COMPUTER UNIT FOR A LIDAR DEVICE, AND LIDAR DEVICE

Abstract

A computer unit for a LiDAR device, which has a laser source configured to emit a laser signal into a transmit path, and a LiDAR sensor arranged in a receive path and configured to detect a laser signal reflected into the receive path. The computer unit is configured to process a multiplicity of laser signal data points of the reflected laser signal. The computer unit is configured to filter halation out of the laser signal data points of the reflected laser signal.

Claims

1.-9. (canceled)

10. A computer unit for a LiDAR device, which has a laser source configured to emit a laser signal into a transmit path, and a LiDAR sensor arranged in a receive path and configured to detect a laser signal reflected into the receive path, the computer unit configured to: process a multiplicity of laser signal data points of the reflected laser signal, including filtering halation out of the laser signal data points of the reflected laser signal.

11. The computer unit as recited in claim 10, wherein the computer unit is configured to identify the laser signal data points that are allocated to a retroreflector.

12. The computer unit as recited in claim 11, wherein the computer unit is configured to calculate a reflectance of each of the laser signal data points using an intensity level later signal data point; compare the calculated reflectance of each laser signal data point with a predetermined reflectance threshold; and allocate the laser signal data points to a retroreflector when the calculated reflectance is greater than the predetermined reflectance threshold.

13. The computer unit as recited in claim 11, wherein the computer unit is configured to identify specific laser signal data points that are located in the same laser signal data point plane as the laser signal data points allocated to the retroreflector and/or are at the same distance from the LiDAR device as the laser signal data points allocated to the retroreflector.

14. The computer unit as recited in claim 13, wherein the computer unit is configured to determine an intensity level of the specific laser signal data points; compare the determined intensity level of each laser signal data point with a predetermined intensity level threshold; and allocate the specific laser signal data points to a real object when the determined intensity level is greater than the predetermined intensity level threshold and/or when the determined intensity level has a discontinuous curve.

15. The computer unit as recited in claim 14, wherein the computer unit is configured to: determine an echo duration of the specific laser signal data points; compare the determined echo duration with a predetermined echo duration threshold; and allocate the specific laser signal data points to a real object if the determined echo duration is greater than the predetermined echo duration threshold.

16. The computer unit as recited in claim 15, wherein the computer unit is configured to identify, out of the multiplicity of laser signal data points, the specific laser signal data points that are not allocated to a real object as being halation laser signal data points and to filter them out.

17. The computer unit as recited in claim 16, wherein the computer unit is configured to determine a probability that laser signal data points allocated to a real object are present in the halation laser signal data points.

18. A LiDAR device, comprising: a laser source configured to emit a laser signal into a transmit path; a LiDAR sensor arranged in a receive path and configured to detect a laser signal reflected into the receive path; and a computer unit configured to process a multiplicity of laser signal data points of the reflected laser signal, including filtering halation out of the laser signal data points of the reflected laser signal.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

[0033] Exemplary embodiments of the present invention will be explained in more detail on the basis of the following description and the figures.

[0034] FIG. 1A is a first front view of a roadway comprising a traffic sign mounted on a gantry.

[0035] FIG. 1B is a first front view of a multiplicity of laser signal data points that encode the view from FIG. 1A.

[0036] FIG. 2A is a second front view of the roadway comprising the traffic sign mounted on the gantry, with a real object arranged below the traffic sign.

[0037] FIG. 2B is a second front view of the multiplicity of laser signal data points that encode the view from FIG. 2A.

[0038] FIG. 3A is a side view of the roadway comprising the traffic sign and the real object from FIG. 2A, with a time-resolved intensity level distribution of the associated laser signal data points in a first configuration.

[0039] FIG. 3B is a side view of the roadway comprising the traffic sign and the real object from FIG. 2A, with a time-resolved intensity level distribution of the associated laser signal data points in a second configuration.

[0040] FIG. 3C is a side view of the roadway comprising the traffic sign and the real object from FIG. 2A, with a time-resolved intensity level distribution of the associated laser signal data points in a third configuration.

DETAILED DESCRIPTION OF EXAMPLE EMBODIMENTS

[0041] The present invention relates to a computer unit for a LiDAR device, which has a laser source configured to emit a laser signal into a transmit path, and a LiDAR sensor arranged in a receive path and configured to detect a laser signal reflected into the receive path, and the computer unit being configured to process a multiplicity of laser signal data points 4 of the reflected laser signal. In this case, a laser signal is emitted using the laser source. Said laser signal is reflected on objects in the surroundings of the LiDAR device. The reflected laser signal is detected in the receive path of the LiDAR device by way of the LiDAR sensor. A multiplicity of laser signal data points 4 mapping the surroundings of the LiDAR device are ascertained from said detected reflected laser signal.

[0042] FIG. 1A shows surroundings of the LiDAR device. These surroundings include a roadway 1 comprising a gantry 2 on which a traffic sign 3 is mounted.

[0043] FIG. 1B then shows the image of said surroundings from FIG. 1A by way of the multiplicity of laser signal data points 4. Each of said laser signal data points 4 has an intensity level 5 that is specified by way of a color code of the different laser signal data points 4. The darker a laser signal data point 4, the higher its associated intensity level 5. This relationship is likewise illustrated in the graph 6. On the horizontal axis, the intensity level 5 is plotted against the height 7. This results in a function 8 of the height 7 depending on the intensity level 5.

