Method and system for detecting objects within the field of view of a lidar device.

Abstract

This enables precise and reliable object detection for detecting objects within the field of view of a LiDAR device. [Solution] The present invention relates to a method for detecting an object within the field of view of a lidar device, wherein the lidar device includes a transmitting unit including a laser source, a receiving unit including a detector, and a processor unit. Multiple light pulses are emitted within the solid angle, The secondary light received from the solid angle is measured by the detector of the receiving unit and converted into a digital signal. Signal aggregation is performed, and coincidence filters are used to identify signal structures that partially match in terms of amplitude. The present invention relates to a method by which signal aggregation is performed by a threshold filter, and a processor unit compares the received or aggregated signal to an applicable threshold, thereby distinguishing the secondary light of the received object from interfering signals such as background noise or interference light, and thus filtering it out.

Description

Technical Field

[0001] The present invention relates to a method and a system for detecting an object within the field of view of a rider device.

Background Art

[0002] Known methods for detecting an object within the field of view by a rider system generally rely on the transmission of optical impulses and the measurement of the reflected secondary light in order to determine the position and distance of the object. However, some of these methods have limitations, especially in dealing with interference signals and the accuracy of object detection.

[0003] A typical rider system transmits optical pulses within a plurality of solid angles and captures the reflected light by a receiving unit. The received light is then converted into a signal and evaluated by a processor unit to determine distance data and position data.

[0004] The methods known from the state of the art are often limited by their inflexible filtering mechanisms, with limited resolution and insufficient adaptability to changing environmental conditions. These problems can reduce the accuracy and reliability of object detection, especially in demanding or dynamic environments.

Summary of the Invention

Problems to be Solved by the Invention

[0005] The problem of the present invention is to provide a method that enables more precise and reliable object detection for detecting an object within the field of view of a rider device.

Means for Solving the Problems

[0006] The present invention relates to a method for detecting an object within the field of view of a lidar (Light Detection and Ranging) device, wherein the lidar device includes a transmitting unit, a receiving unit, and a processor unit. Multiple light pulses are emitted within adjacent solid angles, and the secondary light of the object received from these adjacent solid angles is measured by at least one detector of the receiving unit and converted into individual digital signals for each adjacent solid angle. The signals from adjacent solid angles are aggregated without reducing the angular resolution. During this aggregation, a first filter is used to recognize signal structures in the signals from different solid angles that match at least partially, particularly in terms of their amplitude, in order to identify a particular object. In addition, a threshold filter is used to aggregate the signals, in which the processor unit compares the received and / or aggregated signals with a suitable threshold so that the secondary light of the received object is distinguished from and thus filtered out from interfering signals such as background noise and / or interfering light.

[0007] A lidar device is a technical device for optically capturing an object. This device transmits light pulses and measures the reflected secondary light. A transmitting unit is used to emit light pulses within various solid angles. A receiving unit includes at least one detector that receives the secondary light, which is converted into a digital signal. A processor unit evaluates these signals by aggregating and filtering them. Coincidence filters recognize matching signal structures in signals within adjacent solid angles, while threshold filters separate the received secondary light from interfering signals.

[0008] The advantage of this method lies in the fact that signal aggregation improves the signal-to-noise ratio without reducing angular resolution. This results in more precise object detection because interfering signals are effectively filtered out. The combination of coincidence filters and threshold filters allows for increased sensitivity of the lidar device and improved distinction between relevant signals and background noise.

[0009] It is advantageous that a deflection unit equipped with a rotating mirror can be used to scan the solid angle along the horizontal and / or vertical directions. A deflection unit is a mechanical and optical device that directs light pulses into various solid angles to extend the field of view of a lidar device. A rotating mirror ensures that the light pulses are emitted into the desired solid angle.

[0010] This achieves complete coverage of the LiDAR device's field of view, which improves the recognition of objects at different positions and within solid angles. It is advantageous that the receiving unit may include a detector connected to a single macropixel, in which case at least four, preferably sixteen, subpixels may be interconnected to the single macropixel.

[0011] The signals from subpixels are aggregated into a single common signal from macropixels to improve the signal-to-noise ratio. A single macropixel is a group of subpixels, and the signals from these subpixels are combined to produce a stronger, less noisy signal. This structure allows for higher sensitivity in lidar devices.

[0012] It is advantageous that the light pulse can be emitted as one light impulse per solid angle, or as multiple light impulses per solid angle, preferably at least three light impulses per solid angle.

