def show_simulating_events(events, width=128, height=128):
"""
plot events in three-dimensional space.
Inputs:
-------
events - the dynamic vision sensor.
width - the width of AER sensor.
height - the height of AER sensor.
Outputs:
------
figure - a figure shows events in three-dimensional space.
"""
events_ON, events_OFF = event_processing.On_off_events(events)
fig = plt.figure('{} * {}'.format(width, height))
ax = fig.gca(projection='3d')
ax.scatter(events_ON[1][:] / 10000,
events_ON[2][:],
events_ON[3][:],
c='r',
label='ON',
s=6) # events_ON[1][:]
ax.scatter(events_OFF[1][:] / 10000,
events_OFF[2][:],
events_OFF[3][:],
c='limegreen',
label='OFF',
s=6)
font1 = {'family': 'Times New Roman', 'size': 22}
font2 = {'family': 'Times New Roman', 'size': 17}
font3 = {'family': 'Times New Roman', 'size': 16}
ax.set_xlabel('\n t(ms)', font1) # us
ax.set_ylabel('\n x', font1)
ax.set_zlabel('\n y', font1)
ax.set_xlim([0, 1]) # int(length/1000)
ax.set_ylim([1, width])
ax.set_zlim([1, height])
# print(numpy.linspace(1, width, 5).astype(int))
ax.set_yticks(numpy.linspace(1, width, 5).astype(int))
ax.set_zticks(numpy.linspace(1, height, 5).astype(int))
ax.set_xticks(numpy.linspace(0, 1, 5)) # length
ax.set_xticklabels(ax.get_xticks(), font2)
ax.set_yticklabels(ax.get_yticks(), font2)
ax.zaxis.set_tick_params(labelsize=17) # 13
for t in ax.zaxis.get_major_ticks():
t.label.set_fontname("Times New Roman")
ax.legend(loc='upper center', bbox_to_anchor=(0.76, 0.83), prop=font3)
# ax.set_xticks(numpy.linspace(min(min(events_ON[1][:]), min(events_OFF[1][:])) / 10 ** 6, max(max(events_ON[1][:]), max(events_OFF[1][:])) / 10 ** 6, 5))
# ax.set_ylim([0, width])
# ax.set_zlim([0, height])
# ax.set_xlim([min(min(events_ON[1][:]), min(events_OFF[1][:])) / 10 ** 6, max(max(events_ON[1][:]), max(events_OFF[1][:])) / 10 ** 6])
#ax.legend(loc = 0)
plt.show()
def show_ON_OFF_spikes(events, width=240, height=180):
"""
plot events in three-dimensional space.
Inputs:
-------
events - the dynamic vision sensor.
width - the width of AER sensor.
height - the height of AER sensor.
length - the time length.
Outputs:
------
figure - a figure shows events in three-dimensional space.
"""
new_events = numpy.copy(events)
events[0, :] = new_events[3, :]
events[2, :] = width - new_events[2, :]
events[3, :] = new_events[0, :]
events_ON, events_OFF = event_processing.On_off_events(events)
fig = plt.figure('{} * {}'.format(width, height))
ax = fig.gca(projection='3d')
#ax.scatter(events_ON[3][:]/1000, events_ON[2][:], events_ON[1][:], c='r', label='ON', s= 6) # events_ON[1][:]
ax.scatter(events_OFF[3][:] / 1000,
events_OFF[2][:],
events_OFF[1][:],
c='limegreen',
label='OFF',
s=6)
font1 = {'family': 'Times New Roman', 'size': 20}
font1_x = {'family': 'Times New Roman', 'size': 19}
font2 = {'size': 13}
ax.set_xlabel('t(ms)', font1_x) #us
ax.set_ylabel('x', font1)
ax.set_zlabel('y', font1)
#ax.set_xlim([0, 200]) # int(length/1000)
ax.set_ylim([1, width])
ax.set_zlim([1, height])
# print(numpy.linspace(1, width, 5).astype(int))
ax.set_yticks(numpy.linspace(1, width, 5).astype(int))
ax.set_zticks(numpy.linspace(1, height, 5).astype(int))
#ax.set_xticks(numpy.linspace(0, 3, 5).astype(int)) # length
ax.set_xticks(numpy.linspace(0, 3, 5)) # length
ax.xaxis.set_tick_params(labelsize=14) #12
ax.yaxis.set_tick_params(labelsize=14) #13
ax.zaxis.set_tick_params(labelsize=14) #13
ax.legend(loc='upper center', bbox_to_anchor=(0.77, 0.83), prop=font2)
# ax.set_xticks(numpy.linspace(min(min(events_ON[1][:]), min(events_OFF[1][:])) / 10 ** 6, max(max(events_ON[1][:]), max(events_OFF[1][:])) / 10 ** 6, 5))
# ax.set_ylim([0, width])
# ax.set_zlim([0, height])
# ax.set_xlim([min(min(events_ON[1][:]), min(events_OFF[1][:])) / 10 ** 6, max(max(events_ON[1][:]), max(events_OFF[1][:])) / 10 ** 6])
#ax.legend(loc = 0)
plt.show()
def show_davis240_ON_OFF_spikes(events, width=240, height=180):
"""
plot events in three-dimensional space.
Inputs:
-------
events - the dynamic vision sensor, including polarity(t), timestamp(ts), x coordinate(X) and y coordinate(Y).
width - the width of AER sensor.
height - the height of AER sensor.
length - the time length.
