In [3]:
import cv2
import numpy as np
import matplotlib.pyplot as plt
import os
%matplotlib inline
In [139]:
image = cv2.imread('./1.jpg')
In [140]:
plt.imshow(image)
Out[140]:
In [166]:
image_rgb = cv2.cvtColor(image, cv2.COLOR_BGR2RGB)
In [167]:
plt.imshow(image_rgb)
Out[167]:
In [129]:
image_gray = cv2.cvtColor(image_rgb, cv2.COLOR_RGB2GRAY)
In [130]:
plt.imshow(image_gray, cmap='gray')
Out[130]:
In [131]:
thresh_image = cv2.threshold(image_gray, 100, 255, cv2.THRESH_BINARY)[1]
plt.imshow(thresh_image, cmap='gray')
Out[131]:
In [132]:
find_image, contours, hierarchy = cv2.findContours(thresh_image, cv2.RETR_TREE,cv2.CHAIN_APPROX_SIMPLE)
In [133]:
plt.imshow(find_image, cmap='gray')
Out[133]:
In [134]:
draw_image = cv2.drawContours(find_image, contours, -1, (0, 0, 255), 3)
In [135]:
plt.imshow(draw_image, cmap='gray')
Out[135]:
In [145]:
draw_image = cv2.drawContours(image_rgb, contours, -1, (0, 255, 0), 3)
In [146]:
plt.imshow(draw_image)
Out[146]:
In [162]:
print(contours[4])
In [147]:
len(contours)
Out[147]:
In [148]:
th_area = image_rgb.shape[0] * image_rgb.shape[1] / 100
contours_large = list(filter(lambda c:cv2.contourArea(c) > th_area, contours))
In [151]:
def getRectByPoints(points):
# prepare simple array
points = list(map(lambda x: x[0], points))
points = sorted(points, key=lambda x:x[1])
top_points = sorted(points[:2], key=lambda x:x[0])
bottom_points = sorted(points[2:4], key=lambda x:x[0])
points = top_points + bottom_points
left = min(points[0][0], points[2][0])
right = max(points[1][0], points[3][0])
top = min(points[0][1], points[1][1])
bottom = max(points[2][1], points[3][1])
return (top, bottom, left, right)
def getPartImageByRect(rect):
img = cv2.imread('./1.jpg', 1)
return img[rect[0]:rect[1], rect[2]:rect[3]]
In [168]:
outputs = []
rects = []
approxes = []
for (i,cnt) in enumerate(contours_large):
arclen = cv2.arcLength(cnt, True) # 面積の計算
approx = cv2.approxPolyDP(cnt, 0.02*arclen, True) # 周囲長を計算（要は多角形の辺の総和）
# 小さいやつは除外
if len(approx) < 4:
continue
approxes.append(approx)
x, y, w, h = cv2.boundingRect(cnt)
outputs.append(cv2.rectangle(image_rgb,(x,y),(x+w,y+h), (0,255,0), 2))
In [173]:
len(outputs)
Out[173]:
In [171]:
plt.imshow(outputs[0])
Out[171]:
In [152]:
outputs = []
rects = []
approxes = []
for (i,cnt) in enumerate(contours_large):
arclen = cv2.arcLength(cnt, True) # 面積の計算
approx = cv2.approxPolyDP(cnt, 0.02*arclen, True) # 周囲長を計算（要は多角形の辺の総和）
# 小さいやつは除外
if len(approx) < 4:
continue
approxes.append(approx)
rect = getRectByPoints(approx)
rects.append(rect)
outputs.append(getPartImageByRect(rect))
In [153]:
len(rects)
Out[153]:
In [154]:
rects[0]
Out[154]:
In [155]:
len(outputs)
Out[155]:
In [157]:
plt.imshow(outputs[1])
Out[157]:
In [105]:
draw_image = cv2.drawContours(image_rgb, approxes, -1, (255, 0, 0), 3)
plt.imshow(draw_image)
Out[105]:
In [192]:
image_2 = cv2.imread('./2.jpg')
In [193]:
image_2_gray = cv2.cvtColor(image_2, cv2.COLOR_BGR2GRAY)
In [194]:
plt.imshow(image_2_gray, cmap='gray')
Out[194]:
In [177]:
thresh_image = cv2.threshold(image_2_gray, 50, 255, cv2.THRESH_BINARY)[1]
plt.imshow(thresh_image, cmap='gray')
Out[177]:
In [178]:
im2, contours, hierarchy = cv2.findContours(thresh_image, cv2.RETR_TREE,cv2.CHAIN_APPROX_SIMPLE)
In [179]:
plt.imshow(im2, cmap='gray')
Out[179]:
In [195]:
_image_2 = image_2.copy()
draw_image = cv2.drawContours(_image_2, contours, -1, (0, 255, 0), 3)
In [196]:
plt.imshow(draw_image, cmap='gray')
Out[196]:
In [197]:
image_rgb = cv2.cvtColor(image_2, cv2.COLOR_BGR2RGB)
th_area = image_rgb.shape[0] * image_rgb.shape[1] / 100
contours_large = list(filter(lambda c:cv2.contourArea(c) > th_area, contours))
In [198]:
_image_2 = image_2.copy()
image_rgb = cv2.cvtColor(_image_2, cv2.COLOR_BGR2RGB)
plt.imshow(image_rgb)
Out[198]:
In [199]:
outputs = []
rects = []
approxes = []
for (i,cnt) in enumerate(contours_large):
arclen = cv2.arcLength(cnt, True) # 面積の計算
approx = cv2.approxPolyDP(cnt, 0.02*arclen, True) # 周囲長を計算（要は多角形の辺の総和）
# 小さいやつは除外
if len(approx) < 4:
continue
approxes.append(approx)
x, y, w, h = cv2.boundingRect(cnt)
outputs.append(cv2.rectangle(_image_2,(x,y),(x+w,y+h), (0,255,0), 2))
In [200]:
len(outputs)
Out[200]:
In [201]:
plt.imshow(outputs[0])
Out[201]:
In [59]:
import glob
files = []
for file in glob.glob('./opencv/samples/cpp/left*.jpg'):
files.append(file)
for file in glob.glob('./opencv/samples/cpp/right*.jpg'):
files.append(file)
In [60]:
files
Out[60]:
In [75]:
# termination criteria
criteria = (cv2.TERM_CRITERIA_EPS + cv2.TERM_CRITERIA_MAX_ITER, 30, 0.001)
# prepare object points, like (0,0,0), (1,0,0), (2,0,0) ....,(6,5,0)
objp = np.zeros((6*7,3), np.float32)
objp[:,:2] = np.mgrid[0:7,0:6].T.reshape(-1,2)
# Arrays to store object points and image points from all the images.
objpoints = [] # 3d point in real world space
imgpoints = [] # 2d points in image plane.
for fname in files:
img = cv2.imread(fname)
gray = cv2.cvtColor(img,cv2.COLOR_BGR2GRAY)
# Find the chess board corners
ret, corners = cv2.findChessboardCorners(gray, (7,6),None)
# If found, add object points, image points (after refining them)
if ret == True:
objpoints.append(objp)
corners2 = cv2.cornerSubPix(gray,corners,(11,11),(-1,-1),criteria)
imgpoints.append(corners2)
In [76]:
np.array(imgpoints).shape
Out[76]:
In [77]:
# キャリブレーション
ret, mtx, dist, rvecs, tvecs = cv2.calibrateCamera(objpoints, imgpoints, gray.shape[::-1],None,None)
In [72]:
# 歪み補正
img = cv2.imread(files[1])
h, w = img.shape[:2]
newcameramtx, roi=cv2.getOptimalNewCameraMatrix(mtx,dist,(w,h),1,(w,h))
In [73]:
# 歪補正
dst = cv2.undistort(img, mtx, dist, None, newcameramtx)
# 画像の切り落とし
# x,y,w,h = roi
# dst = dst[y:y+h, x:x+w]
plt.figure(figsize=(20, 10))
plt.subplot(1, 2, 1)
plt.imshow(dst)
plt.subplot(1, 2, 2)
plt.imshow(cv2.imread(files[1]))
Out[73]:
In [74]:
# 歪補正
mapx,mapy = cv2.initUndistortRectifyMap(mtx,dist,None,newcameramtx,(w,h),5)
dst = cv2.remap(img,mapx,mapy,cv2.INTER_LINEAR)
# 画像の切り落とし
x,y,w,h = roi
# dst = dst[y:y+h, x:x+w]
plt.figure(figsize=(20, 10))
plt.subplot(1, 2, 1)
plt.imshow(dst)
plt.subplot(1, 2, 2)
plt.imshow(cv2.imread(files[1]))
Out[74]:
In [78]:
# 歪み補正
img = cv2.imread('./2.jpg')
h, w = img.shape[:2]
newcameramtx, roi=cv2.getOptimalNewCameraMatrix(mtx,dist,(w,h),1,(w,h))
In [80]:
# 歪補正
dst = cv2.undistort(img, mtx, dist, None, newcameramtx)
# 画像の切り落とし
x,y,w,h = roi
dst = dst[y:y+h, x:x+w]
plt.figure(figsize=(20, 10))
plt.subplot(1, 2, 1)
plt.imshow(dst)
plt.subplot(1, 2, 2)
plt.imshow(cv2.imread('./2.jpg'))
Out[80]:
In [81]:
# 歪補正
mapx,mapy = cv2.initUndistortRectifyMap(mtx,dist,None,newcameramtx,(w,h),5)
dst = cv2.remap(img,mapx,mapy,cv2.INTER_LINEAR)
# 画像の切り落とし
x,y,w,h = roi
dst = dst[y:y+h, x:x+w]
plt.figure(figsize=(20, 10))
plt.subplot(1, 2, 1)
plt.imshow(dst)
plt.subplot(1, 2, 2)
plt.imshow(cv2.imread('./2.jpg'))
Out[81]:
In [ ]: