Pranay Thangeda commited on
Commit ·
74f8f8a
1
Parent(s): 8fc0f7c
update usage instructions in readme
Browse files
README.md
CHANGED
|
@@ -170,6 +170,46 @@ stiffness = sample['stiffness']
|
|
| 170 |
scooped_volume = sample['scooped_volume']
|
| 171 |
```
|
| 172 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 173 |
## Dataset Creation
|
| 174 |
|
| 175 |
The dataset was created using the following steps:
|
|
|
|
| 170 |
scooped_volume = sample['scooped_volume']
|
| 171 |
```
|
| 172 |
|
| 173 |
+
### Loading All Data for a Specific Terrain ID
|
| 174 |
+
|
| 175 |
+
If you want to load all samples corresponding to a specific terrain ID, you can filter the dataset using the `filter` method. Here's how you can do it:
|
| 176 |
+
|
| 177 |
+
```python
|
| 178 |
+
# Specify the terrain ID you're interested in
|
| 179 |
+
terrain_id_of_interest = 10 # Replace with the desired terrain ID (1-67)
|
| 180 |
+
|
| 181 |
+
# Filter the dataset to include only samples from the specified terrain
|
| 182 |
+
terrain_samples = dataset.filter(lambda sample: sample['terrain_id'] == terrain_id_of_interest)
|
| 183 |
+
|
| 184 |
+
print(f"Number of samples for terrain {terrain_id_of_interest}: {len(terrain_samples)}")
|
| 185 |
+
for sample in terrain_samples:
|
| 186 |
+
# Load RGB image
|
| 187 |
+
rgb_image = sample['rgb_image'] # PIL Image
|
| 188 |
+
|
| 189 |
+
# Load and reconstruct depth image
|
| 190 |
+
depth_image = sample['depth_image'] # PIL Image
|
| 191 |
+
depth_array = np.array(depth_image).astype(np.float32)
|
| 192 |
+
depth_normalized = depth_array / 65535
|
| 193 |
+
depth_min = sample['depth_min']
|
| 194 |
+
depth_max = sample['depth_max']
|
| 195 |
+
original_depth = depth_normalized * (depth_max - depth_min) + depth_min
|
| 196 |
+
|
| 197 |
+
# Load F/T sensor data
|
| 198 |
+
ft_data = pd.read_csv(sample['ft_csv_path'], header=None).values
|
| 199 |
+
|
| 200 |
+
# Access action parameters
|
| 201 |
+
pixel_x = sample['pixel_x']
|
| 202 |
+
pixel_y = sample['pixel_y']
|
| 203 |
+
yaw = sample['yaw']
|
| 204 |
+
scoop_depth = sample['scoop_depth']
|
| 205 |
+
stiffness = sample['stiffness']
|
| 206 |
+
scooped_volume = sample['scooped_volume']
|
| 207 |
+
|
| 208 |
+
# Perform your analysis or processing here
|
| 209 |
+
# Example: Print scooped volume for each sample
|
| 210 |
+
print(f"Sample {sample['sample_index']}: Scooped volume = {scooped_volume} cubic meters")
|
| 211 |
+
```
|
| 212 |
+
|
| 213 |
## Dataset Creation
|
| 214 |
|
| 215 |
The dataset was created using the following steps:
|