Develop Reference

Solving for intersection points between parametric curves

The parametric equations of the two curves are as follows:
Curve1: r(t) = (2(t-sin(t)),2(1 -cos(t)))
Curve2: s(t) = (2t - sin(t),2 - cos(t))
I need to find the points of intersection in the region [0,4π].
I was able to plot the graph for the mentioned region and observed 4 points of intersection. But I am not able to determine the exact points of intersection.
For non-parametric equations, fsolve from sympy can be used, but the curves which are given in their parametric forms, I am not able to find a workaround.
t = np.arange(-0.25*np.pi,4.25*np.pi,0.01)
rx = np.zeros(len(t))
ry = np.zeros(len(t))
for i in t:
rx = 2*(t - np.sin(t))
ry = 2*(1 - np.cos(t))
sx = np.zeros(len(t))
sy = np.zeros(len(t))
for i in t:
sx = 2*t - np.sin(t)
sy = 2 - np.cos(t)
plt.plot(rx,ry)
plt.plot(sx,sy)

For a given x you can find t for each curve and see if the corresponding y are the same. You can step over the x range with some grid looking for such locations where the thee curves hit and use bisection to zero in on a more precise x. Since you can't solve the parametrix x(t) - x for t, nsolve will have to be used to find an approximate t. Something like this finds values for your 4 roots (confirmed graphically) after correcting your OP equation for Curve1 to be the same as in the code you wrote afterwards.
f = lambda xx: a[1].subs(t, tt)-b[1].subs(t,nsolve(b[0]-xx,tlast))
tlast = 0 # guess for t for a given xx to be updated as we go
tol = 1e-9 # how tight the bounds on x must be for a solution
dx = 0.1
for ix in range(300):
xx = ix*dx
tt=nsolve(a[0]-xx,tlast)
y2 = f(xx)
if ix != 0 and yold*y2 < 0 and tt<4*pi:
tlast = tt # updating guess for t
# bisect for better xx now that bounding xx are found
x1 = xx-dx
x2 = xx
y1 = yold
while x2 - x1 > tol:
xm = (x1 + x2)/2
ym = f(xm)
if ym*y1 < 0:
y2 = ym
x2 = xm
elif ym != 0:
y1 = ym
x1 = xm
else:
break
print(xm) # a solution
yold = y2
I don't know of a more automated way to do this in SymPy.

The two curves don't intersect at the same time (that would be points where sin(t) = cos(t) = 0, which has no solutions). So you really want to know when
R = (2*t1 - 2*sin(t1), 2 - 2*cos(t1))
S = (2*t2 - sin(t2), 2 - cos(t2))
intersect.
This is two equations with two unknowns, so it is straightforward to solve with sympy.nsolve. You have to fiddle with the starting values a bit to find the ones that converge to different solutions. If you know what they are roughly from the graph that is the best place to start.
>>> t1, t2 = symbols('t1 t2')
>>> R = (2*t1 - 2*sin(t1), 2 - 2*cos(t1))
>>> S = (2*t2 - sin(t2), 2 - cos(t2))
>>> nsolve([R[0] - S[0], R[1] - S[1]], [t1, t2], [1, 1])
Matrix([
[ 1.09182358380672],
[0.398264297579454]])
>>> nsolve([R[0] - S[0], R[1] - S[1]], [t1, t2], [5, 5])
Matrix([
[5.19136172337286],
[5.88492100960013]])
>>> nsolve([R[0] - S[0], R[1] - S[1]], [t1, t2], [7, 7])
Matrix([
[7.37500889098631],
[6.68144960475904]])
>>> nsolve([R[0] - S[0], R[1] - S[1]], [t1, t2], [10, 10])
Matrix([
[11.4745470305524],
[12.1681063167797]])

How to plot array with condition (r < 1.5*rs)?

I'm coding photons orbiting a black hole. When the photons go right into the black hole, the trajectory is really weird due to a division by zero. I'd like to ignore the photons of my array with r < 1.5*rs but I don't know how
I've tried using while True and if, it didn't work
h0=[t0, r2, theta, phi2, pt2, pr2, ptheta, pphi2]
T = np.linspace(0, 1000, 9000)
zz=odeint(func, h0, T, args=(rs,))
r22 = zz[:, 1]
theta22 = zz[:, 2]
phi22 = zz[:, 3]
pt22 = zz[:, 4]
pr22 = zz[:, 5]
pphi22 = zz[:, 7]
def sph2cart(r, phi, theta):
X = r * np.cos(phi) * np.sin(theta)
Y = r * np.sin(phi) * np.sin(theta)
Z = r * np.cos(theta)
return(X, Y, Z)
X2, Y2, Z2 = sph2cart(r22, phi22, theta22)
plt.plot(X2, Y2, Z2, 'g')
I don't think you really need the code to help me, but does anyone know how to plot X2, Y2, Z2 for radius r < 1.5*rs (rs is defined in the code)?

