TWI858694B - Lidar system and resolusion improvement method thereof - Google Patents

Abstract

A LiDAR system includes a microcontroller, a laser light source, a lens module and a receiver. The lens module includes a receiver lens module and a laser beam splitter module. The laser beam splitter module includes a diffractive optical element and a collimation lens assembly. The laser light source emits a plurality of laser beams with different wavelengths and includes a light coupler. The light coupler optically couples the laser beams into a collimated light signal. In a shutter time of each subframe of a frame, a plurality of pixels of the receiver receives reflective at least one light signal of the laser light with different wavelengths to obtain a plurality of subframes of environmental images, and takes a distance value represented by the reflective light signals as a distance value of the pixels of the subframe, the microcontroller merges the distance values of the pixels of the plurality of subframes of the environmental images as a final distance value of the frame.

Description

LiDAR system and resolution enhancement method thereof

The present invention relates to a lidar system, and in particular to a lidar system with improved resolution.

In recent years, LiDAR (Light Detection and Ranging) technology has been widely used in the autonomous/semi-autonomous driving and safety warning of automobiles. The main components of LiDAR include: sensors (such as direct time of flight (D-ToF) sensors), laser light sources, scanners and data processors. Current lidar scanning methods can have many forms, such as using an optical phased array (OPA) or a diffractive optical element (DOE) to project a small area of light, using a micro-electromechanical system (MEMS) micro-vibration mirror or a polygon mirror to perform a serpentine back-and-forth scan or oblique scan of a large area, or using a DOE or a multi-point linear arrangement of light sources or multiple reflections to expand the beam and project a linear beam, and use mechanical rotation to scan a large area horizontally. Through the above scanning methods, the sensor can receive the reflected light signal.

However, the laser light source sensing performed in the above scanning method has a small aspect ratio (screen ratio), so it is necessary to continuously receive the reflected light signal at a higher frequency. In contrast, flash LiDAR can achieve high-frequency, high-frame sensing with lower system computing requirements and overall energy consumption by projecting a large area of light spots at one time. In the case of a fixed light spot density, if the imaging resolution of the flash LiDAR can be further improved, the imaging can be made clearer, the accuracy of distance measurement can be improved, and driving safety can be further improved. Therefore, there is an urgent need for a LiDAR system with a higher resolution than existing technologies with the same light spot density, so as to correctly judge the distance and maintain driving safety. In addition, there is an urgent need for a resolution enhancement method to improve the resolution of the lidar system compared to existing technologies with the same light spot density, so as to accurately judge the distance and maintain driving safety.

The main purpose of the present invention is to provide a lidar system with a higher resolution than the existing technology with the same light spot density, so as to accurately judge the distance and maintain driving safety.

In order to achieve the above-mentioned purpose, the present invention provides a lidar system, including: a microcontroller; a laser light source coupled to the microcontroller; a lens module; and a receiver coupled to the microcontroller, wherein the laser light source emits a plurality of laser lights of different wavelengths, and includes an optical coupler and an optical fiber, the optical coupler optically couples the laser lights into a collimated light signal, and transmits the collimated light signal through the optical fiber. The lens module includes a laser spectroscope module and a receiver lens module, the laser spectroscope module receives the laser light emitted from the laser light source, and diverts the laser light into a plurality of diverted lights, and the diverted lights are emitted toward a target. The laser spectroscope module includes a diverting optical element and a collimating lens assembly. The receiver lens module receives a reflected light signal reflected after the diffracted light contacts the target, and sends the reflected light signal to the receiver. The laser light source emits a pulse signal in a cycle time. The microcontroller controls the receiver to open in a sensor shutter time and close in a reset time in each cycle time. In a sensor shutter time of a subframe of a frame, a plurality of pixels of the receiver receive at least one reflected light signal of the laser light of different wavelengths, obtain environmental images of a plurality of subframes, and use the distance values represented by the reflected light signals as the distance values of the pixels of the subframe. The microcontroller fuses the distance values of the pixels of the multiple sub-frames of the environmental image into a final distance value of the frame.

