How to detect text from image using opencv

Well, One way can be you build your own convolution neural network(CNN) that detect letters/alphabet from the images

You can generate your own database using these code, this code will generate the image of particular letter

You may need to modify the code , because it might not having folder at a given path and you may require to deal a bit with hard-code value(static)

import os
import matplotlib.font_manager
from PIL import Image, ImageDraw , ImageFont
    
def rotate_image(image_to_rotate, font_index, parent_folder):
    src_im = Image.open(image_to_rotate)
    angle = 0
    while(angle < 15):
        angle += 5
        size = 320, 320
        dst_im = Image.new("RGBA", (320,320), "black" )
        im = src_im.convert('RGBA')
        rot = im.rotate( angle, expand=1 ).resize(size)
        dst_im.paste( rot, (0, 0), rot )
        r_img_dir = "%s\%s%s%s%s" % ( parent_folder, str(fonts.index(font)).zfill(5), "_rotated_by_", str(angle), ".png")
        dst_im.save(r_img_dir)
        
chars = [] 
for i in range(65,91): # Fix me 
    chars.append(chr(i))

for k in range(10):
    chars.append(k)

for ch in chars: # Remove me [:1]
    char= str(ch) # Fix me use instance of to check if int or Str ..     
    os.makedirs(os.path.join('folder', char)) # Fix me 
    fonts = matplotlib.font_manager.findSystemFonts(fontpaths=None, fontext='ttf')
    p_folder =  "%s\%s" % ( 'folder', char )
    font_count =0 
    for font in fonts:  # Remove me  [:10]
        get_str_for_spt = font[0:17]
        font_name = font.replace(get_str_for_spt,"")
        font_count += 1
        if font_count > 10:
            break
        img = Image.new('RGB', (320, 320), color = (0,0,0))  # make it - 16x16 in place of 32x32
        d = ImageDraw.Draw(img)
        img_font = ImageFont.truetype(font, 300)  # make it 12 in place of 16
        d.text((0,0),char, font=img_font)
        img_save_dir = "%s\%s_%s%s" % ( p_folder, str(fonts.index(font)).zfill(5),font_name, ".png")
        img.save(img_save_dir)
        rotate_image(img_save_dir, str(fonts.index(font)).zfill(5), p_folder)

print("Executed")

You can also take reference from following links

https://www.geeksforgeeks.org/text-detection-and-extraction-using-opencv-and-ocr/

Detect text area in an image using python and opencv

Detect text region in image using Opencv

How to process and extract text from image