Numpy has built in masks which work quite well
r=np.arange(20)
r_mask = np.ma.masked_where(r < 10 , r)
So, since you want to generate X,Y, and Z values only when r<1.5*rs (i think you said it the opposite way at one point, but you'll just need to flip the sign in any case), your code could look like this
import numpy as np
h0=[t0, r2, theta, phi2, pt2, pr2, ptheta, pphi2]
T = np.linspace(0, 1000, 9000)
zz=odeint(func, h0, T, args=(rs,))
r22 = zz[:, 1]
theta22 = zz[:, 2]
phi22 = zz[:, 3]
pt22 = zz[:, 4]
pr22 = zz[:, 5]
pphi22 = zz[:, 7]
def sph2cart(r, phi, theta):
X = r * np.cos(phi) * np.sin(theta)
Y = r * np.sin(phi) * np.sin(theta)
Z = r * np.cos(theta)
return(X, Y, Z)
r22_masked = np.ma.masked_where(r22 > 1.5*rs , r22)
X2, Y2, Z2 = sph2cart(r22_masked, phi22, theta22)
This would generate X2,Y2, and Z2 values only for the unmasked values of r22 (aka where r22 is less than 1.5*rs

pytorch (numpy) calculation about the closest pixels to points

I am trying to solve a complicated problem.
For example, I have a batch of 2D predicted images (softmax output, value between 0 and 1) with size: Batch x H x W and ground truth Batch x H x W
The light gray color pixels are the background with value 0, and the dark gray color pixels are the foreground with value 1. I try to compute the mass center coordinates using scipy.ndimage.center_of_mass on each ground truth image. Then I get the center location point C (red color) for each ground truth. The C points set is Batch x 1.
Now, for each pixel A (yellow color) in the predicted images, I want to get three pixels B1, B2, B3 (blue color) which are the closest to A on the line AC (here C is corresponding location of mass center in ground truth).
I used following code to get the three closest points B1, B2, B3.
def connect(ends, m=3):
d0, d1 = np.abs(np.diff(ends, axis=0))[0]
if d0 > d1:
return np.c_[np.linspace(ends[0, 0], ends[1, 0], m + 1, dtype=np.int32),
np.round(np.linspace(ends[0, 1], ends[1, 1], m + 1))
.astype(np.int32)]
else:
return np.c_[np.round(np.linspace(ends[0, 0], ends[1, 0], m + 1))
.astype(np.int32),
np.linspace(ends[0, 1], ends[1, 1], m + 1, dtype=np.int32)]
So the B points set is Batch x 3 x H x W.
Then, I want to compute like this: |Value(A)-Value(B1)|+|Value(A)-Value(B2)|+|Value(A)-Value(B3)|. The size of the result should be Batch x H x W.
Is there any numpy vectorization tricks that can be used to update the value of each pixel in predicted images? Or can this be solved using pytorch functions? I need to find a method to update the whole image. The predicted image is the softmax output. I cannot use for loop to compute each single value since it will become non-differentiable. Thanks a lot.

