Skip to content

Commit

Permalink
allow to sweep the focus when plotting the phantom source. Also add a…
Browse files Browse the repository at this point in the history 
… depth bar figure.
  • Loading branch information
@relleums
relleums committed Aug 26, 2024
1 parent 538581b commit 5cf0c9b
Show file tree
Hide file tree
Showing 2 changed files with 141 additions and 106 deletions.
245 changes: 140 additions & 105 deletions plenoptics/scripts/plot_phantom_source.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -85,122 +85,157 @@ def gzip_read_raw_sensor_response_into_time_lixel_repr(path):
# -----------------------------------------
systematic_reco_over_true = 1.0169723853658978

# find true depth
# ---------------
TRUE_DEPTH = config["observations"]["phantom"]["phantom_source_meshes_depth"]
object_distances = np.array([TRUE_DEPTH[key] for key in TRUE_DEPTH])
reco_object_distances = systematic_reco_over_true * object_distances

images_dir = os.path.join(out_dir, "images.cache")
os.makedirs(images_dir, exist_ok=True)

img_vmax = 0.0
for obj_idx in range(len(reco_object_distances)):
reco_object_distance = reco_object_distances[obj_idx]

image_path = os.path.join(images_dir, "{:06d}.float32".format(obj_idx))

if os.path.exists(image_path):
img = plenoptics.analysis.image.read_image(path=image_path)
else:
img = plenoptics.analysis.image.compute_image(
light_field_geometry=light_field_geometry,
light_field=phantom_source_light_field,
object_distance=reco_object_distance,
bins=image_bins,
prng=prng,
)
plenoptics.analysis.image.write_image(path=image_path, image=img)

img_vmax = np.max([img_vmax, np.max(img)])

CMAPS = plenoptics.plot.CMAPS
NPIX = 1280
FIG_FILENAME_FORMAT = "instrument_{:s}_cmap_{:s}_{:06d}"

for focus_mode in ["sweep", "on_mesh_structures"]:
if focus_mode == "on_mesh_structures":
# find true depth
# ---------------
TRUE_DEPTH = config["observations"]["phantom"][
"phantom_source_meshes_depth"
]
object_distances = np.array([TRUE_DEPTH[key] for key in TRUE_DEPTH])
else:
object_distances = np.geomspace(2.2e3, 2.2e4, 36)
reco_object_distances = systematic_reco_over_true * object_distances

FIG_FILENAME_FORMAT = "instrument_{:s}_cmap_{:s}_{:06d}.jpg"
focusmode_dir = os.path.join(out_dir, focus_mode)
os.makedirs(focusmode_dir, exist_ok=True)

for cmapkey in CMAPS:
cmap_dir = os.path.join(out_dir, cmapkey)
os.makedirs(cmap_dir, exist_ok=True)
images_dir = os.path.join(focusmode_dir, "images.cache")
os.makedirs(images_dir, exist_ok=True)

for obj_idx in range(len(object_distances)):
image_path = os.path.join(images_dir, "{:06d}.float32".format(obj_idx))
img = plenoptics.analysis.image.read_image(path=image_path)
img_vmax = 0.0
for obj_idx in range(len(reco_object_distances)):
reco_object_distance = reco_object_distances[obj_idx]

fig_filename = FIG_FILENAME_FORMAT.format(
instrument_key, cmapkey, obj_idx
)
fig_path = os.path.join(cmap_dir, fig_filename)
image_path = os.path.join(images_dir, "{:06d}.float32".format(obj_idx))

fig = sebplt.figure(
style={"rows": NPIX, "cols": NPIX, "fontsize": 1.0}
)
ax = sebplt.add_axes(
fig=fig,
span=[0.0, 0.0, 1, 1],
if os.path.exists(image_path):
img = plenoptics.analysis.image.read_image(path=image_path)
else:
img = plenoptics.analysis.image.compute_image(
light_field_geometry=light_field_geometry,
light_field=phantom_source_light_field,
object_distance=reco_object_distance,
bins=image_bins,
prng=prng,
)
plenoptics.analysis.image.write_image(path=image_path, image=img)

img_vmax = np.max([img_vmax, np.max(img)])

for cmapkey in CMAPS:
cmap_dir = os.path.join(focusmode_dir, cmapkey)
os.makedirs(cmap_dir, exist_ok=True)

for obj_idx in range(len(object_distances)):
image_path = os.path.join(
images_dir, "{:06d}.float32".format(obj_idx)
)
img = plenoptics.analysis.image.read_image(path=image_path)

fig_filename = FIG_FILENAME_FORMAT.format(
instrument_key, cmapkey, obj_idx
)
fig_path = os.path.join(cmap_dir, fig_filename + ".jpg")

fig = sebplt.figure(
style={"rows": NPIX, "cols": NPIX, "fontsize": 1.0}
)
ax = sebplt.add_axes(fig=fig, span=[0.0, 0.0, 1, 1])
ax.set_aspect("equal")
cmap = ax.pcolormesh(
np.rad2deg(image_edge_bin["edges"]),
np.rad2deg(image_edge_bin["edges"]),
np.transpose(img) / img_vmax,
cmap=cmapkey,
norm=sebplt.plt_colors.PowerNorm(
gamma=CMAPS[cmapkey]["gamma"],
vmin=0.0,
vmax=1.0,
),
)
sebplt.ax_add_circle(
ax=ax,
x=0.0,
y=0.0,
r=0.5 * max_FoV_diameter_deg,
linewidth=0.33,
linestyle="-",
color=CMAPS[cmapkey]["linecolor"],
num_steps=360 * 5,
)
ax.set_xlim(np.rad2deg(image_edge_bin["limits"]))
ax.set_ylim(np.rad2deg(image_edge_bin["limits"]))
sebplt.ax_add_grid_with_explicit_ticks(
ax=ax,
xticks=image_edge_ticks_deg,
yticks=image_edge_ticks_deg,
linewidth=0.33,
color=CMAPS[cmapkey]["linecolor"],
)
fig.savefig(fig_path)
sebplt.close(fig)

