent'] + "\n"
    return pictransresult
</code>

Result Summary

The JSON response contains fields such as orientation , lanFrom , lanTo , and resRegions with detailed translation content, bounding boxes, and layout information.

Field

Description

orientation

Image orientation

lanFrom

Detected source language

textAngle

Image tilt angle

errorCode

Error code

lanTo

Target language

resRegions

Translated content per region

The author concludes that leveraging an open AI platform makes image recognition and natural language processing straightforward, allowing more time for personal enjoyment.

Project repository: https://github.com/LemonQH/BatchPicTranslate