3 files modified
1 files added
| | |
| | | |
| | | from glob import glob |
| | | |
| | | settings = {"sid_mask": "64xx0xxx", "answer_threshold": 0.25} |
| | | settings = {"sid_mask": "11x0xxxx", "answer_threshold": 0.25} |
| | | classifier = joblib.load(filepath) |
| | | |
| | | #p = Paper(filename="testpage300dpi_scan1.png") |
| | |
| | | "processed_scans/20160408140801098_0004.tif", |
| | | "processed_scans/20160510075445995_0026.tif" |
| | | ] |
| | | p=Paper(filename=pa[9], sid_classifier=classifier, settings=settings) |
| | | #p=Paper(filename=pa[9], sid_classifier=classifier, settings=settings) |
| | | p=Paper(filename='test3.tif', sid_classifier=classifier, settings=settings) |
| | | |
| | | # print(p.QRData) |
| | | # print(p.errors) |
| | |
| | | |
| | | print(p.get_paper_ocr_data()) |
| | | |
| | | |
| | | exit(0) |
| | | filelist = glob("processed_scans/*.tif") |
| | | wrong_sid=0; |
| | | total=0 |
| | |
| | | if total%10 == 0: |
| | | print("Total:{}, wrong SID: {}".format(total,wrong_sid)) |
| | | |
| | | print("Total:{}, wrong SID: {}".format(total,wrong_sid)) |
| | | print("Total:{}, wrong SID: {}".format(total,wrong_sid)) |
| | |
| | | import os |
| | | import pkg_resources |
| | | |
| | | markerfile = '/template.png' # always use slash |
| | | markerfile = '/template-sq.png' # always use slash |
| | | markerfilename = pkg_resources.resource_filename(__name__, markerfile) |
| | | |
| | | |
| | |
| | | return |
| | | self.decodeQRandRotate() |
| | | self.imgTreshold() |
| | | cv2.imwrite('/tmp/debug_threshold.png', self.bwimg) |
| | | skewAngle = 0 |
| | | # try: |
| | | # skewAngle=self.getSkewAngle() |
| | |
| | | self.imgHeight, self.imgWidth = self.img.shape[0:2] |
| | | |
| | | # todo, make better tresholding |
| | | |
| | | def imgTreshold(self): |
| | | (self.thresh, self.bwimg) = cv2.threshold( |
| | | self.img, 128, 255, cv2.THRESH_BINARY | cv2.THRESH_OTSU |
| | | self.img, 128, 255, |
| | | cv2.THRESH_BINARY | cv2.THRESH_OTSU |
| | | ) |
| | | |
| | | def getSkewAngle(self): |
| | |
| | | def locateUpMarkers(self, threshold=0.85, height=200): |
| | | template = cv2.imread(markerfilename, 0) |
| | | w, h = template.shape[::-1] |
| | | crop_img = self.img[0:height, :] |
| | | crop_img = self.bwimg[0:height, :] |
| | | res = cv2.matchTemplate(crop_img, template, cv2.TM_CCOEFF_NORMED) |
| | | loc = np.where(res >= threshold) |
| | | cimg = cv2.cvtColor(crop_img, cv2.COLOR_GRAY2BGR) |
| | |
| | | def locateRightMarkers(self, threshold=0.85, width=200): |
| | | template = cv2.imread(markerfilename, 0) |
| | | w, h = template.shape[::-1] |
| | | crop_img = self.img[:, -width:] |
| | | crop_img = self.bwimg[:, -width:] |
| | | cv2.imwrite('/tmp/debug_right.png', crop_img) |
| | | res = cv2.matchTemplate(crop_img, template, cv2.TM_CCOEFF_NORMED) |
| | | loc = np.where(res >= threshold) |
| | | cimg = cv2.cvtColor(crop_img, cv2.COLOR_GRAY2BGR) |
| | |
| | | loc_filtered_y.append(pt[1]) |
| | | loc_filtered_x.append(pt[0]) |
| | | # order by y coordinate |
