From 02e0f7bc51acfa06e7299919b54b56a3c7eca02b Mon Sep 17 00:00:00 2001 From: Samo Penic <samo.penic@gmail.com> Date: Fri, 16 Nov 2018 18:22:33 +0000 Subject: [PATCH] Initial version of sid recognition. Cutting the numerals is not perfect yet. --- aoiOcr.py | 6 ++ sid_process.py | 33 ++++++++++++++-- .idea/sonarIssues.xml | 5 ++ Ocr.py | 21 +++++++--- 4 files changed, 54 insertions(+), 11 deletions(-) diff --git a/.idea/sonarIssues.xml b/.idea/sonarIssues.xml index c014979..60f73dd 100644 --- a/.idea/sonarIssues.xml +++ b/.idea/sonarIssues.xml @@ -13,6 +13,11 @@ <set /> </value> </entry> + <entry key="/a.dummy"> + <value> + <set /> + </value> + </entry> <entry key="$PROJECT_DIR$/Ocr.py"> <value> <set /> diff --git a/Ocr.py b/Ocr.py index 0da4497..35c6729 100644 --- a/Ocr.py +++ b/Ocr.py @@ -1,17 +1,18 @@ from pyzbar.pyzbar import decode -from sid_process import enhanceSID +from sid_process import getSID import cv2 import numpy as np import math class Paper: - def __init__(self, filename=None): + def __init__(self, filename=None, sid_classifier=None): self.filename = filename self.invalid = None self.QRData = None self.errors = [] self.warnings = [] + self.sid_classifier=sid_classifier if filename is not None: self.loadImage(filename) self.runOcr() @@ -137,8 +138,8 @@ loc_filtered_x, loc_filtered_y = zip( *sorted(zip(loc_filtered_x, loc_filtered_y)) ) - # loc=[loc_filtered_y,loc_filtered_x] - # remove duplicates + # loc=[loc_filtered_y,loc_filtered_x] + # remove duplicates a = np.diff(loc_filtered_x) > 40 a = np.append(a, True) loc_filtered_x = np.array(loc_filtered_x) @@ -213,5 +214,13 @@ self.answerMatrix.append(oneline) def get_enhanced_sid(self): - es= enhanceSID(self.img[int(0.04*self.imgHeight):int(0.08*self.imgHeight), int(0.7*self.imgWidth):int(0.99*self.imgWidth)]) - cv2.imwrite("enhancedSID.png",es) \ No newline at end of file + if self.sid_classifier is None: + return "x" + es = getSID( + self.img[ + int(0.045 * self.imgHeight) : int(0.085 * self.imgHeight), + int(0.7 * self.imgWidth) : int(0.99 * self.imgWidth), + ], + self.sid_classifier, + ) + return es diff --git a/aoiOcr.py b/aoiOcr.py index c8bc97b..72ed3b9 100644 --- a/aoiOcr.py +++ b/aoiOcr.py @@ -1,8 +1,12 @@ from Ocr import Paper +from sklearn.externals import joblib +classifier = joblib.load('filename.joblib') #p=Paper(filename='testpage300dpi_scan1.png') -p=Paper(filename='sizif111.tif') +p=Paper(filename='sizif111.tif', sid_classifier=classifier) +#p=Paper(filename='processed_scans/20141016095134535_0028.tif') + print(p.QRData) print(p.errors) diff --git a/sid_process.py b/sid_process.py index 210cfe7..90d9b33 100644 --- a/sid_process.py +++ b/sid_process.py @@ -1,6 +1,10 @@ import cv2 import numpy as np from skimage import morphology,img_as_ubyte +from sklearn import svm +from sklearn.externals import joblib + + """ (1) The text is an array of chars (in row-major order) where @@ -57,19 +61,18 @@ return np.ones((x, y), np.uint8) -def enhanceSID(image): +def getSID(image, classifier): image=255-image image=img_as_ubyte(image>100) cv2.imwrite("enSID0.png", image) # Remove noise image = cv2.morphologyEx(image, cv2.MORPH_OPEN, kernel(2,2), iterations=1) # Closing. Connect non connected parts - image = cv2.morphologyEx(image, cv2.MORPH_CLOSE, kernel(5, 5), iterations=2) + image = cv2.morphologyEx(image, cv2.MORPH_CLOSE, kernel(5, 3), iterations=4) # Again noise removal after closing + image = cv2.morphologyEx(image, cv2.MORPH_OPEN, kernel(8,8), iterations=1) # Skeletonization - ##For thinning I am using erosion - ##image = cv2.erode(image,kernel(4,4),iterations = 40) image = img_as_ubyte(morphology.thin(image>128)) cv2.imwrite("enSID1.png",image) # Stub removal (might not be necessary if thinning instead of skeletonize is used above @@ -80,4 +83,26 @@ # Thining again image = img_as_ubyte(morphology.skeletonize(image>0.5)) image = cv2.morphologyEx(image, cv2.MORPH_DILATE, kernel(10, 10)) + + im2,ctrs, hier = cv2.findContours(image.copy(), cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE) + sorted_ctrs = sorted(ctrs, key=lambda ctr: cv2.boundingRect(ctr)[0]) + + #classifier = joblib.load('filename.joblib') + + sid_no="" + for i, ctr in enumerate(sorted_ctrs): + # Get bounding box + x, y, w, h = cv2.boundingRect(ctr) + # Getting ROI + if(w<h/2): + sid_no=sid_no+"1" + continue + roi = image[y:y+h, x:x+w] + roi = img_as_ubyte(roi < 128) + roi = cv2.resize(roi,(32,32)) + + #cv2.rectangle(image,(x,y),( x + w, y + h ),(0,255,0),2) + cv2.imwrite('sid_no_{}.png'.format(i), roi) + sid_no=sid_no+str(classifier.predict(roi.reshape(1,-1)/255.0)[0]) + print(sid_no) return image -- Gitblit v1.9.3