[0013] An optical impulse is a short emission of light transmitted by a transmitting unit. Each individual optical pulse consists of a single optical signal, while multiple optical pulses transmit multiple signals within the same solid angle to improve data acquisition.

[0014] In this case, multiple light pulses may be emitted sequentially within the same solid angle. This increases the probability of detecting reflected secondary light, especially in the case of weakly reflective objects, which improves the recognition rate.

[0015] It is advantageous that the processor unit can evaluate aggregated signals to detect objects within adjacent solid angles. The processor unit processes aggregated signals from adjacent solid angles to determine the position and distance of objects within the field of view. In this process, objects are captured as reflected secondary light within the signals from adjacent angles.

[0016] This allows for a comprehensive analysis of signals from various angles, improving the precision of object detection. It is advantageous that signal aggregation can be performed within a single continuous scan cycle, thereby enabling continuous measurement of secondary light within adjacent solid angles of received data.

[0017] A scan cycle is the period during which the field of view of the lidar device is continuously scanned to collect and process data without interruption. This enables real-time monitoring of the field of view, which improves the responsiveness of the lidar device.

[0018] The aggregation of signals within multiple solid angles can be performed by adaptive filtering that matches the characteristics of the received signal, in which case it is advantageous that the individual thresholds of the threshold filter can be fitted to the amplitude of background noise and / or the threshold of the threshold filter can be fitted to the distance of the captured object and / or, in the case of an object recorded at close range, the threshold of the threshold filter can be fitted to the amplitude of interfering light.

[0019] Adaptive filtering is a process in which a filter dynamically adjusts its parameters to match the characteristics of the received signal at any given time in order to maximize the efficiency of signal processing. Threshold filters distinguish secondary light from reflected objects from background noise and interference, and adjust their threshold according to the signal conditions.

[0020] This enables the filter to respond flexibly to changes in background noise, the distance of the object, and the intensity of interfering light, thus improving signal separation in various environments. The processor unit may perform real-time correction of the signal to improve the accuracy of object detection. In this case, it is advantageous for the recorded signal to be inspected by real-time correction for defects caused, for example, by malfunctions of the detector in the receiving unit and / or by malfunctions of the laser in the transmitting unit.

[0021] Real-time correction is a process in which the processor unit inspects the received signal for defects in real time and automatically corrects them. Such defects can be caused by various malfunctions, for example, by a malfunction of the detector in the receiving unit or by a malfunction of the laser in the transmitting unit. Real-time correction makes it possible to immediately identify and remove these defects without interrupting the ongoing signal processing.

[0022] As a result, signal defects caused by malfunctions in the transmitting unit or the receiving unit are corrected in real time, significantly improving the accuracy of object detection. Advantageously, the deflection unit may be configured to deflect the emitted light pulses in a predetermined pattern, preferably in a meandering or spiral pattern, in order to enable a comprehensive scan of the field of view.

[0023] The deflection unit deflects the light pulses in a specific pattern to ensure a systematic scan of the entire field of view of the lidar device. Patterns such as meandering or spiral enable uniform coverage of the field of view.

[0024] This achieves an overall capture of the field of view. The processor unit may include a learning algorithm based on machine learning, and the learning algorithm is trained to detect an object using training data of older measurements. The recognized object is compared with the actually existing object, and the measurements that correctly recognize the object are weighted more strongly, thereby advantageously improving the accuracy of object detection.

[0025] The learning algorithm (artificial intelligence algorithm) is a machine learning system optimized by training data. The algorithm compares the recognized object with the actual object and uses the successfully recognized object to improve the accuracy of the algorithm.

[0026] Thereby, the algorithm learns from the training data and the recognition rate of the algorithm increases, so that the detection accuracy is continuously improved. It is advantageous that the receiving unit may be configured to simultaneously process a signal by a plurality of detectors.

[0027] Simultaneous processing means that the receiving unit simultaneously captures and evaluates a plurality of signals without delaying the processing of individual signals. Thereby, the efficiency of signal processing is increased, which results in faster and more precise object detection.

[0028] Alternatively, signal aggregation may be performed by sequential processing of optical pulses received sequentially in time, and in this case, it is advantageous that the temporal correlation between individual signals is used to recognize an object.

[0029] Sequential processing is a process in which the received optical pulses are processed in the order of their arrival, and the temporal context between signals is utilized for object detection. Thereby, the ability of this system to precisely recognize movements and changes within the field of view is improved.

[0030] Signal aggregation can be performed within adjacent solid angles in the range of 0.01° to 0.1°. For this purpose, in the first step, the measured signals of multiple solid angles are converted into a histogram; in the second step, the signal structure is recognized for each individual solid angle by applying a first threshold filter with a suitable threshold; and in the third step, the recognized signal structures for each individual solid angle are compared by a coincidence filter and / or a second threshold filter with a second suitable threshold is applied, thereby allowing secondary light from the received object to be distinguished from and filtered out from interfering signals such as background noise and / or interference light.

[0031] A histogram is a graphical representation of the distribution of signal values ​​and is used in this case to process signals from various solid angles. A threshold filter recognizes the signal structure, while a coincidence filter checks for the consistency of these structures.

[0032] This effectively filters out interfering signals and correctly identifies the relevant signal structures, thereby improving the precision of object detection. Signals within adjacent solid angles within the range of 0.01° to 0.1° can be aggregated, for which in the first step the measured signals of multiple solid angles are converted into histograms; in the second step the individual histograms of at least two adjacent solid angles are aggregated to produce at least one aggregated histogram; in the third step the signal structure is recognized within the aggregated histogram by applying a first threshold filter with a suitable threshold; and in the fourth step the recognized signal structures for each solid angle are compared by a coincidence filter and / or a second threshold filter with a second suitable threshold is applied so that secondary light from a received object in the signal is distinguished from and thus filtered out from interfering signals such as background noise and / or interfering light.

[0033] The aggregated histogram combines the signal distributions from multiple solid angles and is processed by threshold filters and coincidence filters to extract relevant signals and eliminate interfering signals.

[0034] Aggregating histograms improves detection efficiency and object detection accuracy, thereby optimizing signal processing. A further subject of the present invention is a system comprising a transmitting unit, a receiving unit, a deflection unit, and a processor unit for detecting objects within the field of view of a lidar device. The transmitting unit emits multiple light pulses into adjacent solid angles, and the receiving unit measures the secondary light received from these solid angles and converts it into a digital signal. The signals from adjacent solid angles are aggregated without reducing the angular resolution, and the signals are filtered by a threshold filter. The processor unit compares the received and / or aggregated signals to an applicable threshold to distinguish the echo signal of the object from background noise and filter out the background noise.

[0035] This system enables precise and efficient object detection based on the method described above by aggregating signals from adjacent solid angles and removing interfering signals using a threshold filter, thereby improving the accuracy and reliability of the lidar device.

[0036] The present invention will be explained based on the following drawings. [Brief explanation of the drawing]

[0037] [Figure 1] This is a schematic diagram illustrating the method for detecting objects within the field of view of a LiDAR device. [Figure 2] This is a schematic diagram of the histogram and echograph of the method based on Figure 1. [Figure 3] This is a schematic diagram intended to clarify alternative methods in comparison with Figure 1. [Modes for carrying out the invention]

[0038] Figure 1 shows a schematic diagram illustrating a method for detecting objects within the field of view of a lidar device, which includes a transmitting unit containing a laser, a receiving unit containing a detector, and a processor unit. Within a first solid angle 1 of 0.05°, three light pulses 2 emitted sequentially or simultaneously in time are captured by the detector. Within a second solid angle 3, three light pulses 4 are also captured by the detector. Within a third solid angle 5, three light pulses 6 are also captured by the detector. Subsequently, a first histogram 7 is generated for the measurement data 15 at the first solid angle 1, a second histogram 8 for the measurement data 16 at the second solid angle 3, and a third histogram 9 for the measurement data 17 at the third solid angle 5. This measurement data 15, 16, and 17 is recorded by a single macropixel, in which case at least four, preferably sixteen, subpixels may be interconnected to the single macropixel. Histograms 7, 8, and 9 are graphical representations of the measurement data 15, 16, and 17, where the amplitude of individual pixels is graphically displayed. In a further step, using threshold filters with adaptable thresholds, the secondary light of the received object is determined in the first echograph 10 for the first histogram 7, in the second echograph 11 for the second histogram 8, and in the third echograph 12 for the third histogram 9. In a further step of this method, the first echograph 10 and the second echograph 11 are compared using a coincidence filter, as indicated by the arrows, to generate a first aggregated echograph 13, in which mismatched values ​​of the received secondary light of the object are filtered out, and only matching values ​​of the received secondary light are retained. Correspondingly, the second echograph 11 and the third echograph 12 are compared to generate a second aggregated echograph 14. The use of individual threshold filters and coincidence filters ensures reliable object recognition by the lidar device, while filtering out interfering signals such as background noise and / or stray light.The application of threshold filters to determine echographs 10, 11, and 12, and the subsequent application of coincidence filters to generate aggregated echographs 13 and 14, ensures reliable determination of secondary light of objects and ensures reliable filtering out of background noise and / or interfering light.

[0039] Figure 2 shows a further schematic diagram to clarify the method based on Figure 1, where the first histogram 7 is shown as a graph of the lidar device's measurements. The first graph 20 shows the value of the first light impulse within a first solid angle 1, the second graph 21 shows the measurement of the second light impulse, and the third graph 22 shows the measurement of the third light impulse. The x-axis shows the individual pixels of the lidar device's detector, and the y-axis shows the amplitude of the received light measurement. Correspondingly, the second histogram 8 for a second solid angle 3 is shown as a fourth graph 23 with the value of the first light impulse within the second solid angle 3, as a fifth graph 24 with the value of the second light impulse, and as a sixth graph 25 with the value of the third light impulse. In a further process step, a threshold filter with a suitable threshold 28 is used to generate the first echograph 10 from the first histogram 7 and the second echograph 11 from the second histogram 8. In echographs 10 and 11, the x-axis represents time t related to the passage of time, and the y-axis represents the amplitude A of the measured values. In a further process step, the first echograph 10 and the second echograph 11 are aggregated using a coincidence filter to form a first aggregated echograph 13, in which non-matching secondary light measured values ​​26 are filtered out, while matching secondary light measured values ​​27 are retained, thereby ensuring the detection of the object. In a further process step, an additional second threshold filter may be used to filter out secondary light measured values ​​26 that are below a suitable or defined threshold 29.

[0040] Figure 3 shows a schematic diagram illustrating an alternative embodiment of the present method, which differs from the method based on Figures 1 and 2. In a certain process step, a first aggregated histogram 30 is generated from a first histogram 7 and a second histogram 8, as indicated by the arrows, and a second aggregated histogram 31 is generated from the second histogram 8 and a third histogram 9. This aggregation can be performed, for example, by superimposing the individual pixels or values ​​of these histograms. In this case, in a further process step, a threshold filter generates a first echograph 32 from the first aggregated histogram 30 and a second echograph 33 from the second aggregated histogram 31. In a further step not shown, coincidence filters may be applied to both echographs 32 and 33 to generate aggregated echographs and filter out defective secondary light, as shown in Figures 1 and 2. Instead of the methods based on Figures 1 and 2, the signal data is directly aggregated by applying aggregation of individual histograms 7, 8, and 9 to generate aggregated histograms 30, 31 and by applying threshold filters to generate echographs 32 and 33. This certainly requires higher computational power, but it allows for improved determination of secondary light in the signal and ensures that background noise and / or interfering light are reliably filtered out. [Explanation of Symbols]

[0041] 1, 3, 5 solid angle 2, 4, 6 light pulses 7, 8, 9 Histograms 10, 11, 12 Echograph 13, 14 Aggregated echograph 15, 16, 17 Measurement data 20 Graph 1 21. Second graph 22 Third Graph 23. The fourth graph 24. Graph No. 5 25. Graph No. 6 26. Mismatched measurements 27 Matching measurements 28 Applicable thresholds 29. Applicable or specified thresholds 30, 31 Aggregated histograms 32, 33 Echograph

Claims

1. A method for detecting an object within the field of view of a lidar device, wherein the lidar device includes a transmitting unit including a laser source, a receiving unit including at least one detector, and a processor unit. Multiple light pulses (2, 4, 6) are emitted within adjacent solid angles (1, 3, 5), The secondary light of the object received from the adjacent solid angles (1, 3, 5) is measured by at least one detector of the receiving unit and converted into individual digital signals for each of the adjacent solid angles. The aggregation of adjacent solid angle signals (1, 3, 5) is performed without reducing the angular resolution, and during the aggregation, for a particular object, a first filter is used to recognize signal structures that, in particular with respect to their amplitude, are at least partially identical among the individual signals of different solid angles. A method wherein the aggregation of the signal is further performed by a threshold filter, in which the processor unit compares the received and / or aggregated signal with a threshold, thereby distinguishing and filtering out secondary light from background noise and / or interfering signals in the signal.

