As the ToF distance measurement relies on the reflection of light sent out by the camera, any additional light, e.g., artificial light sources or sunlight, can influence the measurement results.
While the camera is able to measure the ambient light and then subtract it from the total light energy the pixels receive, there is a limit to the camera's ability to compensate for ambient light. The reason is that a pixel can only accept a certain amount of electrical charge. If this capacity is used up by ambient light, less space is available for the intended light pulse reflection and, as a consequence, the signal-to-noise ratio decreases.
To mitigate the influence of ambient light, blaze cameras have an optical bandpass filter. This filter only allows light of the same spectrum as the camera's light source (940 nm) to pass through to the sensor. You should avoid any artificial light sources in the vicinity of your camera that emit light in that spectrum as this will lead to noise in the acquired images. Similarly, operating another blaze camera at the same time will also cause interference. For measures how to avoid this, see Working with Multiple Cameras.
Sunlight is a minor problem as this isn't very strong in the 940-nm spectrum.
For tips how to best plan your camera setup in order to avoid overexposure, see Hardware Installation (blaze).
Reflections inside the camera lens or behind it, can create scattered light. Even though great care is taken during the manufacture of blaze cameras, e.g., by using special non-reflective coating for internal components, these reflections can't be eliminated completely.
Scattered light is also caused by bright surfaces and objects in the immediate vicinity of the light source. These surfaces and objects don't even need to be in the direct field of view of the camera to cause problems. Just by placing the camera in the middle of a tabletop, the light from the camera may be reflected by the tabletop straight back into the lens and can thus skew the distance measurement. The following figure illustrates these effects.
To combat the effects of scattered light, Basler has taken the following measures:
- Careful selection of the lens
- Limited aperture of the light source
The cover glass in front of the VCSELs must be clean.
For tips how to best plan your camera setup in order to avoid the creation of scattered light, see Hardware Installation (blaze).
Temperature has a significant impact on the accuracy of distance measurements. There are different aspects of this to consider:
- Camera requires stable temperature.
Measurement accuracy fluctuates with changes in temperature, even if the overall temperature is within the allowed temperature range. To achieve best possible results, the temperature should be kept stable. The optimum temperature is 22 °C. For recommendations on how to achieve this, see Providing Heat Dissipation (blaze).
- Camera requires warm-up time.
The camera needs approx. 20 minutes to achieve a stable operating temperature. As a stable temperature is a prerequisite for accurate measurements, allow 20 minutes before starting to acquire images.
- Observe the operating temperature range specified in the "Environmental Requirements" section of your camera model topic.
At temperatures above 50 °C, reliable camera operation can't be guaranteed anymore. At housing temperatures below 0 °C, the camera won't start.
The blaze cameras are equipped with an over temperature protection mechanism. Therefore, even if you're operating the camera at temperatures outside the recommended range, the risk of damage or injuries is low.
The camera's Thermal Drift Correction feature can be used to counter the influence that the temperature has on measurement accuracy.
Multiple Reflections (Multipath)#
For accurate distance measurement, only light reflected once delivers reliable data. Any light that has been reflected several times, by other objects in the camera's field of view or the environment in general, can falsify the measurement. Concave forms, like corners of a room or the inside of a coffee cup, are particularly problematic as the light pulse can bounce back and forth between the different surfaces, thus increasing the time until it is received by the sensor.
Mirrors and highly-reflective surfaces, e.g., lacquered tabletops, can also lead to multiple reflections or can even deflect the light pulse completely. The following figure illustrates these effects.
For tips how best to plan your camera setup in order to avoid multiple reflections, see Hardware Installation (blaze).
Reflectivity of Target Objects#
The reflectivity of the target object also has an influence on the accuracy of the measurement. Two aspects need to be considered:
- Reflection type determined by the surface quality of the target object
- Target object color
Two kinds of reflection can occur:
- Diffuse reflection
This occurs in matte objects (e.g., wood or paper). Reflectance is uniform which means that the light pulses are reflected equally at all angles. In contrast to specular reflections described below, this kind of reflection is preferable because the intensity of the light reflected back into the sensor isn't influenced by the angle.
- Specular reflection
This occurs in glossy objects (e.g., polished metal or, in extreme cases, mirrors) and transparent materials. Reflection will occur according to the law of reflection which states that incoming light of a certain direction will be reflected into one outgoing direction, with incoming and outgoing light pulses at equal angles to the perpendicular of the object surface. This means that the reflected light pulse will follow a different path than the light pulse sent out by the camera which can lead to multiple reflections.
The other danger is that the reflected light hits the sensor directly which occurs when the light pulse hits the perpendicular of the object surface. This can lead to the pixels becoming saturated with the energy from a single light pulse.
For the above reasons, accurate distance measurement of glossy or transparent materials is difficult.
Target Object Color#
Bright objects provide better results as, generally speaking, they are able to reflect more light, whereas with dark objects a certain amount of light is absorbed which will therefore not be returned back to the sensor.
Another aspect to consider is the distance of the object from the sensor. While bright objects generally yield more accurate results, if they are placed close to the sensor, they reflect back so much light that oversaturation occurs. Once pixels become oversaturated, they don't contribute to the distance measurement anymore. As a result, the measurement quality decreases.
You also need to be aware that the camera works with near-infrared light, and it is difficult to predict whether an object actually reflects the near-infrared light well enough. Therefore, the distinction made above can only serve as a general rule of thumb.
The intensity image can help you to assess the accuracy of your measurements. Objects that are well illuminated without being under- or oversaturated in the intensity image, will produce accurate measurement results.
The problem of ambiguity arises because light can travel beyond the range that the camera is specified for and the camera considers this space an undefined area. The effect is that at the end of the non-ambiguity range a new measurement interval starts again at 0 m. This is inherent in the camera because of the timing pattern of the light pulses.
This means that objects outside the range will also reflect light back to the camera. To deal with this, the camera "folds back" the results into the non-ambiguity range.
When you're operating blaze cameras in one of the standard operating modes, the non-ambiguity range is 0–30 m. Beyond 30 m, not much light will be reflected back to the camera. Therefore, ambiguity isn't much of a problem.
If you're using the Fast Mode, however, the non-ambiguity range is 0–1.5 m in Short Range and 0–10 m in Long Range. This means that only distances of objects measured in these ranges reflect the objects' true distance.
To better understand the problem, consider the following example:
Objects at 1 m outside the range will appear at 1 m within the non-ambiguity range. In other words, in Long Range, if an object is 11 m away from the camera, its distance is measured as 1 m.
The following figure illustrates this.