# focus bar plot
# --------------
fig = sebplt.figure(
style={"rows": NPIX, "cols": NPIX // 4, "fontsize": 2.0}
)
ax = sebplt.add_axes(
fig=fig,
span=[0.75, 0.1, 0.2, 0.8],
style={"spines": ["left"], "axes": ["y"], "grid": False},
)
ax.set_ylim([0, 2.5e1])
ax.set_xlim([0, 1])
ax.set_ylabel("depth / km")
ax.plot(
[0, 1],
[
object_distances[obj_idx] * 1e-3,
object_distances[obj_idx] * 1e-3,
],
color="white",
linewidth=2,
)
fig.savefig(os.path.join(cmap_dir, fig_filename + ".bar.jpg"))
sebplt.close(fig)

# colormap
# --------
fig_cmap = sebplt.figure(
style={"rows": 120, "cols": 1280, "fontsize": 1}
)
ax.set_aspect("equal")
cmap = ax.pcolormesh(
np.rad2deg(image_edge_bin["edges"]),
np.rad2deg(image_edge_bin["edges"]),
np.transpose(img) / img_vmax,
cmap=cmapkey,
norm=sebplt.plt_colors.PowerNorm(
gamma=CMAPS[cmapkey]["gamma"],
vmin=0.0,
vmax=1.0,
),
ax_cmap = sebplt.add_axes(fig_cmap, [0.1, 0.8, 0.8, 0.15])
ax_cmap.text(0.5, -4.7, r"intensity$\,/\,$1")
sebplt.plt.colorbar(
cmap, cax=ax_cmap, extend="max", orientation="horizontal"
)
sebplt.ax_add_circle(
ax=ax,
x=0.0,
y=0.0,
r=0.5 * max_FoV_diameter_deg,
linewidth=0.33,
linestyle="-",
color=CMAPS[cmapkey]["linecolor"],
num_steps=360 * 5,
)
ax.set_xlim(np.rad2deg(image_edge_bin["limits"]))
ax.set_ylim(np.rad2deg(image_edge_bin["limits"]))
sebplt.ax_add_grid_with_explicit_ticks(
ax=ax,
xticks=image_edge_ticks_deg,
yticks=image_edge_ticks_deg,
linewidth=0.33,
color=CMAPS[cmapkey]["linecolor"],
)
fig.savefig(fig_path)
sebplt.close(fig)

# colormap
# --------
fig_cmap = sebplt.figure(style={"rows": 120, "cols": 1280, "fontsize": 1})
ax_cmap = sebplt.add_axes(fig_cmap, [0.1, 0.8, 0.8, 0.15])
ax_cmap.text(0.5, -4.7, r"intensity$\,/\,$1")
sebplt.plt.colorbar(
cmap, cax=ax_cmap, extend="max", orientation="horizontal"
)
fig_cmap_filename = "cmap_{:s}.jpg".format(cmapkey)
fig_cmap.savefig(os.path.join(cmap_dir, fig_cmap_filename))
sebplt.close(fig_cmap)

# avg image
# ---------
avg_img = np.zeros(shape=(NPIX, NPIX, 3), dtype=np.float32)
for obj_idx in range(len(object_distances)):
fig_filename = FIG_FILENAME_FORMAT.format(
instrument_key, cmapkey, obj_idx
fig_cmap_filename = "cmap_{:s}.jpg".format(cmapkey)
fig_cmap.savefig(os.path.join(cmap_dir, fig_cmap_filename))
sebplt.close(fig_cmap)

# avg image
# ---------
avg_img = np.zeros(shape=(NPIX, NPIX, 3), dtype=np.float32)
for obj_idx in range(len(object_distances)):
fig_filename = (
FIG_FILENAME_FORMAT.format(instrument_key, cmapkey, obj_idx)
+ ".jpg"
)
fig_path = os.path.join(cmap_dir, fig_filename)
avg_img += skimage.io.imread(fig_path)
fig_filename = "instrument_{:s}_cmap_{:s}_average.jpg".format(
instrument_key, cmapkey
)
fig_path = os.path.join(cmap_dir, fig_filename)
avg_img += skimage.io.imread(fig_path)
fig_filename = "instrument_{:s}_cmap_{:s}_average.jpg".format(
instrument_key, cmapkey
)
fig_path = os.path.join(cmap_dir, fig_filename)
avg_img /= len(object_distances)
avg_img = avg_img.astype(np.uint8)
skimage.io.imsave(fig_path, avg_img)
avg_img /= len(object_distances)
avg_img = avg_img.astype(np.uint8)
skimage.io.imsave(fig_path, avg_img)


fig_filename = "phantom_source_meshes.jpg"
Expand Down
2 changes: 1 addition & 1 deletion plenoptics/version.py
View file
Original file line number Diff line number Diff line change
@@ -1 +1 @@
__version__ = "0.2.5"
__version__ = "0.2.6"

0 comments on commit 5cf0c9b

Please sign in to comment.