AFBR-S50 API Reference Manual v1.5.6
AFBR-S50 Time-of-Flight Sensor SDK for Embedded Software
Loading...
Searching...
No Matches
Pixel Binning Algorithm

Pixel Binning Algorithm (PBA) parameter definitions and API functions. More...

Collaboration diagram for Pixel Binning Algorithm:

Data Structures

struct  argus_cfg_pba_t
 The pixel binning algorithm settings data structure. More...
 

Enumerations

enum  argus_pba_flags_t {
  PBA_ENABLE = 1U << 0U ,
  PBA_ENABLE_GOLDPX_PRIORITY_MODE = 1U << 4U ,
  PBA_ENABLE_GOLDPX_FALLBACK_MODE = 1U << 5U ,
  PBA_ENABLE_MIN_DIST_SCOPE = 1U << 6U
}
 Enable flags for the pixel binning algorithm. More...
 
enum  argus_pba_averaging_mode_t {
  PBA_SIMPLE_AVG = 1U ,
  PBA_LINEAR_AMPLITUDE_WEIGHTED_AVG = 2U
}
 The averaging modes for the pixel binning algorithm. More...
 

Detailed Description

Pixel Binning Algorithm (PBA) parameter definitions and API functions.


Defines the generic Pixel Binning Algorithm (PBA) setup parameters and data structure.

The PBA module contains filter algorithms that determine the pixels with the best signal quality and extract an 1D distance information from the filtered pixels by averaging them in a specified way.

Basically, the Pixel Binning Algorithm is a multi-stage filter:

  1. A fixed pre-filter mask is applied to statically disable specified pixels.
  2. A relative and absolute amplitude filter is applied in the second stage. The relative filter is determined by a ratio of the maximum amplitude off all available (i.e. not filtered in stage 1) pixels. Pixels that have an amplitude below the relative threshold are dismissed. The same holds true for the absolute amplitude threshold. All pixel with smaller amplitude are dismissed.
    Note that the absolute amplitude threshold is disabled if the Golden Pixel (see below) is also disabled in order to prevent invalid filtering for multi-pixel devices.
    The relative threshold is useful to setup a distance measurement scenario. All well illuminated pixels are selected and considered for the final 1D distance. The absolute threshold is used to dismiss pixels that are below the noise level. The latter would be considered for the 1D result if the maximum amplitude is already very low.
    Those threshold are implemented using a hysteresis behavior. For its configuration, see the following parameters:
  3. An absolute minimum distance filter is applied in addition to the amplitude filter. This removes all pixel that have a lower distance than the specified threshold. This is used to remove invalid pixels that can be detected by a physically not correct negative distance.
    For its configuration, see the following parameters:
  4. A distance filter is used to distinguish pixels that target the actual object from pixels that see the brighter background, e.g. white walls. Thus, the pixel with the minimum distance is referenced and all pixel that have a distance between the minimum and the given minimum distance scope are selected for the 1D distance result. The minimum distance scope is determined by an relative (to the current minimum distance) and an absolute value. The larger scope value is the relevant one, i.e. the relative distance scope can be used to heed the increasing noise at larger distances.
    For its configuration, see the following parameters:
  5. If all of the above filters fail to determine a single valid pixel, the Golden Pixel is used as a fallback value. The Golden Pixel is the pixel that sits right at the focus point of the optics at large distances. Thus, it is expected to have the best signal at large distances.
    For its configuration, see the following parameters:
  6. In order to avoid unwanted effects from "out-of-focus" pixels in application that require a smaller focus, the Golden Pixel Priority Mode prioritizes a valid signal on the central Golden Pixel over other pixels. That is, while the Golden Pixel has a reasonable signal strength, it is the only pixel considered for the 1D result.
    For its configuration, see the following parameters:

After filtering is done, there may be more than a single pixel left to determine the 1D signal. Therefore several averaging methods are implemented to obtain the best 1D result from many pixels. See argus_pba_averaging_mode_t for details.

Enumeration Type Documentation

◆ argus_pba_averaging_mode_t

The averaging modes for the pixel binning algorithm.


Enumerator
PBA_SIMPLE_AVG 

Evaluate the 1D range from all available pixels using a simple average.

PBA_LINEAR_AMPLITUDE_WEIGHTED_AVG 

Evaluate the 1D range from all available pixels using a linear amplitude weighted averaging method. Formula: x_mean = sum(x_i * A_i) / sum(A_i)

◆ argus_pba_flags_t

Enable flags for the pixel binning algorithm.


Determines the pixel binning algorithm feature enable status.

     - [0]: #PBA_ENABLE: Enables the pixel binning feature.
     - [1]: reserved
     - [2]: reserved
     - [3]: reserved
     - [4]: #PBA_ENABLE_GOLDPX_PRIORITY_MODE: Enables the Golden Pixel
                                              priority mode feature.
     - [5]: #PBA_ENABLE_GOLDPX_FALLBACK_MODE: Enables the Golden Pixel
                                              fallback mode feature.
     - [6]: #PBA_ENABLE_MIN_DIST_SCOPE: Enables the minimum distance
                                        scope feature.
     - [7]: reserved
     .
Enumerator
PBA_ENABLE 

Enables the pixel binning feature.

PBA_ENABLE_GOLDPX_PRIORITY_MODE 

Enables the Golden Pixel Priority Mode. If enabled, the Golden Pixel is prioritized over other Pixels as long as it has a good signal (determined by #

PBA_ENABLE_GOLDPX_FALLBACK_MODE 

Enables the Golden Pixel Fallback Mode. If enabled, the Golden Pixel is used as a last fallback pixel to obtain a valid signal from. This is recommended for all non-multi pixel devices whose TX field-of-view is aligned to target the Golden Pixel in factory calibration.

PBA_ENABLE_MIN_DIST_SCOPE 

Enables the minimum distance scope filter.