def main():
parser = argparse.ArgumentParser(description='Initial rectangle localization.')
parser.add_argument('input', type=argparse.FileType('r'))
parser.add_argument('img_index', type=int)
parser.add_argument('px', type=int)
parser.add_argument('py', type=int)
args = parser.parse_args()
## Object that provides the images and contains information such
## as the camera parameters
ds = rectrckr.DataSource(args.input)
imgana = ds.get_image(args.img_index)
img = imgana.img
px, py = args.px, args.py
## Get data to plot the image levels and derivatives over the
## vertical and horizontal lines crossing (px,py)
lh = img[py,:]
lv = img[:,px]
lhd = array([
norm(array(lowlevel.gradient(img, x, py)))**2
for x in mgrid[2:img.shape[1]-2]])
lvd = array([
norm(array(lowlevel.gradient(img, px, y)))**2
for y in mgrid[2:img.shape[0]-2]])
## Extract edgels
z = imgana.extract_edgels(px,py)
edges = array([
[z[0],py],
[z[1],py],
[px,z[2]],
[px,z[3]],
])
##
## Plot image and the extracted edges
ion()
figure(1)
imshow(img, cmap=cm.gray)
plot(px,py, 'bo')
plot(edges[:,0], edges[:,1], 'ro')
## Plot image levels and derivatives
figure(2)
subplot(2,1,1)
plot(lh, 'b-')
twinx()
plot(mgrid[2:img.shape[1]-2], lhd, 'r-+')
subplot(2,1,2)
plot(lv)
twinx()
plot(mgrid[2:img.shape[0]-2], lvd, 'r-+')
code.interact(local=locals())
def estimate_direction_at_point(self, px, py):
return numpy.array(lowlevel.gradient(self.img, px, py))