In order to achieve the above-mentioned purpose, the present invention provides a method for improving the resolution of the above-mentioned lidar system, the method comprising: setting the diffraction optical element as a movable part with the function of rotation and/or reciprocating movement. Under the conditions of multiple rotation angles or reciprocating movement positions, multiple sub-frame environmental images are obtained. Each pixel of the reflected light signals in the environmental image represents a sub-range value, and each sub-frame environmental image is composed of multiple sub-range values to form a three-dimensional image with depth information. After eliminating abnormal sub-frames, the remaining multiple sub-frame environmental images are fused. If the same pixel has multiple sub-range values, they are averaged or selected. If the same pixel has only one sub-range value, the sub-range value is selected. If the pixel has no range value, the maximum or minimum value (0) in the range measurement range is selected as the final range value of the three-dimensional image of the frame.

The effect of the present invention is that, under the condition of fixed light spot density, the imaging resolution of the flash lidar is further improved, the imaging is clearer, the accuracy of distance measurement is improved, and driving safety is further improved.

100: LiDAR system

101:Microcontroller

102: Laser light source

104: Laser light

106: Lens module

108: Receiver lens module

110: Laser spectrometer module

112: Receiver

120: Target

122: Image field

124: Field of view

126: Reflected light

202: Concave lens

204: Convex lens

206:Diffraction optical element

208: Concave mirror

210: Collimator lens set

221,222,223,224:Laser light

230: Optocoupler

240: Optical fiber

a: Seam width

θ ₁ : diffraction angle

L: Projection distance

y ₁ ,y ₂ ,...-y ₁ ,-y ₂ ,...: dark pattern position

λ ₁ ,λ ₂ ,λ ₃ : wavelength

302: Laser light

304:Diffraction optical element

306a,306b,306c: Point cloud

310a, 310b: diffraction optical element

502: Collimator lens set

5021: Concave lens

5022: Convex lens

504:Diffraction optical element

506:Laser light

508:Diffuse light

512: Collimator lens set

5121: Concave lens

5122: Convex lens

514:Diffraction optical element

516:Laser light

518:Diffuse light

520: Concave mirror

PW: Pulse Width

T: cycle time

SS:Sensor shutter time

R: Reset time

Ts: start time

Tl: End time

800:Method

802,804,806: Steps

901,902,903:Laser light

911,912,913:Laser light

921: Laser light

A,B,C,D,E: Sampling blocks

FIG1 is a schematic diagram of a lidar system of the present invention; FIG2A and FIG2B are schematic diagrams of the internal structure of some components shown in FIG1; FIG2C and FIG2D are schematic diagrams of single slit diffraction; FIG2E shows the overlapping of point clouds generated by laser lights of different wavelengths; FIG3A and FIG3B show the operation of the diffraction optical element; FIG4 shows the operation of the present invention at different distances; FIG5A is a collimator lens group configuration according to the present invention; FIG5B is another collimator lens group configuration according to the present invention method; FIG. 6 is an example timing diagram according to the present invention; FIG. 7 is another example timing diagram according to the present invention; FIG. 8 is a flow chart of the resolution enhancement method according to the present invention; FIG. 9A, FIG. 9B and FIG. 9C are example timing diagrams of the present invention; FIG. 10A is a real environment image; FIG. 10B, FIG. 10C and FIG. 10D are sampling examples of FIG. 10A; and FIG. 11A, FIG. 11B, FIG. 11C, FIG. 11D, FIG. 11E and FIG. 11F are sampling situations of different subframes in the same frame.

The following is a more detailed description of the implementation of the present invention with the help of diagrams and component symbols, so that those who are familiar with the technology can implement it accordingly after reading this manual.

The present invention provides a lidar system with improved resolution and a method for improving the resolution of the lidar system. By projecting laser lights of multiple different wavelengths and rotating or oscillating the diffraction optical element, the resolution of the imaging can be increased while the light spot density remains unchanged.

