VL53L8CX and sunlight

[Samb22]

Hello, 
I have tested, the VL53L8CX tof sensor in interior (fixation at 3m10), and the visualization of the matrices works very well, in an empty scene or with a person pass under the sensor. 

When my sensor is fixed at the same height, but light up with daylight, next to the window, my matrices are noised, and when a person pass under it, there are some noise with the passage of the person. 

For example, the sensor create lots of pixels with this situation rather than the other. 
How to reduce this noise and this sensibility to the daylight ? 

[Bin]

Hi:

What you observe is expected in a strong daylight environment, especially near a window. Higher ambient IR can reduce the signal-to-noise ratio and may generate more noisy zones or additional responses in the matrix.

To improve the behavior, we recommend the following:

1. Use 4x4 resolution

This is generally more robust than 8x8 under strong ambient light.

vl53l8cx_set_resolution(&Dev, VL53L8CX_RESOLUTION_4X4);

2. Use Continuous mode

Continuous mode usually provides better ambient light immunity than Autonomous mode.

vl53l8cx_set_ranging_mode(&Dev, VL53L8CX_RANGING_MODE_CONTINUOUS);

3. Reduce the ranging frequency

For example, try 5 Hz or 10 Hz.

vl53l8cx_set_ranging_frequency_hz(&Dev, 5);

4. Increase the sharpener

This can help reduce the spreading of a target into neighboring zones.

vl53l8cx_set_sharpener_percent(&Dev, 20);

5. Use Strongest target order

This is often more stable in difficult conditions.

vl53l8cx_set_target_order(&Dev, VL53L8CX_TARGET_ORDER_STRONGEST);

6. If Autonomous mode must be used, try increasing the integration time

For example:

vl53l8cx_set_integration_time_ms(&Dev, 10);

or try higher values such as 15 ms or 20 ms.

7. Filter the output using output date

Also, if possible, please check the mechanical setup to reduce direct or reflected daylight entering the FoV.

Hope these can help you, Thanks~

Best regards.

Bin FAN

[Samb22]

Hello, 

Thank you for your valuables answers. 
I succeed to filter the sunlight that comes in my sensor, in modifiying the integration time at 30ms. 

But i noticed that the frequency is lower now. In fact, i record data during 30s and i have only 233 matrices. Also when i don’t have integration time in my soft i was at 15Hz with around 450 matrices for 30s (450/30=15). 
Why this parameter influence the frequency ? 
And was is really integration time in case of our TOF? 

Thanks and regards 

Samuel

[Bin]

Hi:

if the ranging mode is VL53L8CX_RANGING_MODE_CONTINUOUS, integration time (Timing Budget) setting decides the frequence.

you can refer to the below diagram, it means when the first data ready, the operation of the second frame of data will immediately begin, and so on...

But if the ranging mode is VL53L8CX_RANGING_MODE_AUTONOMOUS, for example the below diagram: set the frequency 15hz means inter-measurement time is 66.67ms, the integration time (Timing Budget) setting less than the period for save power.

 

Best Regards,

Bin FAN