Datasets:
File size: 11,447 Bytes
4e6a175 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 | #!/usr/bin/env python3
"""Sample synchronized RGB/IR frame pairs.
Each RGB frame is paired with the IR frame nearest to ``rgb_time + ir_offset``.
Matches farther than ``--max-gap`` seconds are omitted.
Usage:
python scripts/sample_dataset.py \
--out-root ./Data --start 300 --step 300
"""
import argparse
import csv
import logging
import math
import os
import sys
import tempfile
from pathlib import Path
import cv2
FIELDS = [
"pair_id", "unit", "rgb_video", "ir_video", "ir_offset",
"sample_index", "rgb_frame", "rgb_time_s", "ir_frame", "ir_time_s",
"t_ir_target_s", "nearest_gap_s", "rgb_path", "ir_path", "ir_ok",
]
log = logging.getLogger("sample_dataset")
def open_video(path, kind):
cap = cv2.VideoCapture(str(path))
if not cap.isOpened():
raise RuntimeError(f"cannot open {kind} video: {path}")
return cap
def write_png(path, frame):
if not cv2.imwrite(str(path), frame):
raise RuntimeError(f"cannot write PNG: {path}")
def read_ir(cap, path, index, expected, previous_time=None):
ok, frame = cap.read()
if not ok:
if expected > 0 and index < expected:
raise RuntimeError(f"IR decode ended at frame {index}: {path}")
return None
if expected > 0 and index >= expected:
raise RuntimeError(f"IR decoded more than {expected} reported frames: {path}")
timestamp = cap.get(cv2.CAP_PROP_POS_MSEC) / 1000.0
if not math.isfinite(timestamp):
raise RuntimeError(f"non-finite IR timestamp at frame {index}: {path}")
if previous_time is not None and timestamp <= previous_time:
raise RuntimeError(f"unordered IR timestamp at frame {index}: {path}")
return index, frame, timestamp
def sample_pair(pair, output, unit_dirs, start, step, max_gap):
pair_id, unit = pair["pair_id"], pair["unit"]
rgb_path = unit_dirs[unit] / pair["rgb_video"]
ir_path = unit_dirs[unit] / pair["ir_video"]
rgb_dir, ir_dir = output / "rgb", output / "ir"
rgb_dir.mkdir(parents=True)
ir_dir.mkdir()
rows = []
rgb = open_video(rgb_path, "RGB")
try:
ir = open_video(ir_path, "IR")
try:
frame_count = int(rgb.get(cv2.CAP_PROP_FRAME_COUNT))
ir_count = int(ir.get(cv2.CAP_PROP_FRAME_COUNT))
next_ir_index = 0
current = read_ir(ir, ir_path, next_ir_index, ir_count)
if current is None:
raise RuntimeError(f"IR video decoded no frames: {ir_path}")
next_ir_index += 1
previous = None
previous_target = None
for index in range(start, frame_count, step):
if not rgb.set(cv2.CAP_PROP_POS_FRAMES, index):
raise RuntimeError(f"cannot seek {rgb_path} to frame {index}")
ok, rgb_frame = rgb.read()
if not ok:
raise RuntimeError(f"cannot read {rgb_path} at frame {index}")
rgb_time = rgb.get(cv2.CAP_PROP_POS_MSEC) / 1000.0
if not math.isfinite(rgb_time):
raise RuntimeError(f"non-finite RGB timestamp: {rgb_path}, frame {index}")
target = rgb_time + pair["ir_offset"]
if previous_target is not None and target < previous_target:
raise RuntimeError(f"RGB timestamps moved backwards: {rgb_path}")
previous_target = target
while current is not None and current[2] <= target:
previous = current
current = read_ir(
ir, ir_path, next_ir_index, ir_count, previous[2]
)
next_ir_index += 1
if previous is None:
continue
if current is None:
if target > previous[2]:
break
nearest = previous
else:
nearest = (
current
if current[2] - target < target - previous[2]
else previous
)
ir_index, ir_frame, ir_time = nearest
gap = abs(target - ir_time)
if gap > max_gap:
continue
name = f"f{index:05d}.png"
write_png(rgb_dir / name, rgb_frame)
write_png(ir_dir / name, ir_frame)
rows.append({
"pair_id": pair_id,
"unit": unit,
"rgb_video": pair["rgb_video"],
"ir_video": pair["ir_video"],
"ir_offset": pair["ir_offset"],
"sample_index": index,
"rgb_frame": index,
"rgb_time_s": round(rgb_time, 4),
"ir_frame": ir_index,
"ir_time_s": round(ir_time, 4),
"t_ir_target_s": round(target, 4),
"nearest_gap_s": round(gap, 4),
"rgb_path": (Path(unit) / pair_id / "rgb" / name).as_posix(),
"ir_path": (Path(unit) / pair_id / "ir" / name).as_posix(),
"ir_ok": True,
})
finally:
ir.release()
finally:
rgb.release()
log.info("%s: sampled %d frame pairs", pair_id, len(rows))
return rows
def load_pairs(sync_csv, unit_dirs):
with sync_csv.open(newline="") as file:
reader = csv.DictReader(file)
required = {"pair_id", "unit", "rgb_video", "ir_video", "ir_offset"}
if not reader.fieldnames or not required.issubset(reader.fieldnames):
raise RuntimeError(f"invalid sync.csv header: {sync_csv}")
source = list(reader)
pairs, seen = [], set()