Referring to FIG. 1 , the present invention provides a lidar system 100, including a microcontroller (MCU) 102, a laser light source 102, a lens module 106, and a receiver 112. The lens module 106 includes a receiver lens module 108 and a laser spectrometer module 110. The laser light source 102 and the receiver 112 are coupled to the microcontroller 101.

In order to measure the distance between the target 120 and the lidar system 100, first, the microcontroller 101 controls the laser light source 102 to emit laser light 104. Then, the laser spectroscope module 110 scatters the laser light 104 into a plurality of light spots, which are distributed in the field of image (FOI) 122, and the field of image 122 completely covers the target 120. Subsequently, after the light spots contact the target 120, they are reflected into a plurality of reflected lights 126, which are distributed in the field of view (FOV) 124. The receiver lens module 108 receives the reflected light 126 and transmits a signal to the receiver 112. The receiver 112 transmits the received signal to the microcontroller 101 for subsequent image analysis.

Please refer to FIG. 2A. The receiver lens module 108 in FIG. 1 includes a lens module composed of at least one concave lens 202 and at least one convex lens 204. The concave lens 202 and the convex lens 204 form a focusing lens group, which can focus the reflected light 126 in FIG. 1 so as to transmit the optical signal to the receiver 112. The laser spectroscope module 110 in FIG. 1 includes a diffractive optical element (DOE) 206, a concave mirror 208 and a collimation lens group (collimation lens) 210. Among them, the diffractive optical element 206 has the function of rotation or oscillation. The operation of the laser spectroscope module 110 will be described in detail below.

Please refer to FIG. 2B . The laser light source 102 in FIG. 1 can emit multiple (for example, but not limited to four) laser lights 221, 222, 223, and 224 of different wavelengths. The laser light source 102 includes an optical coupler 230 and an optical fiber 240. The laser lights 221, 222, 223, and 224 can be infrared lights, for example, infrared lights with wavelengths of 850nm, 905nm, 940nm, and 1064nm. The laser lights 221, 222, 223, and 224 can be emitted simultaneously or sequentially. The optical coupler 230 optically couples the laser lights 221, 222, 223, and 224 into a single collimated light signal, which is transmitted to the lens module via the optical fiber 240.

而暗紋位置y_n為：y _n =L tan θ _n =Lnλ/a因此，透過發射不同波長λ的雷射光，可產生具有不同繞射點間距的點雲。如圖2D所示，波長λ₁、λ₂及λ₃的雷射光產生的點雲具有不同的繞射點間距，故當波長λ₁、λ₂及λ₃的雷射光被光學耦合為單一光訊號射出時，由波長分別為λ₁、λ₂及λ₃的雷射光所產生的單狹縫繞射圖像具有不同的繞射點間距，而該等單狹縫繞射圖像可互相填補空隙，有效增加點雲密度，進而增進成像解析度。如圖2E所示，波長λ₁、λ₂及λ₃的雷射光產生的點雲會落在不同位置，當波長λ₁、λ₂及λ₃的雷射光被光學耦合為單一光訊號射出時，可造成該等點雲疊合，有效增加點雲密度，進而增進成像解析度。此一方法適合使用於動態偵測之情境。 According to the single-slit diffraction principle, the diffraction angle θ _n is related to the slit width a and the wavelength λ: a sinθ _n = nλ,n =±1,±2,±3... Please refer to Figure 2C. When the projection distance L is fixed, the dark pattern positions y ₁ ,y ₂ ,... and -y ₁ ,-y ₂ ,... are related to the diffraction angle θ _n . When the wavelength λ is in the infrared range (about the order of 1000nm) and the slit width a is in the order of mm, the diffraction angle θ _n is extremely small. At this time:

The dark pattern position _yn is: yn = _{Ltanθn =} Lnλ/aTherefore , by emitting laser light of different wavelengths λ, point clouds with different diversion point spacings can be generated. As shown in FIG2D, the point clouds generated by the laser light of wavelengths _λ1 , _λ2 and _λ3 have different diversion point spacings. Therefore, when the laser light of wavelengths _λ1 , _λ2 and _λ3 are optically coupled as a single light signal and emitted, the single slit diversion images generated by the laser light of wavelengths _λ1 , _λ2 and _λ3 have different diversion point spacings, and these single slit diversion images can fill the gaps with each other, effectively increasing the point cloud density and thus improving the imaging resolution. As shown in FIG2E , the point clouds generated by the laser lights of wavelengths λ ₁ , λ ₂ and λ ₃ fall at different positions. When the laser lights of wavelengths λ ₁ , λ ₂ and λ ₃ are optically coupled as a single light signal, the point clouds can be superimposed, effectively increasing the point cloud density and thus improving the imaging resolution. This method is suitable for use in dynamic detection scenarios.

Please refer to FIG. 3A . When laser light 302 is directed to diffraction optical element 304 , diffraction optical element 304 diffracts laser light 302 into thousands to tens of thousands of light spots. These light spots form point clouds 306a , 306b and 306c at different distances. Point cloud 306a is closest to diffraction optical element 304 , has the most dense light spots and covers the smallest area; while point cloud 306c is farthest from diffraction optical element 304 , has the least dense light spots and covers the largest area. Diffraction optical element 304 may be, for example, HCPDOE ^™ of Tyrafos Semiconductor Co., Ltd., but the present invention is not limited thereto.

Please refer to FIG. 3B. Through rotation or oscillation, the diffraction optical elements 310a and 310b can be two states of the diffraction optical element 304 at different times. As shown in FIG. 3B, through the rotation and/or reciprocating movement of the diffraction optical elements 310a and 310b, the light spot position of the diffraction optical element 310b can cover the gap between the light spots of the diffraction optical element 310a, and vice versa. Therefore, in each subframe of a plurality of subframes in a frame, the light spot projection position of the diffraction optical element 304 will change. By fusing a plurality of subframes, the light spot illumination area can be increased without increasing the light spot density, thereby improving the imaging resolution. This method is suitable for use in static detection scenarios.

Since the point cloud coverage area shown in FIG3A is proportional to the square of the distance, when the distance is far, the point cloud coverage area will expand rapidly, causing the light energy per unit area to decrease, thereby resulting in insufficient reflected light intensity. A significant increase in the intensity of the laser light 302 may result in a decrease in the life of the device and may easily cause damage to the human eye. Therefore, please refer to FIG4 , the focal length adjustable lens module composed of at least one concave lens 202 and at least one convex lens 204 can adjust the field of view size according to the detection range, so that the light energy per unit area at different distances (e.g., 15 meters, 40 meters, 100 meters, 200 meters and 300 meters) is roughly equal, to prevent the situation where the reflected light intensity is insufficient when the distance is far. Alternatively, a plurality of lens modules with fixed focal lengths may be used, each lens module including at least one concave lens 202 and at least one convex lens 204, and the lens modules may be switched according to the ranging range to adjust the field of view.

One way to achieve the configuration shown in FIG. 4 is to use a collimator set to converge the coverage area of the diffracted light within a certain range. By adjusting the focal length, the collimator set can adjust the divergence angle of the outgoing parallel light and adjust the image field range of the projected light spot according to the ranging range to achieve the effect shown in FIG. 4. A plurality of collimators with fixed focal lengths can be used, and the collimators can be switched according to the ranging range to adjust the image field range. Alternatively, a collimator set with a variable focal length can be used, and the collimators can be switched according to the ranging range to adjust the image field range. Please refer to FIG. 5A , a collimator lens set is configured to place the collimator lens set 502 directly in front of a rotatable or oscillating diffraction optical element 504, wherein the mirror surface of the collimator lens set 502 is perpendicular to the incident direction of the laser light 506. As shown in FIG. 5A , the collimator lens set 502 can converge the diffraction light 508 emitted by the diffraction optical element 504 to be substantially parallel to each other, so that the unit area light energy of the diffraction light 508 at different distances is substantially maintained equal. In one embodiment, the collimator lens set 502 includes a concave lens 5021 and a convex lens 5022, wherein the distance between the concave lens 5021 and the convex lens 5022 can be adjusted to control the divergence angle.