As suggested by #Matin, you could consider Bresenham's algorithm to get your points on the AC line.
A simplistic PyTorch implementation could be as follows (directly adapted from the pseudo-code here ; could be optimized):
import torch
def get_points_from_low(x0, y0, x1, y1, num_points=3):
dx = x1 - x0
dy = y1 - y0
xi = torch.sign(dx)
yi = torch.sign(dy)
dy = dy * yi
D = 2 * dy - dx
y = y0
x = x0
points = []
for n in range(num_points):
x = x + xi
is_D_gt_0 = (D > 0).long()
y = y + is_D_gt_0 * yi
D = D + 2 * dy - is_D_gt_0 * 2 * dx
points.append(torch.stack((x, y), dim=-1))
return torch.stack(points, dim=len(x0.shape))
def get_points_from_high(x0, y0, x1, y1, num_points=3):
dx = x1 - x0
dy = y1 - y0
xi = torch.sign(dx)
yi = torch.sign(dy)
dx = dx * xi
D = 2 * dx - dy
y = y0
x = x0
points = []
for n in range(num_points):
y = y + yi
is_D_gt_0 = (D > 0).long()
x = x + is_D_gt_0 * xi
D = D + 2 * dx - is_D_gt_0 * 2 * dy
points.append(torch.stack((x, y), dim=-1))
return torch.stack(points, dim=len(x0.shape))
def get_points_from(x0, y0, x1, y1, num_points=3):
is_dy_lt_dx = (torch.abs(y1 - y0) < torch.abs(x1 - x0)).long()
is_x0_gt_x1 = (x0 > x1).long()
is_y0_gt_y1 = (y0 > y1).long()
sign = 1 - 2 * is_x0_gt_x1
x0_comp, x1_comp, y0_comp, y1_comp = x0 * sign, x1 * sign, y0 * sign, y1 * sign
points_low = get_points_from_low(x0_comp, y0_comp, x1_comp, y1_comp, num_points=num_points)
points_low *= sign.view(-1, 1, 1).expand_as(points_low)
sign = 1 - 2 * is_y0_gt_y1
x0_comp, x1_comp, y0_comp, y1_comp = x0 * sign, x1 * sign, y0 * sign, y1 * sign
points_high = get_points_from_high(x0_comp, y0_comp, x1_comp, y1_comp, num_points=num_points) * sign
points_high *= sign.view(-1, 1, 1).expand_as(points_high)
is_dy_lt_dx = is_dy_lt_dx.view(-1, 1, 1).expand(-1, num_points, 2)
points = points_low * is_dy_lt_dx + points_high * (1 - is_dy_lt_dx)
return points
# Inputs:
# (#todo: extend A to cover all points in maps):
A = torch.LongTensor([[0, 1], [8, 6]])
C = torch.LongTensor([[6, 4], [2, 3]])
num_points = 3
# Getting points between A and C:
# (#todo: what if there's less than `num_points` between A-C?)
Bs = get_points_from(A[:, 0], A[:, 1], C[:, 0], C[:, 1], num_points=num_points)
print(Bs)
# tensor([[[1, 1],
# [2, 2],
# [3, 2]],
# [[7, 6],
# [6, 5],
# [5, 5]]])
Once you have your points, you could retrieve their "values" (Value(A), Value(B1), etc.) using torch.index_select() (note that as of now, this method only accept 1D indices, so you need to unravel your data). All things put together, this would look like something such as the following (extending A from shape (Batch, 2) to (Batch, H, W, 2) is left for exercise...)
# Inputs:
# (#todo: extend A to cover all points in maps):
A = torch.LongTensor([[0, 1], [8, 6]])
C = torch.LongTensor([[6, 4], [2, 3]])
batch_size = A.shape[0]
num_points = 3
map_size = (9, 9)
map_num_elements = map_size[0] * map_size[1]
map_values = torch.stack((torch.arange(0, map_num_elements).view(*map_size),
torch.arange(0, -map_num_elements, -1).view(*map_size)))
# Getting points between A and C:
# (#todo: what if there's less than `num_points` between A-C?)
Bs = get_points_from(A[:, 0], A[:, 1], C[:, 0], C[:, 1], num_points=num_points)
# Get map values in positions A:
A_unravel = torch.arange(0, batch_size) * map_num_elements
A_unravel = A_unravel + A[:, 0] * map_size[1] + A[:, 1]
values_A = torch.index_select(map_values.view(-1), dim=0, index=A_unravel)
print(values_A)
# tensor([ 1, -4])
# Get map values in positions A:
A_unravel = torch.arange(0, batch_size) * map_num_elements
A_unravel = A_unravel + A[:, 0] * map_size[1] + A[:, 1]
values_A = torch.index_select(map_values.view(-1), dim=0, index=A_unravel)
print(values_A)
# tensor([ 1, -78])
# Get map values in positions B:
Bs_flatten = Bs.view(-1, 2)
Bs_unravel = (torch.arange(0, batch_size)
.unsqueeze(1)
.repeat(1, num_points)
.view(num_points * batch_size) * map_num_elements)
Bs_unravel = Bs_unravel + Bs_flatten[:, 0] * map_size[1] + Bs_flatten[:, 1]
values_B = torch.index_select(map_values.view(-1), dim=0, index=Bs_unravel)
values_B = values_B.view(batch_size, num_points)
print(values_B)
# tensor([[ 10, 20, 29],
# [-69, -59, -50]])
# Compute result:
res = torch.abs(values_A.unsqueeze(-1).expand_as(values_B) - values_B)
print(res)
# tensor([[ 9, 19, 28],
# [ 9, 19, 28]])
res = torch.sum(res, dim=1)
print(res)
# tensor([56, 56])

How to generate squares (randomly located, equally sized, randomly rotated) that don't intersect each other?

I've been working on generating a layer of randomly rotated and placed squares on a 1x1 grid. I have been able to generate a single square that is randomly placed and rotated on the grid, but I'm not sure how to improve the code to generate more random squares that do not intersect with each other. Current code seen below:
Example of my One Randomized Square
from math import cos, pi, sin
from random import randint
from matplotlib.mlab import frange
from matplotlib.pyplot import plot, axis, show
def flake_position_layer1(): #Determines the initial position of one corner of the square
x0 = randint(0, 100) / 100
y0 = randint(0, 100) / 100
theta = randint(0, 90) * pi / 180 #Angle of rotation for the square
return x0, y0, theta
def flake_shape(): #generates the other 3 corners of the square
x0, y0, z, theta = flake_position_layer1()
x1 = x0 + (0.1 * cos(theta))
x2 = x1 + (0.1 * cos((90 * pi/180) + theta))
x3 = x2 + (0.1 * cos((180 * pi/180) + theta))
y1 = y0 + (0.1 * sin(theta))
y2 = y1 + (0.1 * sin((90 * pi/180) + theta))
y3 = y2 + (0.1 * sin((180 * pi/180) + theta))
return x0, x1, x2, x3, y0, y1, y2, y3
def display(): #connects the 4 corners on a plot
x0, x1, x2, x3, y0, y1, y2, y3 = flake_shape()
return plot([x0, x1, x2, x3, x0], [y0, y1, y2, y3, y0])
display()
axis([0,1,0,1]) #1x1 grid
show()
I do not have a CS background (I'm an environmental engineering major) and I am extremely inexperienced with coding. Please give me any recommendations that you may have for me to try and tackle this problem with!

