Document full ZED depth availability and synchronization stages
Browse files
README.md
CHANGED
|
@@ -7,7 +7,7 @@ viewer: false
|
|
| 7 |
|
| 8 |
This dataset release contains raw recordings for **GRADE: Single-Frame Generative Radar Depth Estimation Under Visual Degradation**, accepted to **ACM MobiCom 2026**. See the [paper](https://arxiv.org/abs/2609.10756) and [project website](https://phi-lab-rice.github.io/GRADE/).
|
| 9 |
|
| 10 |
-
The recordings were made with a TI IWR1843BOOST radar (with DCA1000EVM), a ZED 2i stereo camera, and a DJI Action 5 Pro camera. Evaluation sequences also used a MAX30105 smoke sensor. **All 70 public sequences
|
| 11 |
|
| 12 |
## Folders and splits
|
| 13 |
|
|
@@ -35,17 +35,18 @@ Each sequence folder on the Hub also has:
|
|
| 35 |
```text
|
| 36 |
zed_video_anonymized.mkv calibrated rectified-left ZED RGB, defaced, 1280x720
|
| 37 |
dji_video_anonymized.mkv fisheye-corrected and cropped DJI RGB, defaced, 1280x720
|
|
|
|
| 38 |
```
|
| 39 |
|
| 40 |
Both videos use **lossy H.265/HEVC Main profile at CRF 14**, with 8-bit YUV 4:2:0 pixels in Matroska. Faces were blurred before this encoding. The original anonymized lossless RGB files were used as the transcoding inputs. Three original DJI recordings (`brk-2`, `brk-3`, and `brk-basement`) have 8, 3, and 6 untimed terminal video frames, respectively. Their anonymized videos omit those frames during CRF encoding, so every released DJI frame has a matching entry in its original DJI timestamp H5. No timestamps were removed from the H5 files.
|
| 41 |
|
| 42 |
-
|
| 43 |
|
| 44 |
The local preparation folders retain the original DJI recordings and ZED SVO2 files for preservation. **Do not include those originals in a public upload.**
|
| 45 |
|
| 46 |
## Process a sequence
|
| 47 |
|
| 48 |
-
`processing_code/processor.py` first matches the H5 sensor timestamps, then extracts synchronized radar, ZED RGB/depth, and DJI RGB frames. It saves one output folder per sequence. `processing_code/utils/` contains the sensor extractors, radar processing, synchronization, and point-cloud code. The processor requires both anonymized videos and the matching `zed_depth.h5`
|
| 49 |
|
| 50 |
1. From this repository's root folder, create a Python environment and install the dependencies:
|
| 51 |
|
|
@@ -57,7 +58,7 @@ The local preparation folders retain the original DJI recordings and ZED SVO2 fi
|
|
| 57 |
|
| 58 |
The public processor does not require the ZED SDK or an NVIDIA GPU. It uses OpenCV to decode the HEVC videos and h5py to read precomputed metric depth.
|
| 59 |
|
| 60 |
-
2. Process one evaluation sequence
|
| 61 |
|
| 62 |
```bash
|
| 63 |
python processing_code/processor.py \
|
|
|
|
| 7 |
|
| 8 |
This dataset release contains raw recordings for **GRADE: Single-Frame Generative Radar Depth Estimation Under Visual Degradation**, accepted to **ACM MobiCom 2026**. See the [paper](https://arxiv.org/abs/2609.10756) and [project website](https://phi-lab-rice.github.io/GRADE/).
|
| 9 |
|
| 10 |
+
The recordings were made with a TI IWR1843BOOST radar (with DCA1000EVM), a ZED 2i stereo camera, and a DJI Action 5 Pro camera. Evaluation sequences also used a MAX30105 smoke sensor. **All 70 public sequences include anonymized ZED and DJI RGB videos and the corresponding full-frame ZED depth.** The original ZED SVO2 and DJI recordings are not published.
|
| 11 |
|
| 12 |
## Folders and splits
|
| 13 |
|
|
|
|
| 35 |
```text
|
| 36 |
zed_video_anonymized.mkv calibrated rectified-left ZED RGB, defaced, 1280x720
|
| 37 |
dji_video_anonymized.mkv fisheye-corrected and cropped DJI RGB, defaced, 1280x720
|
| 38 |
+
zed_depth.h5 full ZED-frame depth before cross-sensor synchronization
|
| 39 |
```
|
| 40 |
|
| 41 |
Both videos use **lossy H.265/HEVC Main profile at CRF 14**, with 8-bit YUV 4:2:0 pixels in Matroska. Faces were blurred before this encoding. The original anonymized lossless RGB files were used as the transcoding inputs. Three original DJI recordings (`brk-2`, `brk-3`, and `brk-basement`) have 8, 3, and 6 untimed terminal video frames, respectively. Their anonymized videos omit those frames during CRF encoding, so every released DJI frame has a matching entry in its original DJI timestamp H5. No timestamps were removed from the H5 files.
|
| 42 |
|
| 43 |
+
Each sequence includes `zed_depth.h5`: full-frame ZED depth spatially aligned with the ZED RGB video at 1280×720, **before synchronization with radar and DJI**. The file has a `depth_mm` dataset of shape `(frames, 720, 1280)`, dtype `uint16`, with no HDF5 compression, matching the radar H5 compression setting. Zero marks invalid depth; positive values are millimeters rounded to 1 mm. Depth was computed from the original SVO2 before RGB face blurring. The ZED camera clock is `camera_timestamps_*.h5`; there is no separate SVO2 timestamp file. Frame index `i` in that H5 corresponds to frame `i` in the ZED video and depth H5, up to the exported frame count. The processor trims surplus terminal timestamps to that count. It then selects the ZED frame indices matched to radar and DJI timestamps; the output `zed_depth.npy` contains only these synchronized frames.
|
| 44 |
|
| 45 |
The local preparation folders retain the original DJI recordings and ZED SVO2 files for preservation. **Do not include those originals in a public upload.**
|
| 46 |
|
| 47 |
## Process a sequence
|
| 48 |
|
| 49 |
+
`processing_code/processor.py` first matches the H5 sensor timestamps, then extracts synchronized radar, ZED RGB/depth, and DJI RGB frames. It saves one output folder per sequence. `processing_code/utils/` contains the sensor extractors, radar processing, synchronization, and point-cloud code. The processor requires both anonymized videos and the matching `zed_depth.h5`, all included in this release. Download those files together with the sequence timestamp and radar H5 files before processing.
|
| 50 |
|
| 51 |
1. From this repository's root folder, create a Python environment and install the dependencies:
|
| 52 |
|
|
|
|
| 58 |
|
| 59 |
The public processor does not require the ZED SDK or an NVIDIA GPU. It uses OpenCV to decode the HEVC videos and h5py to read precomputed metric depth.
|
| 60 |
|
| 61 |
+
2. Process one downloaded evaluation sequence:
|
| 62 |
|
| 63 |
```bash
|
| 64 |
python processing_code/processor.py \
|