Referring to FIG. 5B , another collimator lens set configuration is to place the collimator lens set 512 in front of the concave mirror 520, so that the concave mirror 520 collects the diffracted light emitted by the rotatable or oscillating diffracting optical element 514. As shown in FIG. 5B , the diffracting optical element 514 diffracts the laser light 516 into a plurality of diffracted light beams 518, which are reflected toward the collimator lens set 512 through the concave mirror 520. Subsequently, the collimator lens set 512 converges the diffracted light beams 518 to be substantially parallel to each other, so that the light energy per unit area of the diffracted light beams 518 at different distances is substantially maintained equal. In one embodiment, the collimator lens assembly 512 includes a concave lens 5121 and a convex lens 5122, wherein the distance between the concave lens 5121 and the convex lens 5122 can be adjusted to control the divergence angle. Compared with the configuration shown in FIG. 5A, this configuration can collect diffracted light at a larger angle, thereby increasing the projected light energy per unit area without increasing the laser light intensity.

In the use scenario of vehicle automatic driving, when the vehicle is driving, the interference signals that the lidar system 100 may receive include scanning lasers from vehicles in the opposite lane ahead, front directional pulse lasers from vehicles in the opposite lane ahead, scanning lasers from vehicles in the same lane ahead, and rear directional pulse lasers from vehicles in the same lane ahead. Therefore, such interference signals must be eliminated by appropriate methods to accurately measure distance and maintain driving safety.

When the laser light source 102 of FIG. 1 emits a pulse signal, in order to eliminate interference signals, the microcontroller 101 can activate or deactivate the receiver 112 according to the ranging range, so that the receiver 112 only receives the reflected light signal within the ranging range. For example, if the object to be measured is 300 meters away, the time required from the laser light source 102 emitting a pulse signal to the receiver 112 receiving the reflected light signal is 2μs (R=ct/2, where R is the distance, c is the speed of light 3×10 ⁸ m/s, and t is time (seconds)). Therefore, the receiver 112 can be activated synchronously with the laser light source 102 within a cycle time, the sensing time is 2μs, and the rest of the time is turned off to prevent receiving interference signals. Referring to FIG. 6 , the laser source (TX) emits a pulse signal with a pulse width PW at a cycle time T. The receiver (RX) opens at a sensor shutter time SS within the cycle time T and closes at a reset time R, where T=SS+R. The sensor shutter time SS and the reset time R are determined according to the range. In one embodiment, when the range is 300 meters, the sensor shutter time SS is 2μs, the reset time R is 2μs, the cycle time T is 4μs, and the pulse width PW is 100ns. At this time, the receiver (RX) can receive reflected light signals within a range of 0 to 300 meters, and the theoretical frame rate (number of scans) can be as high as 1/T=2.5×10 ⁵ f/s.

Please refer to FIG7. In addition to the upper limit of the ranging range, the lower limit of the ranging range can also be set for the receiver (RX) by adjusting the sensor shutter time SS. In FIG7, the start time Ts of the sensor shutter time SS is determined according to the lower limit of the ranging range, and the end time Tl is determined according to the upper limit of the ranging range. In one embodiment, when the ranging range is 90 to 300 meters, the start time Ts is 600ns, the end time Tl is 2μs, the sensor shutter time SS is 1400ns, the reset time R is 2μs, the cycle time T is 4μs, and the pulse width PW is 100ns. At this time, the receiver (RX) can receive the reflected light signal within 90 to 300 meters, and the theoretical frame rate is 1/T=2.5×10 ⁵ f/s.

