Items with no label
공지
FPGA community forums and blogs on community.intel.com are migrating to the new Altera Community and are read-only. For urgent support needs during this transition, please visit the FPGA Design Resources page or contact an Altera Authorized Distributor.
3338 토론

Change the Depth map accuracy with SDK2.0

cxinl
초급자
5,313 조회수

Hello.

I have a question that can I Change the Depth map accuracy like this:

rs2::pipeline pipe;

rs2::pipeline_profile selection = pipe.start();

rs2::device selected_device = selection.get_device();

auto depth_sensor = selected_device.first();

if (depth_sensor.supports(RS2_OPTION_EMITTER_ENABLED))

{

depth_sensor.set_option(RS2_OPTION_EMITTER_ENABLED, 1.f); // Enable emitter

depth_sensor.set_option(RS2_OPTION_EMITTER_ENABLED, 0.f); // Disable emitter

}

if (depth_sensor.supports(RS2_OPTION_LASER_POWER))

{

// Query min and max values:

auto range = depth_sensor.get_option_range(RS2_OPTION_LASER_POWER);

depth_sensor.set_option(RS2_OPTION_LASER_POWER, range.max); // Set max power

depth_sensor.set_option(RS2_OPTION_LASER_POWER, 0.f); // Disable laser

}

Does depth_sensor.set_option(RS2_OPTION_LASER_POWER, range.max) can change dehpth accruacy. The depth image is 16 bit ,and D400 camera can detect distance of 65m,

if I set range.max to 2.0m, dose the max value 65536 I get from depth img represent 2.0m? thanks.

0 포인트
10 응답
MartyG
명예로운 기여자 III
2,615 조회수

The depth scale on the 400 Series cameras, which influences accuracy, can be changed with RS2_OPTION_DEPTH_UNITS

0 포인트
cxinl
초급자
2,615 조회수

Thanks.

Default scale is 0.01, and 65535 represent 65m, is that mean I changed with RS2_OPTION_DEPTH_UNITS to 0.001, 65536 then represent 6.5m?

0 포인트
MartyG
명예로운 기여자 III
2,615 조회수

It is ... slightly complicated. The answer linked to below by Intel support agent Yermi is the best explanation that I know of.

0 포인트
cxinl
초급자
2,615 조회수

Thanks.

I use API

float rs2_depth_frame_get_distance(const rs2_frame* frame_ref, int x, int y, rs2_error** error);

In default, if the return value of rs2_depth_frame_get_distance represent the real distance in Meter, for example, return 1.5, is this reprensent 1.5m?

or I must multiply the scale factor of 0.01?

0 포인트
MartyG
명예로운 기여자 III
2,615 조회수

That question is outside of my knowledge, unfortunately. I will refer it to RealSense programming expert jb455 in the hope that they can provide useful insight.

0 포인트
cxinl
초급자
2,615 조회수
0 포인트
jb455
소중한 기여자 II
2,615 조회수

rs2_depth_frame_get_distance does indeed return values in metres. If you get values from the 16 bit depth image directly, you'll need to multiply by the depth unit to get the depth in metres.

Since the D400 cameras are stereo, they only need the projector to help when looking at a region with little colour texture like a plain wall or something, then the projector adds texture to the scene so the cameras can pick out points more easily. So changing the laser power will improve depth accuracy in some circumstances, but if you're outside where IR from the sun will overpower the projector, or too far away from a wall that the projector pattern is too dim by the time it reaches it (I don't know what the max range of the projector is), the projector won't do much to help. Changing the laser power value therefore won't necessarily directly affect the depth range.

As Marty pointed out, changing the depth unit will alter the maximum reported range as well as the reported resolution (though not necessarily accuracy - just because it's reported to more decimal places doesn't automatically make it a more reliable value but in my experience at short range, it generally looks pretty good down to 0.1mm resolution), and you can clip the depth data values on the top and bottom ends using depth clamp max/min (see viewer app>advanced controls>depth table) - but I don't think that will affect the values left over in the middle at all. To improve the reliability of the depth values, you could look at the post-processing options to alleviate random errors.

cxinl
초급자
2,615 조회수

Thanks to jb455.

I use 415 and 435 to reconstuct face 3D model, and I found it can't see the eyes area clearly, so I think maybe the depth resolution is not good

0 포인트
jb455
소중한 기여자 II
2,615 조회수

Are you creating the model from a single frame or using multiple frames? If the latter, perhaps the algorithm you're using has some sort of resolution parameter?

Using a single frame with depth unit set to 0.0001 (0.1mm) I can get much better resolution than you have there.

0 포인트
cxinl
초급자
2,615 조회수

Oh, the later method. I just use default configs ,so I don't know wether I need do some works to set the depth unit . My resolution both rgb and depth are 640x480

0 포인트
응답