| | | loc_filtered_y, loc_filtered_x = zip( |
| | | *sorted(zip(loc_filtered_y, loc_filtered_x)) |
| | | ) |
| | | try: |
| | | loc_filtered_y, loc_filtered_x = zip( |
| | | *sorted(zip(loc_filtered_y, loc_filtered_x)) |
| | | ) |
| | | except: |
| | | self.yMarkerLocations=[np.array([1,1]),np.array([1,2])] |
| | | return self.yMarkerLocations |
| | | # loc=[loc_filtered_y,loc_filtered_x] |
| | | # remove duplicates |
| | | a = np.diff(loc_filtered_y) > 40 |
| | |
| | | es, err, warn = getSID( |
| | | self.img[ |
| | | int(0.04 * self.imgHeight) : int(0.095 * self.imgHeight), |
| | | int(0.7 * self.imgWidth) : int(0.99 * self.imgWidth), |
| | | int(0.65 * self.imgWidth) : int(0.95 * self.imgWidth), |
| | | ], |
| | | self.sid_classifier, |
| | | sid_mask, |
| | |
| | | data = qrdata.split(",") |
| | | retval = { |
| | | "exam_id": int(data[1]), |
| | | "page_no": int(data[3])+1, |
| | | "page_no": int(data[3]), |
| | | "paper_id": int(data[2]), |
| | | "faculty_id": int(data[0]), |
| | | "sid": None |
| | | } |
| | | if len(data) > 4: |
| | | retval["sid"] = data[4] |
| | |
| | | data["errors"] = self.errors |
| | | data["warnings"] = self.warnings |
| | | data["up_position"] = ( |
| | | list(self.xMarkerLocations[1] / self.imgWidth), |
| | | list(self.yMarkerLocations[1] / self.imgHeight), |
| | | list(self.xMarkerLocations[0] / self.imgWidth), |
| | | list(self.xMarkerLocations[1] / self.imgHeight), |
| | | ) |
| | | data["right_position"] = ( |
| | | list(self.xMarkerLocations[1] / self.imgWidth), |
| | | list(self.yMarkerLocations[0] / self.imgWidth), |
| | | list(self.yMarkerLocations[1] / self.imgHeight), |
| | | ) |
| | | data["ans_matrix"] = ( |
| | |
| | | |
| | | def find_biggest_blob(image, original_image,sid_mask): |
| | | if sid_mask[0] == "1": |
| | | move_left = 45 |
| | | move_left = 35 |
| | | elif sid_mask[0] == "x": |
| | | move_left = 55 |
| | | move_left = 40 |
| | | else: |
| | | move_left = 0 |
| | | # Remove noise |
| | | image2 = cv2.morphologyEx( |
| | | original_image, cv2.MORPH_OPEN, kernel(2, 2), iterations=7 |
| | | original_image, cv2.MORPH_OPEN, kernel(2, 2), iterations=3 |
| | | ) |
| | | # find biggest block of pixels |
| | | image1 = cv2.morphologyEx(image2, cv2.MORPH_DILATE, kernel(5, 25), iterations=4) |
| | |
| | | sid_err = [] |
| | | image = 255 - image |
| | | image_original = image.copy() |
| | | image = img_as_ubyte(image > 100) |
| | | image = img_as_ubyte(image > 70) |
| | | cv2.imwrite("/tmp/enSID0.png", image) |
| | | |
| | | # Remove noise |
| | | image = cv2.morphologyEx(image, cv2.MORPH_OPEN, kernel(2, 2), iterations=3) |
| | | #image = cv2.morphologyEx(image, cv2.MORPH_OPEN, kernel(2, 2), iterations=3) |
| | | |
| | | # Closing. Connect non connected parts |
| | | image = cv2.morphologyEx(image, cv2.MORPH_CLOSE, kernel(5, 3), iterations=4) |
| | | image = cv2.morphologyEx(image, cv2.MORPH_CLOSE, kernel(5, 1), iterations=4) |
| | | |
| | | # Again noise removal after closing |
| | | # image = cv2.morphologyEx(image, cv2.MORPH_OPEN, kernel(8, 8), iterations=1) |