def read_user_dots(path, number_points, x, y, w, h, shape):
dco = DetectContours()
points = []
for i in range(0, number_points):
#if i < 10:
# i = str(0) + str(i)
print path + "/" + str(i + 1) + ".jpg"
img = cv2.imread(path + "/" + str(i + 1) + ".jpg")
# make image for userpt_locations
blank = np.zeros(shape)
f_x = 1.0 * w / img.shape[1]
f_y = 1.0 * h / img.shape[0]
print img.shape
print f_y, f_x
print w / img.shape[1], h / img.shape[0]
sys.stdout.flush()
dots = cv2.resize(img, (None),
fx=f_x,
fy=f_y,
interpolation=cv2.INTER_LINEAR)
blank[y:int(y + h), x:int(x + w)] = dots
print blank.dtype
sys.stdout.flush()
a = dco.getContours(np.asarray(blank, np.uint8))
#dci.get_circles(img)
points.append(dco.getCentroids(a))
#points.append(dci.get)
return points
def read_cam_dots(path, number_points):
dch = DetectChanges()
dco = DetectContours()
points = []
bg = cv2.imread(path + "/0.jpg")
for i in range(0, number_points):
img = cv2.imread(path + "/" + str(i + 1) + ".jpg")
a = dch.getContours(bg, img)
points.append(dco.getCentroids(a))
return points
def read_dots(path, number_points):
dco = DetectContours()
points = []
for i in range(0, number_points):
#if i < 10:
# i = str(0) + str(i)
img = cv2.imread(path + "/" + str(i + 1) + ".jpg")
a = dco.getContours(img)
#dci.get_circles(img)
points.append(dco.getCentroids(a))
#points.append(dci.get)
return points
import cv2
import cv2.cv as cv
import numpy as np
import math
import sys
import os
import imutils
from util import *
from getDots import *
from detectPattern import DetectContours
from detectPattern import DetectCircles
from findHomography import FindHomography
dco = DetectContours()
dci = DetectCircles()
fh = FindHomography()
# TODO: automate getting the centroids of all the original calibration images
orig12 = [(30, 30), (283, 30), (536, 30), (789, 30), (30, 310), (283, 310),
(536, 310), (789, 310), (30, 589), (283, 589), (536, 589),
(789, 589)]
orig18 = [(30, 30), (182, 30), (334, 30), (486, 30), (638, 30), (789, 30),
(30, 310), (182, 310), (334, 310), (486, 310), (638, 310),
(789, 310), (30, 589), (182, 589), (334, 589), (486, 589),
(638, 589), (789, 589)]
#(1) translation from original to points -> get in between pixels from bilinear interpolation
#(2) homography from 4 points (1245) (2356)
import numpy as np
import math
import sys
import os
import imutils
from util import *
from getDotsOld import *
from detectPattern import DetectContours
from detectPattern import DetectCircles
from findHomography import FindHomography
#from computeMapping import interpolate_colour, reverse_warp_helper, warp_image
# 1) check that i warp rectangle to camera image
number_points = 28
dco = DetectContours()
fh = FindHomography()
cam_points = []
cam_shape = []
x = 100
y = 50
w = 300
h = 200
#userpt_locations = read_user_dots(os.getcwd() + "/user/generated/" + sys.argv[3], number_points, x,y,w,h,
# cv2.imread(os.getcwd() + "/user/camera/" + sys.argv[3] + "/0.jpg").shape)
#origpoints = read_dots(os.getcwd() + "/user/generated/" + sys.argv[3], number_points) #camera points
for i in range(0, number_points):
#if i < 10:
# i = str(0) + str(i)
img = cv2.imread("images/" + sys.argv[1] + "/" + str(i) + ".jpg")