[0044] In FIG. 1B, it can also be seen that the size of the traffic sign 3 is overestimated in the image in the vertical by way of the laser signal data points 4. This is caused by “halation,” which is crosstalk with adjacent laser signal data points 4. It can be particularly pronounced in highly reflective or retroreflective objects such as a traffic sign 3. As can also be seen in FIG. 1B, the traffic sign 3 appears in the image as a wall on the roadway 1, so to speak, as a result of the laser signal data points 4. However, this erroneous depiction owing to halation is dangerous in terms of the road safety of motor vehicles driving autonomously on the roadway 1. For instance, an autonomously driving motor vehicle would have to trigger emergency braking in view of a “wall” on the roadway 1. This “false” reaction jeopardizes road safety and has to be avoided.

[0045] Another potential road safety hazard is illustrated in FIG. 2A. The roadway 1 having the gantry 2 and the traffic sign 3 is shown again. Now, though, a real object 9 is arranged on the roadway 1 below the traffic sign 3. This real object 9 may pose a risk to a vehicle driving autonomously. The autonomously driving vehicle may have to avoid the object 9 in some circumstances. For this purpose, though, the real object 9 also has to be reliably identified in the laser signal data points 4.

[0046] In particular, however, it can be seen in FIG. 2B that a real object 9 arranged below the traffic sign 3 (i.e., close to a retroreflector) may in some circumstances fall into the region of the halation in the laser signal data points 4. This poses the risk that the real object 9 will no longer be recognized in the halation. A graph 6 again shows the relationship between the height 7 and the intensity level 5. The associated function 8 shows a small plateau, corresponding to the real object 9, when the values for the height 7 are low.

[0047] In sum, therefore, there is a need both to eliminate halation (in particular in retroreflectors) and, at the same time, to retain the laser signal data points 4 that can be allocated to a real object 9 (in the halation region).

[0048] The computer unit according to the present invention is now configured to filter halation out of the laser signal data points 4 of the reflected laser signal. How this can be done will be described below in relation to FIG. 3A-3C.

[0049] In a first step, the laser signal data points 4 that correspond to a retroreflector (traffic sign 3) are identified. This is done by analyzing the intensity level 5 of the reflected laser signal using the computer unit. It is initially assumed that a retroreflector is larger than one pixel in the laser signal data points 4. The intensity level 5 then measured using the LiDAR sensor is proportional to the emitted laser power, the reflectance of the retroreflector, and the square of the distance between the LiDAR sensor and the retroreflector. In other words, the reflectance of the retroreflector can then be deduced from the measured intensity level 5. The reflectance of the retroreflector is calculated accordingly for each laser signal data point 4 using the computer unit. The calculated value is then compared with the value expected from a 100%-reflective Lambertian target. Retroreflectors have the unique property whereby their apparent Lambertian reflectance is much higher than 100%. Typical values are between 1,000% and 100,000%. All the laser signal data points 4 having a reflectance much higher than 100% are therefore identified as retroreflectors.

[0050] In a second step, all the specific laser signal data points 4 located in the same laser signal data point plane as the retroreflector are identified. In addition, all the specific laser signal data points 4 that are at the same distance from the LiDAR sensor as the laser signal data points 4 allocated to the retroreflector are identified. All these specific laser signal data points 4 are halation candidates. Three different cases may now occur.

[0051] In a first case (in FIG. 3A), an echo duration 10 of the reflected laser signal is greater than a distance between the real object 9 and the traffic sign 3. The figure shows the intensity level distribution 11 of the real object 9 and the intensity level distribution 12 of the traffic sign 3 as a function of time 13. In this case, the distance between the intensity level distribution 11 of the real object 9 and the intensity level distribution 12 of the traffic sign 3 is greater than the echo duration 10. Two separate laser signal data points 4 are detected.

[0052] In a second case (in FIG. 3B), the distance between the real object 9 and the traffic sign 3 is less than an echo duration 10 of the reflected laser signal. The figure shows the intensity level distribution 11 of the real object 9 and the intensity level distribution 12 of the traffic sign 3 as a function of time 13. In this case, the distance between the intensity level distribution 11 of the real object 9 and the intensity level distribution 12 of the traffic sign 3 is less than the echo duration 10. A single laser signal data point 4 is detected. However, this laser signal data point 4 has a greater echo duration than that of the intensity level distribution 11 of the real object 9 and that of the intensity level distribution 12 of the traffic sign 3. A superposed intensity level distribution 14 is measured.

[0053] In a third case (in FIG. 3C), the real object 9 and the traffic sign 3 are at the same distance from the LiDAR device. A single laser signal data point 4 is detected. However, this laser signal data point 4 has a greater peak intensity level than that of the intensity level distribution 11 of the real object 9 and that of the intensity level distribution 12 of the traffic sign 3. A superposed intensity level distribution 14 is measured.

[0054] The two parameters of echo duration 10 and intensity level distribution 11, 12 thus allow real objects 9 below the traffic sign 3 to be identified. In this case, it is also possible to take advantage of the fact that laser signal data points 4 for halation exhibit behavior as illustrated in FIGS. 1B and 2B. In this case, the intensity level 5 for laser signal data points 4 of this kind drops in the vertical starting from the traffic sign 3. This is exemplified in the function 8. This behavior is typical of halation since halation is reduced as the distance from the retroreflector increases. Real objects 9 do not exhibit this kind of behavior. In this case, discontinuous behavior is expected in the intensity level 5. As a result, halation can be identified.

[0055] In a third step, all the laser signal data points 4 that exhibit the behavior in terms of intensity level 5 that is typical of halation and those that have an echo duration equal to that of the reflected laser signal can then be filtered out. All the laser signal data points 4 having a discontinuous intensity level distribution are retained.

[0056] Although the present invention has been illustrated and described in more detail using preferred exemplary embodiments, the present invention is not limited to the disclosed examples and a person skilled in the art may derive other variations therefrom without departing from the scope of the present invention.