Smart Librarian by Python Optical Character Recognition

Smart Librarian.pdf

2.69MB

 

from PIL import Image
import http.client, urllib.request, urllib.parse, urllib.error, base64, json
import cv2
import numpy as np
import winsound
def order_points(pts):

    rect = np.zeros((4, 2), dtype = "float32")
    s = pts.sum(axis = 1)
    rect[0] = pts[np.argmin(s)]
    rect[2] = pts[np.argmax(s)]

    diff = np.diff(pts, axis = 1)
    rect[1] = pts[np.argmin(diff)]
    rect[3] = pts[np.argmax(diff)]

    return rect

def print_text(json_data):
    result = json.loads(json_data)
    for l in result['regions']:
        for w in l['lines']:
            line = []
            for r in w['words']:
                line.append(r['text'])
            global t1
            t1=(''.join(line))

                
            
    return

def ocr_project_oxford(headers, params, data):
    conn = http.client.HTTPSConnection('westcentralus.api.cognitive.microsoft.com')
    conn.request("POST", "/vision/v1.0/ocr?%s" % params, data, headers)
    response = conn.getresponse()
    data = response.read().decode()
    #print (data)
    print_text(data)
    
    conn.close()
    return


def auto_scan_image_via_webcam():


    try:
        cap=cv2.VideoCapture(0)
        screenCnt=''
        
    except:
        print('cannot open camera')
        return
    while True:
        global frame
        ret, frame1 = cap.read()
        
        frame=frame1[10:800, 50:350]
        if not ret:
            print('cannot open camera')
            break
        
        k = cv2.waitKey(10)
        if k == 27:
            break

        gray = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY)
        gray = cv2.GaussianBlur(frame, (5, 5), 0)
        edged = cv2.Canny(frame, 80, 200)


        (_, cnts, _) = cv2.findContours(edged.copy(), cv2.RETR_LIST, cv2.CHAIN_APPROX_SIMPLE)
        cnts = sorted(cnts, key = cv2.contourArea, reverse = True)[:5]
        
        # loop over the contours
        for c in cnts:
            # approximate the contour
            peri = cv2.arcLength(c, True)
            approx = cv2.approxPolyDP(c, 0.02 * peri, True)
        
            

            # if our approximated contour has four points, then we
            # can assume that we have found our screen
            if len(approx) == 4:
                contourSize = cv2.contourArea(approx)
                camSize = frame.shape[0] * frame.shape[1]
                ratio = contourSize / camSize
                ###print (contourSize)
                ##print (camSize)
                ##print (ratio)
                screenCnt = []
                
                if ratio > 0.01:
                    screenCnt=approx
                    
                break
            
        if len(screenCnt)==0:
            cv2.imshow('webcam', frame)
            continue

        else:
            cv2.drawContours(frame,[screenCnt], -1, (0,255,0), 5)
            cv2.imshow('webcam', frame)

            rect = order_points(screenCnt.reshape(4, 2))
            (topLeft, topRight, bottomRight, bottomLeft) = rect
    
            w1 = abs(bottomRight[0] - bottomLeft[0])
            w2 = abs(topRight[0] - topLeft[0])
            h1 = abs(topRight[1] - bottomRight[1])
            h2 = abs(topLeft[1] - bottomLeft[1])
            maxWidth = max([w1, w2])
            maxHeight = max([h1, h2])
    
            dst = np.float32([[0,0], [maxWidth-1,0], 
                                  [maxWidth-1,maxHeight-1], [0,maxHeight-1]])
    
            M = cv2.getPerspectiveTransform(rect, dst)
            warped = cv2.warpPerspective(frame, M, (maxWidth, maxHeight))

            warped=cv2.cvtColor(warped,cv2.COLOR_BGR2GRAY)
            warped = cv2.adaptiveThreshold(warped, 255, cv2.ADAPTIVE_THRESH_MEAN_C, cv2.THRESH_BINARY, 21, 10)
            
            break
    
    cap.release()

    cv2.imwrite('scannedbooktag.png', warped)

    headers = {
        # Request headers
        'Content-Type': 'application/octet-stream',
        'Ocp-Apim-Subscription-Key': '8b19b33ce85d4fda96c9ffd825105795',
    }
    params = urllib.parse.urlencode({
        # Request parameters
        'language': 'en',
        'detectOrientation ': 'true',
    })
    data = open('scannedbooktag.png', 'rb').read()
    
    try:
        image_file = 'scannedbooktag.png'
        ocr_project_oxford(headers, params, data)
    except Exception as e:
        print(e)

    
if __name__ == '__main__':
 

    frame=''
    while True :

        print ("Enter the book/book tag you want to find:")
        t2=input('')
     ##########sectionA #########
        if t2== ("1") or t2 == ("Control Systems Engineering") or t2 == ('GR439478') or t2==('gr439478'):
            a='GR439478'
            print ("You entered "+t2+',It located "section A", book tag : '+a)
            break
        elif t2== ("2") or t2 == ('제어시스템공학') or t2 == ('NA793292') or t2 ==('na793292'):
            a='NA793292'
            print ("You entered "+t2+',located "section A", book tag : '+a)
            break
        elif t2== ("3") or t2 == ('이동통신공학') or t2 == ('ZD276468') or t2==('zd276468'):
            a='ZD276468'
            print ("You entered "+t2+',located "section A", book tag : '+a)
            break
        elif t2== ("4") or t2 == ('고체전자공학') or t2 == ('UZ637949') or t2 ==('uz637949'):
            a='UZ637949'
            print ("You entered "+t2+',located "section A", book tag : '+a)
            break
    #############sectionB##########
        elif t2== ("5")  or t2== ("전자회로") or t2 == ('FU462694') or t2==('fu462694'):
            a='FU462694'
            print ("You entered "+t2+',located "section B", book tag : '+a)
            break
    ########sectionC#########
        elif t2== ("6") or t2== ("선형대수") or t2 == ('LR336246') or t2==('lr336246'):
            a='LR336246'
            print ("You entered "+t2+',located "section C", book tag : '+a)
            break
        elif t2== ("7") or t2== ("이공계대학수학") or t2 == ('EP398946') or t2 ==('ep398946'):
            a='EP398946'
            print ("You entered "+t2+',located "section C", book tag : '+a)
            break
        elif t2== ("8") or t2== ("수학의 정석") or t2 == ('AQ682989') or t2==('aq682989'):
            a='AQ682989'
            print ("You entered "+t2+',located "section C", book tag : '+a)
            break
     
        else:
           
            print('no information')
            

        
    
    
    while True :
        
        t1=''

        
        auto_scan_image_via_webcam()
        
        
        
            
        if t1==a :
            print ("recognized tag :"+t1)
            break

        
    #빈칸 입력되면 아주짧게 삡 잘못된 인식이면 삐입 맞으면 삡삡
                
        elif t1=='':
            print ("Please focus")
            duration = 400
            freq=150
            winsound.Beep(freq, duration)
        else:
            print ("recognized tag :" +t1+ ", does not match")
            duration = 400
            freq=440
            winsound.Beep(freq, duration)

    print ("Found it!")
    cv2.putText(frame, 'THIS ONE!', (50, 100), cv2.FONT_HERSHEY_SIMPLEX, 1.5, (0, 255, 0), 4)
    cv2.imshow("webcam", frame)
    duration = 400
    freq=800
    winsound.Beep(freq, duration)  
    winsound.Beep(freq, duration)
    





    cv2.waitKey(0)
    cv2.destroyAllWindows()
    cv2.waitKey(1)

professor : seungeun lee

member : yoojin kim, jihye son