Outputs:
------
figure - a figure shows events in three-dimensional space.
"""
events[2, :] = width - events[2, :]
events_ON, events_OFF = event_processing.On_off_events(events)
fig = plt.figure('{} * {}'.format(width, height))
ax = fig.gca(projection='3d')
#ax.scatter(events_ON[1][:]/1000, events_ON[2][:], events_ON[3][:], c='r', label='ON', s= 6) # events_ON[1][:]
#ax.scatter(events_OFF[1][:]/1000, events_OFF[2][:], events_OFF[3][:], c='limegreen', label='OFF', s= 6)
ax.scatter(events_ON[1][:] / 1000,
events_ON[2][:],
events_ON[3][:],
c='red',
label='ON',
s=3) # events_ON[1][:]
ax.scatter(events_OFF[1][:] / 1000,
events_OFF[2][:],
events_OFF[3][:],
c='mediumblue',
label='OFF',
s=3)
font1 = {'family': 'Times New Roman', 'size': 22}
font2 = {'family': 'Times New Roman', 'size': 17}
font3 = {'family': 'Times New Roman', 'size': 16}
ax.set_xlabel('\n t(ms)', font1) # us
ax.set_ylabel('\n x', font1)
ax.set_zlabel('\n y', font1)
#ax.set_xlim([0, 200]) # int(length/1000)
ax.set_ylim([1, width])
ax.set_zlim([1, height])
# print(numpy.linspace(1, width, 5).astype(int))
ax.set_yticks(numpy.linspace(1, width, 2).astype(int))
ax.set_zticks(numpy.linspace(1, height, 2).astype(int))
ax.set_xticks(numpy.linspace(0, 4, 5)) # bicylce 3, night_drive 6, sun 4.
ax.set_xticklabels(ax.get_xticks(), font2)
ax.set_yticklabels(ax.get_yticks(), font2)
ax.zaxis.set_tick_params(labelsize=17) # 13
for t in ax.zaxis.get_major_ticks():
t.label.set_fontname("Times New Roman")
#ax.legend(loc='upper center', bbox_to_anchor=(0.77, 0.83), prop=font2)
plt.show()
def show_noise_events(true_events, events, width=346, height=240):
"""
plot events in three-dimensional space.
Inputs:
-------
true events - the true event signal.
events - events include true event signal and noise events.
width - the width of AER sensor.
height - the height of AER sensor.
Outputs:
------
figure - a figure shows events in three-dimensional space.
"""
index = numpy.array(
numpy.all((events[:, None, 1:] == true_events[None, :, 1:]),
axis=-1).nonzero()).T.tolist()
noise_index = numpy.array(index)
true_events = numpy.transpose(true_events)
events = numpy.transpose(events)
noise_events = numpy.copy(events)
for i in range(events.shape[1]):
for j in range(true_events.shape[1]):
if numpy.array_equal(events[1:, i], true_events[1:, j]):
numpy.delete(noise_events, i, 1)
new_events = numpy.copy(true_events)
new_events[0, :] = true_events[3, :]
new_events[1, :] = true_events[0, :]
new_events[2, :] = true_events[1, :]
new_events[3, :] = true_events[2, :]
events_ON, events_OFF = event_processing.On_off_events(new_events)
fig = plt.figure('{} * {}'.format(width, height))
ax = fig.gca(projection='3d')
ax.scatter(events_ON[1][:],
events_ON[2][:],
events_ON[3][:],
c='r',
label='ON',
s=6) # events_ON[1][:]
ax.scatter(events_OFF[1][:],
events_OFF[2][:],
events_OFF[3][:],
c='limegreen',
label='OFF',
s=6)
font1 = {'family': 'Times New Roman', 'size': 20}
font1_x = {'family': 'Times New Roman', 'size': 19}
font2 = {'size': 13}
ax.set_xlabel('t(us)', font1_x) # us
ax.set_ylabel('x', font1)
ax.set_zlabel('y', font1)
ax.set_xlim([0, max(new_events[1][:])]) # int(length/1000)
ax.set_ylim([1, width])
ax.set_zlim([1, height])
# print(numpy.linspace(1, width, 5).astype(int))
ax.set_yticks(numpy.linspace(1, width, 5).astype(int))
ax.set_zticks(numpy.linspace(1, height, 5).astype(int))
ax.set_xticks(numpy.linspace(0, max(new_events[1][:]), 5)) # length
ax.xaxis.set_tick_params(labelsize=14) # 12
ax.yaxis.set_tick_params(labelsize=14) # 13
ax.zaxis.set_tick_params(labelsize=14) # 13
ax.legend(loc='upper center', bbox_to_anchor=(0.77, 0.83), prop=font2)
# ax.set_xticks(numpy.linspace(min(min(events_ON[1][:]), min(events_OFF[1][:])) / 10 ** 6, max(max(events_ON[1][:]), max(events_OFF[1][:])) / 10 ** 6, 5))
# ax.set_ylim([0, width])
# ax.set_zlim([0, height])
# ax.set_xlim([min(min(events_ON[1][:]), min(events_OFF[1][:])) / 10 ** 6, max(max(events_ON[1][:]), max(events_OFF[1][:])) / 10 ** 6])
# ax.legend(loc = 0)
plt.show()