2. The method according to claim 1, wherein the first filter is a coincidence filter, and the aggregation of signals within a plurality of solid angles (1, 3, 5) is performed by adaptive filtering that matches the characteristics of the received signal, in which case the individual thresholds of the threshold filter are adapted to the amplitude of the background noise and / or the thresholds of the threshold filter are adapted to the distance of the captured object and / or, in the case of an object recorded at close range, the thresholds of the threshold filter are adapted to the amplitude of interfering light.

3. The method according to claim 1 or 2, wherein a deflection unit equipped with a rotating mirror is used to scan a solid angle (1, 3, 5) along the horizontal and / or vertical directions.

4. The method according to any one of claims 1 to 3, wherein the receiving unit includes a detector, i.e., subpixels connected to a single macropixel, with respect to at least four subpixels, preferably at least sixteen subpixels, connected to a single macropixel, and the signals of the subpixels are aggregated into one common signal of the macropixel, thereby improving the signal-to-noise ratio.

5. The method according to any one of claims 1 to 4, wherein the optical pulses (2, 4, 6) are emitted as one optical impulse for each solid angle (1, 3, 5), or as multiple optical impulses for each solid angle, preferably at least three optical impulses for each solid angle.

6. The method according to any one of claims 1 to 5, wherein the processor unit evaluates the aggregated signals to detect an object within adjacent solid angles (1, 3, 5).

7. The method according to any one of claims 1 to 6, wherein the aggregation of signals within one scan cycle is performed, which involves continuous measurement of secondary light within the adjacent solid angles (1, 3, 5) received.

8. The method according to any one of claims 1 to 7, wherein the processor unit performs real-time correction of the signal to improve the accuracy of object detection.

9. The method according to any one of claims 1 to 8, wherein the processor unit includes a machine learning-based learning algorithm, the learning algorithm is trained to detect objects using training data from older measurements, recognized objects are compared to actual existing objects, and measurements that correctly recognize objects are given greater weight, thereby improving the accuracy of object detection.

10. The method according to any one of claims 1 to 9, wherein the receiving unit is configured to process a signal simultaneously with a plurality of detectors.

11. The method according to any one of claims 1 to 10, wherein the aggregation of signals is performed by sequential processing of the optical pulses (2, 4, 6) received in succession in time, and in this case, the time correlation between the individual signals is used to recognize an object.

12. The method according to any one of claims 1 to 11, wherein the aggregation of signals within adjacent solid angles (1, 3, 5) in the range of 0.01° to 0.1° is performed, for this purpose, in a first step, the measured signals of a plurality of solid angles (1, 3, 5) are converted into histograms (7, 8, 9); in a second step, for each individual solid angle (1, 3, 5), the signal structure is recognized by applying a first threshold filter having a suitable threshold; and in a third step, the recognized signal structures for each individual solid angle (1, 3, 5) are compared by a coincidence filter and / or a second threshold filter having a second suitable threshold is applied so that secondary light of a received object in the signal is distinguished from and filtered out from interfering signals such as background noise and / or interfering light.

13. The method according to any one of claims 1 to 12, wherein the signals within adjacent solid angles (1, 3, 5) in the range of 0.01° to 0.1° are aggregated, for this purpose, in a first step, the measured signals of a plurality of solid angles (1, 3, 5) are converted into histograms (7, 8, 9); in a second step, the individual histograms (7, 8, 9) of at least two adjacent solid angles (1, 3, 5) are aggregated to generate at least one aggregated histogram (30, 31); in a third step, the signal structure is recognized within the aggregated histogram (30, 31) by applying a first threshold filter having a suitable threshold; and in a fourth step, the recognized signal structure for each solid angle (1, 3, 5) is compared by a coincidence filter and / or a second threshold filter having a second suitable threshold is applied so that secondary light of a received object in the signal is distinguished from and filtered out from interfering signals such as background noise and / or interfering light.

14. A system for detecting objects within the field of view of a lidar device, A transmitting unit equipped with at least one laser source, A receiving unit equipped with at least one detector, Deflection unit and, It includes a processor unit, The system is configured such that multiple light pulses (2, 4, 6) are emitted within adjacent solid angles (1, 3, 5). The secondary light received from the adjacent solid angles (1, 3, 5) is measured by at least one detector of the receiving unit and is configured to be converted into individual digital signals for each of the adjacent solid angles (1, 3, 5). The aggregation of the signals of adjacent solid angles (1, 3, 5) is performed without reducing the angular resolution. The signal aggregation is performed by a threshold filter. A system in which the processor unit compares the received and / or aggregated signals to an applicable threshold in order to distinguish the echo signal of an object from background noise, and therefore to filter out the background noise.

Images