- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Hello
I am porting my prototype application to use realsense SR300 instead of another commercial camera. I run it on linux and I use OpenCV libraries. I used the librealsense from github to acquire raw frames from the Intel camera and it works.
The problem that I have right now is the noise in the realsense camera. So far, I have been always using the non-adaptive background model in my application, and it was working fine for me. Unfortunately, after porting it to SR300 the level of noise is so big (around 75 units), that the background model doesn't work anymore.
What am I missing? Is it the issue of camera options configuration? Or should we use some other way of background extraction/filtering?
I use the following setup:
1. I use ubuntu 14.04 laptop with Intel i7 processor and Realsense SR300 via USB3.0
2. I use librealsense library https://github.com/IntelRealSense/librealsense
3. The realsense options are as follows:
color_backlight_compensation 0
color_brightness 0
color_contrast 50
color_exposure 156
color_gain 64
color_gamma 300
color_hue 0
color_saturation 64
color_sharpness 50
color_white_balance 4600
color_enable_auto_exposure 1
color_enable_auto_white_balance 1
f200_laser_power 15
f200_accuracy 3
f200_motion_range 50
f200_filter_option 7
f200_confidence_threshold 3
sr300_auto_range_enable_motion_versus_range 1
sr300_auto_range_enable_laser 1
sr300_auto_range_min_motion_versus_range 180
sr300_auto_range_max_motion_versus_range 605
sr300_auto_range_start_motion_versus_range 303
sr300_auto_range_min_laser 2
sr300_auto_range_max_laser 16
sr300_auto_range_start_laser -1
sr300_auto_range_upper_threshold 1250
sr300_auto_range_lower_threshold 650
frames_queue_size 20
hardware_logger_enabled 0
Link Copied

- Subscribe to RSS Feed
- Mark Topic as New
- Mark Topic as Read
- Float this Topic for Current User
- Bookmark
- Subscribe
- Printer Friendly Page