Math background
1. Algebra
1st degree function (or [Wikipedia]: Linear function) is a function ([Wikipedia]: function) whose:
Expression can be written as: f(x) = a * x + b (a and b constants, x variable)
Graphical representation is a straight line in the xOy plane ([Wikipedia]: Cartesian coordinate system)
2. Plane Geometry
A plane consists of an (infinite) number of points: let's refer to a point by its coordinates which can be referred to as:
abscissa or horizontal coordinate or (simply) x
ordinate or vertical coordinate or (simply) y
The points in the plane spread across 2 dimensions
In the plane, every point can be uniquely identified by its x and y
Some of the points in the plane might have some common characteristics: e.g. a bunch of points that are on a straight line... a point that is on a straight line satisfies the line straight line equation (which is an expression generally defined as ${function (from previous paragraph) result} = ${value})
So, in short: for a point P0(x0, y0), if y0 == f(x0), the point is located on that straight line (and more: depending on y0 being greater / lower than f(x0), P0 is located above / below the straight line in the xOy plane). Again, I want to state that for non linear functions, y = f(x) still applies (as it's the general equation formula), but other operators (e.g. <, >) don't
! Important Note !: everything discussed here applies to a variety of functions (don't want to say all), but I'm limiting to linear functions, for clarity's sake
A straight line is determined by 2 distinct points (e.g. P0(x0, y0), P1(x1, y1)) - the equation for that straight line would be y = a * x + b (in our example: y = ((y0 - y1) / (x0 - x1)) * x + (y0 - x0 * ((y0 - y1) / (x0 - x1)))); !! Of course it worth mentioning the "vertical" (parallel to Oy) line which is !! not a function !!
Example: having 2 distinct parallel (non Bolyai :) ) lines: f0(x) = a * x + b0 and f1(x) = a * x + b1 (a is the same for both lines - this is the condition for them to be parallel) and an external point P0(x0, y0) (that obviously doesn't belong to any of the lines). How to determine if P0 is between the 2 lines? Well, the point must be above (the lower) one and below the other (the higher one). Translated into math (considering f0 being the lower one):
y0 > f0(x0) (y0 - f0(x0) > 0)
y0 < f1(x0) (y0 - f1(x0) < 0)
From the above observations (and there may be more wisdom), this is the condition that the point coordinates should satisfy: (y0 - f0(x0)) * (y0 - f1(x0)) < 0
Going further: a square consists of 2 sets of parallel lines (its sides); if a point is between each lines pair, then the point is in the square.
code00.py:
#!/usr/bin/env python3
import sys
from random import random, seed
from math import pi, sin, cos, sqrt
import matplotlib.pyplot as plt
pi_2 = pi / 2
MINX = MINY = 0
MAXX = MAXY = 1
DEFAULT_SIDE = 0.1
DEFAULT_SAFETY_MARGIN = DEFAULT_SIDE * sqrt(2)
MAX_SQUARES = 30
__global_generation_counter = 0
def get_func_deg1(p0, p1):
(x0, y0), (x1, y1) = p0, p1
if x0 == x1:
return None
a = (y0 - y1)/(x0 - x1)
b = y0 - x0 * a
return lambda x: a * x + b
def is_point_in_square(p, sq):
x, y = p
p0, p1, p2, p3 = sq
side_func0 = get_func_deg1(p0, p1)
side_func1 = get_func_deg1(p1, p2)
side_func2 = get_func_deg1(p2, p3)
side_func3 = get_func_deg1(p3, p0)
if not side_func0 or not side_func1 or not side_func2 or not side_func3:
xmin = min(p0[0], p2[0])
xmax = max(p0[0], p2[0])
ymin = min(p0[1], p2[1])
ymax = max(p0[1], p2[1])
return xmin <= x <= xmax and ymin <= y <= ymax
return ((y - side_func0(x)) * (y - side_func2(x))) <= 0 and \
((y - side_func1(x)) * (y - side_func3(x))) <= 0
def squares_overlap(square0, square1):
for p0 in square0:
if is_point_in_square(p0, square1):
return True
for p1 in square1:
if is_point_in_square(p1, square0):
return True
xc0 = (square0[0][0] + square0[2][0]) / 2
yc0 = (square0[0][1] + square0[2][1]) / 2
if is_point_in_square((xc0, yc0), square1):
return True
# The "reverse center check" not needed, since squares are congruent
"""
xc1 = (square1[0][0] + square1[2][0]) / 2
yc1 = (square1[0][1] + square1[2][1]) / 2
if is_point_in_square((xc1, yc1), square0):
return True
"""
return False
def __generation_monitor():
global __global_generation_counter
__global_generation_counter += 1
def generate_random_point(minx=MINX, miny=MINY, maxx=MAXX, maxy=MAXY, safety_margin=DEFAULT_SAFETY_MARGIN):
if maxx - minx < 2 * safety_margin or maxy - miny < 2 * safety_margin:
print("MUEEE")
safety_margin = 0
x = safety_margin + random() * (maxx - minx - 2 * safety_margin)
y = safety_margin + random() * (maxy - miny - 2 * safety_margin)
__generation_monitor()
return x, y
def generate_random_angle(max_val=pi_2):
return random() * max_val
def generate_random_square(side=DEFAULT_SIDE, squares_to_avoid=()):
while 1:
restart = False
x0, y0 = generate_random_point()
angle = generate_random_angle()
x1 = x0 + side * cos(angle)
y1 = y0 + side * sin(angle)
angle += pi_2
x2 = x1 + side * cos(angle)
y2 = y1 + side * sin(angle)
angle += pi_2
x3 = x2 + side * cos(angle)
y3 = y2 + side * sin(angle)
ret = (x0, y0), (x1, y1), (x2, y2), (x3, y3)
for square in squares_to_avoid:
if squares_overlap(ret, square):
restart = True
if restart:
continue
return ret
def square_to_plot(square):
xs, ys = zip(square[0], square[1], square[2], square[3])
return xs + (xs[0],), ys + (ys[0],)
def main():
seed()
squares = list()
allow_overlapping = False # CHANGE to True to allow square to overlap
for _ in range(MAX_SQUARES):
#print("Generating:", _)
if allow_overlapping:
square = generate_random_square()
else:
square = generate_random_square(squares_to_avoid=squares)
squares.append(square)
plot_squares = tuple()
for sq in squares:
plot_squares += square_to_plot(sq)
print("STATS:\n Squares: {}\n Allow overlapping: {}\n Generated values: {}".format(MAX_SQUARES, allow_overlapping, __global_generation_counter))
plt.plot(*plot_squares)
plt.axis([MINX, MAXX, MINY, MAXY])
plt.show()
if __name__ == "__main__":
print("Python {:s} on {:s}\n".format(sys.version, sys.platform))
main()
Notes:
I didn't work with matplotlib before (actually, I pip installed it for this task)
General comments:
A point is represented by a tuple representing its coordinates: (x, y)
A square is a tuple consisting of 4 points (p0, p1, p2, p3)
get_func_deg1:
Returns the function that describes the line that contains the 2 points given as arguments
If the 2 points are on a line that is parallel to Oy (there's no "normal" function to describe it), simply return None
is_point_in_square:
Determines if a point is inside a square
Uses the logic explained above, except
For the special case when the square edges are parallel to Ox and Oy when it uses some simple arithmetic operations
squares_overlap:
Determines whether 2 squares overlap (I'm sure there are faster "algorithms")
Checks if any of the 1st square corners are inside the 2nd one
The other way around: checks if any of the 2nd square corners are inside the 1st one
Since the above 2 checks are not enough (imagine a regular [Wikipedia]: Octagon and unifying its vertices every 2nd one: there will be 2 squares neither having its corners in the other one, but sharing their "central" areas), also check that one square's center is inside the other one
generate_random_point:
Generates a point in the given bounding box
safety_margin specifies the (minimum) distance that the generated point should be away from any of the bounding box sides, in order that any square that will have this point as a corner, would fit entirely in the bounding box
generate_random_angle:
Generates a random angle between 0 and (π / 2)
generate_random_square:
Generates a random point, a random angle, and constructs a square starting from there, using the specified side
squares_to_avoid is a list of squares. After the square is generated, it is checked against every square from that list. If the 2 squares overlap, the square is regenerated
square_to_plot:
Converts a square (from a tuple of points) to matplotlib format (2 tuples consisting of xs and ys with the 1st element duplicated as the last)
main:
The main function
__generation_monitor (0):
Internal function used for profiling
In order to change the number of squares, modify MAX_SQUARES
Output:
[cfati#CFATI-5510-0:e:\Work\Dev\StackOverflow\q046081491]> "e:\Work\Dev\VEnvs\py_064_03.05.04_test0\Scripts\python.exe" code00.py
STATS:
Squares: 30
Allow overlapping: False
Generated values: 1135
Few words about the squares generation
As seen in the output, for 30 displayed squares, 1135 were generated (at this run). That is because they were overlapping
If changing from main allow_overlapping = True, Generated values in the output will match the number of squares (MAX_SQUARES)
If increasing MAX_SQUARES to values let's say higher than 50, the number of generated values will increase exponentially (so will the time needed to generate them), and the chance that the program will enter an infinite loop (because it won't be able to generate a square that doesn't overlap to another one) will grow as well

