First you need to go to the official website to download the code library
Volcengine official code warehouse see volc-sdk-python, GitHub - volcengine/volc-sdk-python
Here are the libraries we need to call, which can be directly downloaded and placed locally
Download the volcengine file directly and put it locally
Then there is our code section
First, we need to convert the image we need for text recognition to base64 encoding
Then we need the AK and SK of our Volcano Engine account, which can be viewed on the Volcano Engine console. If you need help here, you can directly consult the customer service of Volcano Engine.
The last word is to extract the text
Not much to say, just look at the code
import base64
from volcengine.visual.VisualService import VisualService
img_path = 'D:\sdcard\lanya\ceshi.jpg' # 图片
with open(img_path, 'rb') as f:
image_data = f.read()
base64_data = base64.b64encode(image_data) # 图片转base64编码
visual_service = VisualService()
visual_service.set_ak('这里是你的AK')
visual_service.set_sk('这里是你的SK')
form = dict()
form["image_base64"] = base64_data
resp = visual_service.ocr_normal(form)
imageText = resp.get('data').get('line_texts') # 获取文字
print(imageText)
num_list_string = " ".join('%s' % id for id in imageText)
res = num_list_string.find("车辆") # 寻找图片里的车辆这两个字
print(res)
Below is the image I want to identify
The result of running the code is as follows:
The result is exactly what we want!
There will be a small pit here:
If there is an error when the script is running, the error is reported in Util.py, and you need to download crypto. If so, change the folder name to a capital C. After downloading, it should be crypto, and change it to Crypto.
Continue to run after modification
At this time we need to download pycryptodome
Download command pip install pycryptodome
Then run it again and it will be OK without error.