RuntimeError: Failed to init API, possibly an invalid tessdata path: C:\Users\hp\Anaconda3\/tessdata/

user_13554219 0
Viewed 2620

I am using windows 10 and tesserocr version is 2.4. Want to detect text from an image and then the language of that text.

While running this piece of code:

from tesserocr import PyTessBaseAPI
import argparse
parser = argparse.ArgumentParser("Enter Image Path")
parser.add_argument('-i' , '--image' , help="Image Path")
args, unknown = parser.parse_known_args()
images = ['b.jpg']
images[0] = args.image
count = 0
count2 = 0
count3 = 0
with PyTessBaseAPI() as api:
    for img in images:
        api.Init(lang = 'eng')
        api.SetImageFile(img)
        # print api.AllWordConfidences()
        arr = list(api.AllWordConfidences())
        sumarr = sum(arr) / float(len(arr))

Getting below error :

<ipython-input-3-152e90ab44c3> in <module>
     13 count2 = 0
     14 count3 = 0
---> 15 with PyTessBaseAPI() as api:
     16    for img in images:
     17        api.Init(lang = 'eng')

tesserocr.pyx in tesserocr._tesserocr.PyTessBaseAPI.__cinit__()

tesserocr.pyx in tesserocr._tesserocr.PyTessBaseAPI._init_api()

RuntimeError: Failed to init API, possibly an invalid tessdata path: C:\Users\hp\Anaconda3\/tessdata/

Anyone aware of this issue ?

tesseract language-detection ocr python
3 Answers

Add an environment variable on your system to point to the tessdata directory

enter image description here

You can also pass it in to PyTessBaseAPI as per the docs:

Args:
    path (str): The name of the parent directory of tessdata. Must end in /.

Should be:

 with PyTessBaseAPI(path='./') as api:

this is if you have the trainedata in the same folder as the script, otherwise substitute './' with your path

user_45126460

For others who still got this issue if you are using another language other than English, you might need to pass lang parameter to tessaocr like this

with PyTessBaseAPI(path='./',lang='tur') as api:
        api.SetImage(imageI)
        ocr = api.GetUTF8Text()
        confidence = api.AllWordConfidences()
        print(ocr,confidence)
        return ocr
user_64822630

For Windows when we are installing tesserocr library, by default it is going in the below path: default path-C:\Users\sagar\Anaconda_New\envs\Image_to_OCR\share/tessdata*

On executing our code we are getting this error : RuntimeError: Failed to init API, possibly an invalid tessdata path: C:\Users\sagar\Anaconda_New\envs\Image_to_OCR/tessdata/

Solution to this problem is cut the tessdata folder from the default path mentioned above and paste it in this path: C:\Users\sagar\Anaconda_New\envs\Image_to_OCR.

If you follow this steps your problem will be solved.

user_116734330
Related