for row in source:
pair = {key: (row[key] or "").strip() for key in required - {"ir_offset"}}
pair_id = pair["pair_id"]
if not pair_id or Path(pair_id).name != pair_id:
raise RuntimeError(f"unsafe pair_id: {pair_id!r}")
if pair_id in seen:
raise RuntimeError(f"duplicate pair_id: {pair_id}")
if pair["unit"] not in unit_dirs:
raise RuntimeError(f"unknown unit for {pair_id}: {pair['unit']!r}")
offset = row["ir_offset"].strip()
if offset.upper() == "N/A":
log.info("%s: skipped because IR offset is N/A", pair_id)
continue
try:
pair["ir_offset"] = float(offset)
except ValueError as error:
raise RuntimeError(f"invalid IR offset for {pair_id}") from error
if not math.isfinite(pair["ir_offset"]):
raise RuntimeError(f"non-finite IR offset for {pair_id}")
for key in ("rgb_video", "ir_video"):
path = unit_dirs[pair["unit"]] / pair[key]
if not path.is_file():
raise RuntimeError(f"missing input for {pair_id}: {path}")
seen.add(pair_id)
pairs.append(pair)
return pairs
def write_manifest(path, rows):
with path.open("w", newline="") as file:
writer = csv.DictWriter(file, fieldnames=FIELDS)
writer.writeheader()
writer.writerows(rows)
file.flush()
os.fsync(file.fileno())
def commit(stage, pairs, out_root):
"""Install staged RGB/IR directories, rolling them back if commit fails."""
installed = []
out_root.mkdir(parents=True, exist_ok=True)
try:
for pair in pairs:
relative = Path(pair["unit"]) / pair["pair_id"]
for kind in ("rgb", "ir"):
source = stage / "new" / relative / kind
target = out_root / relative / kind
backup = stage / "old" / relative / kind
target.parent.mkdir(parents=True, exist_ok=True)
backup.parent.mkdir(parents=True, exist_ok=True)
had_old = target.exists()
if had_old:
if target.is_symlink() or not target.is_dir():
raise RuntimeError(f"unexpected output path: {target}")
target.rename(backup)
try:
source.rename(target)
except Exception:
if had_old:
backup.rename(target)
raise
installed.append((source, target, backup, had_old))
os.replace(stage / "samples.csv", out_root / "samples.csv")
except Exception:
for source, target, backup, had_old in reversed(installed):
source.parent.mkdir(parents=True, exist_ok=True)
target.rename(source)
if had_old:
backup.rename(target)
raise
def parse_args():
parser = argparse.ArgumentParser(description=__doc__)
parser.add_argument(
"--raw-root", type=Path, default=Path("RawVideos"),
help="root directory containing the raw unit imagery",
)
parser.add_argument(
"--sync-csv", type=Path, default=Path("sync.csv"),
help="CSV describing synchronized RGB/IR video pairs",
)
parser.add_argument(
"--out-root", type=Path, default=Path("Data"),
help="destination directory for sampled frames and samples.csv",
)
parser.add_argument(
"--start", type=int, default=300,
help="zero-based index of the first RGB frame to sample (default: 300)",
)
parser.add_argument(
"--step", type=int, default=300,
help="number of RGB frames between samples (default: 300)",
)
parser.add_argument(
"--max-gap", type=float, default=10.0,
help="maximum allowed RGB/IR timestamp gap in seconds (default: 10.0)",
)
args = parser.parse_args()
if args.start < 0:
parser.error("--start must be non-negative")
if args.step <= 0:
parser.error("--step must be positive")
if not math.isfinite(args.max_gap) or args.max_gap < 0:
parser.error("--max-gap must be a finite, non-negative number")
return args
def main():
args = parse_args()
unit_dirs = {
"Canopy": args.raw_root / "Livingston Place 18 19 20 March 2026"
/ "Canopy Unit 18 March 2026" / "DCIM" / "100MEDIA",
"Graminoid": args.raw_root / "Livingston Place 18 19 20 March 2026"
/ "Graminoid Unit 20 March 2026" / "DCIM" / "100MEDIA",
}
pairs = load_pairs(args.sync_csv, unit_dirs)
if not pairs:
raise RuntimeError("sync.csv contains no pairs")
args.out_root.parent.mkdir(parents=True, exist_ok=True)
with tempfile.TemporaryDirectory(
prefix=".sample-dataset-", dir=args.out_root.parent,
ignore_cleanup_errors=True,
) as temporary:
stage = Path(temporary)
rows = []
for pair in pairs:
output = stage / "new" / pair["unit"] / pair["pair_id"]
rows.extend(
sample_pair(
pair, output, unit_dirs, args.start, args.step, args.max_gap
)
)
write_manifest(stage / "samples.csv", rows)
commit(stage, pairs, args.out_root)
log.info("complete: %d pairs, %d samples", len(pairs), len(rows))
if __name__ == "__main__":
logging.basicConfig(level=logging.INFO, format="%(levelname)s: %(message)s")
try:
main()
except Exception as error:
log.error("%s", error)
sys.exit(1)
|