Okay, here is what I've come up with with a little help from the shapely package. Installation help is at the bottom here. The ultimate result:
Code walkthrough
distance is a helper function using the Point class from shapely to find the distance between two coordinates. Just a helper function for later.
Square instantiations a new polygon. It has 4 corners, each an (x,y) pair, one coordinate for its center, and a scalar value equal to half the distance of its diagonal.
test_overlap is pretty self explanatory by title. But logically what it does is this: find the distance from center-to-center between the two shapes. Then find the sum of the half-diagonal of each shape. If the center-to-center distance is greater than the sum, the squares cannot overlap.
Squares starts out with an empty container (empty list) and attempts to add squares to it. But for each possible new addition, it firsts tests that there is no overlap with existing squares.
Code
import math
import random
from shapely.geometry import Polygon, Point
def distance(a, b):
return Point(a).distance(Point(b))
class Square(object):
def __init__(self):
self.x0, self.y0 = random.random(), random.random()
theta = random.randint(0, 90) * math.pi / 180 # Angle of rotation
self.x1 = self.x0 + (0.1 * math.cos(theta))
self.x2 = self.x1 + (0.1 * math.cos((90 * math.pi/180) + theta))
self.x3 = self.x2 + (0.1 * math.cos((180 * math.pi/180) + theta))
self.y1 = self.y0 + (0.1 * math.sin(theta))
self.y2 = self.y1 + (0.1 * math.sin((90 * math.pi/180) + theta))
self.y3 = self.y2 + (0.1 * math.sin((180 * math.pi/180) + theta))
self.corners = ((self.x0, self.y0), (self.x1, self.y1),
(self.x2, self.y2), (self.x3, self.y3))
#property
def center(self):
"""(x, y) of the center of the polygon."""
return Polygon(self.corners).centroid.coords[0]
#property
def half_diag(self):
"""The distance of 1/2 the shape's diagonal (center-to-corner)."""
p0, p1, p2, p3 = self.corners
return 0.5 * distance(p0, p1) * math.sqrt(2)
def test_overlap(square1, square2):
"""Do two shapes overlap?
Note this is a 'conservative' test. May return True if they do not
(false positive), but will never return False if they do (false negative).
"""
# Distance between two centers
ctc = distance(square1.center, square2.center)
# Sum of half-diagonals
halfdiags = square1.half_diag + square2.half_diag
res = ctc < halfdiags
return res
class Squares(object):
def __init__(self):
self.squares = []
def add_square(self):
new_square = Square()
if not self.squares:
# Initial empty list/container - just add without any tests
self.squares.append(new_square)
else:
while True:
# Test that new_square overlaps with existing
res = [test_overlap(square, new_square) for square in self.squares]
if any(res):
# We have at least 1 case of overlap (1 True)
new_square = Square()
else:
# Safe to add
self.squares.append(new_square)
break
def plot_squares(self):
for square in self.squares:
(x0, y0), (x1, y1), (x2, y2), (x3, y3) = square.corners
plt.plot([x0, x1, x2, x3, x0], [y0, y1, y2, y3, y0])
Example
import itertools
%matplotlib inline
for _ in itertools.repeat(None, 10):
# Add 10 squares; you could also just build this into the class
sqs.add_square()
sqs.plot_squares()
Installing shapely
Install the Anaconda distribution if you don't already. Then just use conda-forge to install shapely. From cmd run:
conda config --add channels conda-forge
conda install shapely
Glaring deficiency
At a certain point, your container of squares gets filled up and there is minimal room left to add new shapes. Even if there is available space, the function is basically trial-and-error, so will take long at high shape counts. That happens at around 20-25 squares (in a 1.0x1.0 box) at the moment.

