Get raw depth data through python

[purpleyoda1]

Hi! Is it possible to access raw depth data through the python API? I followed the example code for getting depth maps, but the resulting maps seem to have been run through several filters when comparing it to the raw stream in realsense viewer. I tried accessing the depth sensor or the pipelines processing block, but its difficult to figure out what methods and options are available for the different objects. Is there a way to disable the filters in python?

I apologize if this is actually a really simple case or i have missed something obvious (im quite the novice), but i have been trying on my own for a while now. Any help is greatly appreciated!

My current function can be seen below, although it doesn't run of course

def capture_and_save_raw_depth_map(output_path):

Configure realsense pipeline

pipeline = rs.pipeline()
config = rs.config()
config.enable_stream(rs.stream.depth, 424, 240, rs.format.z16, 30)

# Start stream
profile = pipeline.start(config)

# Get depth sensor
depth_sensor = profile.get_device().first_depth_sensor()
depth_sensor.set_option(rs.filter_option, 0)

# Disable post processing filters
processing_blocks = profile.get_device().query_processing_blocks()
for block in processing_blocks:
    try:
        block.set_option(rs.option.filter_option, 0)
    except:
        pass

try:
    # Skip first 30 frames to stabilize
    for i in range(30):
        pipeline.wait_for_frames()

    # Get depth frame
    frames = pipeline.wait_for_frames()
    depth_frame = frames.get_depth_frame()
    if not depth_frame:
        raise Exception("Could not get depth frame")

    # Convert to numpy array and save
    depth_map = np.asanyarray(depth_frame.get_data())
    cv2.imwrite(output_path, depth_map)
    print(f"Depth image saved to {output_path}")

finally:
    pipeline.stop()

[MartyG-RealSense]

Hi @purpleyoda1 A key reason why the realsense-viewer depth images can look different to Python ones is that realsense-viewer applies a range of post-processing filters and depth colorization settings by default. A Python script applies no filtering by default - filters have to be deliberately programmed into the script.

So you do no need to attempt to disable post-processing filters in Python because they are already inactive by default.

[purpleyoda1]

Oh wow i see, thanks! I assumed they were on based on visual inspection of the outputs. Thank you so much for the quick response!