Please refer to FIG8 . In a frame of an environment image including a plurality of (at least three, for example, six) subframes, method 800 obtains the environment image of the plurality of subframes, compares the average distance values of the plurality of sampling blocks of each subframe, and the microcontroller fuses the distance values of the pixels of the environment image of the plurality of subframes into a final distance value of the frame. In step 802 , a laser light source sequentially emits a plurality of laser light signals to obtain the environment image of the plurality of subframes, wherein within a sensor shutter time of a subframe of a frame, a plurality of pixels of the receiver receive at least one reflected light signal of the laser lights of different wavelengths. In each subframe, a plurality of laser lights of different wavelengths can be emitted simultaneously and coupled into a single optical signal, or can be emitted in sequence. In step 804, the distance values represented by the reflected light signals (the distance value can be obtained according to the time of flight (ToF) of the reflected light) are used as the distance values of the pixels of the subframe. In the environment image including a plurality of sampling blocks, the average distance values of the plurality of sampling blocks of the subframes are batch compared (see Table 1, Table 2 and Table 3 below). In step 806, according to the results of the batch comparison in step 804, the microcontroller eliminates abnormal subframes and merges normal subframes into the final distance value of the frame. Among them, "fusion" can be performed by taking the average, superposition, selection or other methods.

Please refer to FIG. 9A and FIG. 9B. As mentioned above, in the case of dynamic detection, laser lights of different wavelengths can be emitted simultaneously or sequentially. FIG. 9A shows the situation where laser lights of different wavelengths are emitted simultaneously. In an example where a frame includes six subframes, in each subframe, the laser light source (TX) emits laser lights 901, 902, and 903 of different wavelengths simultaneously with a cycle time T, and the receiver (RX) is turned on within the sensor shutter time SS to receive the reflected light signals of the laser lights 901, 902, and 903, and the reflected light signals are sent to the microcontroller to calculate the fused sub-range value. In the example shown in FIG9A , the position of the reflected light signal of the fifth subframe in the sensor shutter time SS is different from that of the remaining subframes, so the microcontroller can eliminate the sub-distance value of the fifth subframe and then merge the sub-distance values of the remaining subframes to form the final distance value of the frame. FIG9B shows the situation where laser lights of different wavelengths are emitted in sequence. In an example where a frame includes six subframes, in each subframe, a laser light source (TX) sequentially emits laser lights 911, 912, and 913 of different wavelengths with a cycle time T, and a receiver (RX) is turned on within a sensor shutter time SS to receive reflected light signals of laser lights 911, 912, or 913, and the reflected light signals are sent to a microcontroller to calculate the sub-range values represented by the reflected light signals. In the first subframe, laser light 911 is emitted. In the second subframe, laser light 912 is emitted. In the third subframe, laser light 913 is emitted. In the fourth subframe, laser light 911 is emitted. In the fifth subframe, laser light 912 is emitted. In the sixth subframe, laser light 913 is emitted. In this way, the microcontroller can calculate the sub-range value of each subframe, wherein the first and fourth subframes are associated with laser light 911, the second and fifth subframes are associated with laser light 912, and the third and sixth subframes are associated with laser light 913. In the example shown in FIG9B , the position of the reflected light signal of the fifth subframe in the sensor shutter time SS is different from that of the remaining subframes (i.e., the flight time is different), so the average range value of the fifth subframe will be different from that of the remaining subframes, so the microcontroller can eliminate the sub-range value of the fifth subframe, and then merge the sub-range values of the remaining subframes to become the final range value of the frame.