You need a function to determine whether two cube intersect.
from math import cos, pi, sin
from random import random
from matplotlib.mlab import frange
from matplotlib.pyplot import plot, axis, show,axes
LEN = 0.1
def rotate(point, theta):
x = point[0]
y = point[1]
x_ = x * cos(theta) + y * sin(theta)
y_ = - x * sin(theta) + y * cos(theta)
return x_, y_
class CUBE(object):
def __init__(self, x, y, theta):
self.corner = [(LEN / 2, LEN / 2),
(-LEN / 2, LEN / 2),
(-LEN / 2, -LEN / 2),
(LEN / 2, -LEN / 2)
]
self.theta = theta
self.x = x
self.y = y
for i in range(4):
self.corner[i] = rotate(self.corner[i], theta)
self.corner[i] = (self.corner[i][0] + x,
self.corner[i][1] + y)
def is_include(cube, point):
point = [point[0] - cube.x, point[1] - cube.y]
point = rotate(point, -cube.theta)
if (point[0] < -LEN / 2
or point[0] > LEN / 2
or point[1] < -LEN / 2
or point[1] > LEN / 2
):
return False
else:
return True
def is_intersect(cube1, cube2):
if (any([is_include(cube1, point) for point in cube2.corner])
or any([is_include(cube2, point) for point in cube1.corner])
or is_include(cube1, (cube2.x, cube2.y))):
return True
else:
return False
def plot_cube(cube,n):
plot(
[cube.corner[i][0] for i in [0, 1, 2, 3, 0]],
[cube.corner[i][1] for i in [0, 1, 2, 3, 0]])
ax = axes()
ax.text(cube.x,cube.y,str(n))
def display(cubelist): # connects the 4 corners on a plot
for i,cube in enumerate(cubelist):
plot_cube(cube,i)
axis([0, 1, 0, 1]) # 1x1 grid
show()
cubelist = []
for i in range(100):
x0 = random()
y0 = random()
theta = random() * pi
cube = CUBE(x0, y0, theta)
if any(is_intersect(cube,cb) for cb in cubelist):
continue
else:
cubelist.append(cube)
display(cubelist)

Finding intersection between straight line and contour

I am trying to find the intersection point of a straight(dashed red) with the contour-line highlighted in red(see plot). I used .get_paths in the second plot to isolate said contour line form the others(second plot).
I have looked at a contour intersection problem, How to find all the intersection points between two contour-set in an efficient way, and have tried to use it as a base but have not been able to reproduce anything useful.
http://postimg.org/image/hz01fouvn/
http://postimg.org/image/m6utofwb7/
Does any one have any ideas?
relevant functions to recreate plot,
#for contour
def p_0(num,t) :
esc_p = np.sum((((-1)**n)*(np.exp(t)**n)*((math.factorial(n)*((n+1)**0.5))**-1)) for n in range(1,num,1))
return esc_p+1
tau = np.arange(-2,3,0.1)
r=[]
p1 = p_0(51,tau)
p2 = p_0(51,tau)
for i in p1:
temp_r=i/p2
r.append(temp_r)
x,y= np.meshgrid(tau,tau)
cs = plt.contour(x, y, np.log(r),50,colors='k')
whichContour =20
pa = CS.collections[whichContour].get_paths()[0]
v = pa.vertices
xx = v[:, 0]
yy = v[:, 1]
plt.plot(xx, yy, 'r-', label='Crossing contour')
#straight line
p=0.75
logp = (np.log(p*np.exp(tau)))
plt.plot(tau,logp)
Current attempt,
import matplotlib
import numpy as np
import matplotlib.cm as cm
import matplotlib.mlab as mlab
import matplotlib.pyplot as plt
import math
def intercepting_line() :
matplotlib.rcParams['xtick.direction'] = 'out'
matplotlib.rcParams['ytick.direction'] = 'out'
#fake data
delta = 0.025
x = np.arange(-3.0, 3.0, delta)
y = np.arange(-2.0, 2.0, delta)
X, Y = np.meshgrid(x, y)
Z1 = mlab.bivariate_normal(X, Y, 1.0, 1.0, 0.0, 0.0)
Z2 = mlab.bivariate_normal(X, Y, 1.5, 0.5, 1, 1)
Z = 10.0 * (Z2 - Z1)
#plot
cs = plt.contour(X,Y,Z)
whichContour = 2 # change this to find the right contour lines
#get the vertices to calculate an intercept with a line
p = cs.collections[whichContour].get_paths()[0]
#see: http://matplotlib.org/api/path_api.html#module-matplotlib.path
v = p.vertices
xx = v[:, 0]
yy = v[:, 1]
#this shows the innermost ring now
plt.plot(xx, yy, 'r--', label='inner ring')
#fake line
x = np.arange(-2, 3.0, 0.1)
y=lambda x,m:(m*x)
y=y(x,0.9)
lineMesh = np.meshgrid(x,y)
plt.plot(x,y,'r' ,label='line')
#get the intercepts, two in this case
x, y = find_intersections(v, lineMesh[1])
print x
print y
#plot the intercepting points
plt.plot(x[0], y[0], 'bo', label='first intercept')
#plt.plot(x[1], y[1], 'rs', label='second intercept')
plt.legend(shadow=True, fancybox=True, numpoints=1, loc='best')
plt.show()
#now we need to calculate the intercept of the vertices and whatever line
#this is pseudo code but works in case of two intercepting contour vertices
def find_intersections(A, B):
# min, max and all for arrays
amin = lambda x1, x2: np.where(x1<x2, x1, x2)
amax = lambda x1, x2: np.where(x1>x2, x1, x2)
aall = lambda abools: np.dstack(abools).all(axis=2)
slope = lambda line: (lambda d: d[:,1]/d[:,0])(np.diff(line, axis=0))
x11, x21 = np.meshgrid(A[:-1, 0], B[:-1, 0])
x12, x22 = np.meshgrid(A[1:, 0], B[1:, 0])
y11, y21 = np.meshgrid(A[:-1, 1], B[:-1, 1])
y12, y22 = np.meshgrid(A[1:, 1], B[1:, 1])
m1, m2 = np.meshgrid(slope(A), slope(B))
m1inv, m2inv = 1/m1, 1/m2
yi = (m1*(x21-x11-m2inv*y21) + y11)/(1 - m1*m2inv)
xi = (yi - y21)*m2inv + x21
xconds = (amin(x11, x12) < xi, xi <= amax(x11, x12),
amin(x21, x22) < xi, xi <= amax(x21, x22) )
yconds = (amin(y11, y12) < yi, yi <= amax(y11, y12),
amin(y21, y22) < yi, yi <= amax(y21, y22) )
return xi[aall(xconds)], yi[aall(yconds)]
At the moment it finds intersecting points but only where the line is uniform, the main reason why I cannot find a solution here is that I dont understand the original authors train of thinking here,
yi = (m1*(x21-x11-m2inv*y21) + y11)/(1 - m1*m2inv)
xi = (yi - y21)*m2inv + x21

Use shapely can find the intersection point, than use the point as the init guess value for fsolve() to find the real solution:
#for contour
def p_0(num,t) :
esc_p = np.sum((((-1)**n)*(np.exp(t)**n)*((math.factorial(n)*((n+1)**0.5))**-1)) for n in range(1,num,1))
return esc_p+1
tau = np.arange(-2,3,0.1)
x,y= np.meshgrid(tau,tau)
cs = plt.contour(x, y, np.log(p_0(51, y)/p_0(51, x)),[0.2],colors='k')
p=0.75
logp = (np.log(p*np.exp(tau)))
plt.plot(tau,logp)
from shapely.geometry import LineString
v1 = cs.collections[0].get_paths()[0].vertices
ls1 = LineString(v1)
ls2 = LineString(np.c_[tau, logp])
points = ls1.intersection(ls2)
x, y = points.x, points.y
from scipy import optimize
def f(p):
x, y = p
e1 = np.log(0.75*np.exp(x)) - y
e2 = np.log(p_0(51, y)/p_0(51, x)) - 0.2
return e1, e2
x2, y2 = optimize.fsolve(f, (x, y))
plt.plot(x, y, "ro")
plt.plot(x2, y2, "gx")
print x, y
print x2, y2
Here is the output:
0.273616328952 -0.0140657435002
0.275317387697 -0.0123646847549
and the plot:

See your contour lines as polylines and plug the vertex coordinates into the implicit line equation (F(P) = a.X + b.Y + c = 0). Every change of sign is an intersection, computed by solving 2x2 linear equations. You need no sophisticated solver.
If you need to detect the contour lines simultaneously, it is not much more complicated: consider the section of the terrain by a vertical plane through the line. You will obtain altitudes by linear interpolation along the edges of the grid tiles that are crossed. Finding the intersections with the grid is closely related to the Bresenham line drawing algorithm.
Then what you get is a profile, i.e. a function of a single variable. Locating the intersections with the horizontal planes (iso-values) is also done by detecting changes of sign.

This is a way that I used to solve this problem
def straight_intersection(straight1, straight2):
p1x = straight1[0][0]
p1y = straight1[0][1]
p2x = straight1[1][0]
p2y = straight1[1][1]
p3x = straight2[0][0]
p3y = straight2[0][1]
p4x = straight2[1][0]
p4y = straight2[1][1]
x = p1y * p2x * p3x - p1y * p2x * p4x - p1x * p2y * p4x + p1x * p2y * p3x - p2x * p3x * p4y + p2x * p3y * p4x + p1x * p3x * p4y - p1x * p3y * p4x
x = x / (p2x * p3y - p2x * p4y - p1x * p3y + p1x * p4y + p4x * p2y - p4x * p1y - p3x * p2y + p3x * p1y)
y = ((p2y - p1y) * x + p1y * p2x - p1x * p2y) / (p2x - p1x)
return (x, y)

While this question is old now, i want to share my answer for this problem for future generations to come.
Actually, there is two more solutions i found to this problem that is relatively efficient.
First solution is using pointPolygonTest in opencv recursively like that.
# Enumerate the line segment points between 2 points
for pt in zip(*line(*p1, *p2)):
if cv2.pointPolygonTest(conts[0], pt, False) == 0: # If the point is on the contour
return pt
Second solution, you can simply draw the contour and line and make a np.logical_and() to get the answer
blank = np.zeros((1000, 1000))
blank_contour = drawContour(blank.copy(), cnt[0], 0, 1, 1)
blank_line = cv2.line(blank.copy(), line[0], line[1], 1, 1)
intersections = np.logical_and(blank_contour, black_line)
points = np.where(intersections == 1)