Please refer to FIG. 9C . As described above, in the static detection scenario, by rotating or oscillating, in each of the multiple subframes of a frame, by setting the diffraction element 304 as a movable part with the function of rotation and/or reciprocating movement, the light spot projection position of the diffraction optical element 304 will change. In FIG. 9C , under the conditions of multiple rotation angles or reciprocating positions, multiple subframe environmental images are taken, and the light spot projection position of the laser light 921 will change in each subframe. Each pixel of each reflected light signal in the environmental image represents a sub-range value, and each subframe environmental image is composed of multiple sub-range values to form a three-dimensional image with depth information. After that, the microcontroller removes the abnormal subframes and fuses the remaining multiple subframes of the environment image. If the same pixel has multiple distance values, the distance values are averaged or selected. If the same pixel has only one distance value, the sub-distance value is selected. If the pixel has no distance value, the maximum value (e.g., 500 meters) or the minimum value (e.g., 0) in the distance measurement range is selected to calculate the final distance value of the three-dimensional image of the frame. For example, in one embodiment, the following Tables 1A to 1F represent the first to sixth subframes of the same sampling block in the same frame, respectively, wherein each small square represents a pixel, and the one with a value represents the sub-distance value measured by the pixel of the subframe, while the one without a value represents no sub-distance value. For each subframe, the average distance value of the pixels with sub-distance values is taken, and the abnormal subframes are eliminated according to the average distance value of each subframe. From Table 1A to Table 1F, it can be seen that the average distance value of the sixth subframe is obviously different from that of the other subframes, so the sixth subframe is regarded as an abnormal value and eliminated. Then, as shown in Table 1G, each pixel of the normal subframe (the first to the fifth subframe, i.e., Table 1A to Table 1E) is superimposed, and then the average value of each pixel with sub-distance value after superposition is taken as the final distance value of this sampling block in this frame. During superposition, if the same pixel has multiple sub-range values, the average or the maximum value is selected. If the pixel does not have a sub-range value, the minimum value in the ranging range (e.g., 0) is selected. Alternatively, if the same pixel has multiple sub-range values, the average or the minimum value is selected. If the pixel does not have a sub-range value, the maximum value in the ranging range (e.g., 500 or 1000) is selected. In the example shown in FIG. 9C , the sub-range value of the fifth sub-frame is different from that of the remaining sub-frames, so the microcontroller can eliminate the sub-range value of the fifth sub-frame, and then merge the remaining sub-frames to calculate the final range value of the frame.

Figure 10A is a real environment image. In order to improve the computing efficiency, it is not necessary to measure the distance of every pixel on the entire image, but several blocks can be sampled for distance measurement. Each sampling block includes a plurality of pixels, such as 10×10 pixels. The sampled pixels should not be too many, for example, not more than 10% of the total number of pixels, to improve the computing efficiency. Figure 10B shows an example of sampling two blocks. Figure 10C shows an example of sampling five blocks. Figure 10D shows an example of sampling nine blocks. The number of sampling blocks should not be less than five to better grasp the environmental information. In a normal subframe, if the sampling blocks with normal distance values exceed a certain ratio (e.g. 80% or 88.9%, where 80% means that when five blocks are sampled, one sampling block is allowed to have an abnormal distance value, and 88.9% means when nine blocks are sampled, one sampling block is allowed to have an abnormal distance value), otherwise it is considered an abnormal subframe.

Please refer to FIG. 11A to FIG. 11F. In an embodiment in which a frame includes six subframes, the six subframes are FIG. 11A, FIG. 11B, FIG. 11C, FIG. 11D, FIG. 11E and FIG. 11F in sequence. Among them, the second subframe (FIG. 11B) and the sixth subframe (FIG. 11F) have external light intrusion. In order to effectively filter out the interfered subframes, the distance values measured by each pixel in each sampling block in each subframe can be fused as the sub-distance value of the sampling block in the subframe, and then the six sub-distance values of the same sampling block in the six subframes are compared with each other to filter out abnormal values. In one embodiment, the method of filtering out abnormal values is to calculate the average value (μ), standard deviation (σ), upper threshold and lower threshold of the six sub-range values in the same sampling block within six sub-frames, wherein the upper threshold is the average value plus a number of standard deviations (μ+nσ), and the lower threshold is the average value minus a number of standard deviations (μ-nσ), wherein the size of n is determined according to experimental data and actual needs, and can be an integer or non-integer, such as (but not limited to) 1 or 1.5. In the embodiments shown in Tables 2, 3 and 4 below, n=1 is used as an example for explanation, but the present invention is not limited thereto. Then, the subframes with sub-distance values greater than the upper threshold or less than the lower threshold are eliminated, and the remaining subframes with similar sub-distance values are merged as the final distance value of the frame.

Tables 2, 3, and 4 show possible sensing results. In the example shown in Table 2, in the first subframe, there is no obstacle in front of sampling block A. At this time, the distance of sampling block A is considered to be the farthest distance (e.g., 500 meters). In the second subframe, there is external light intrusion into sampling block A, and in the sixth subframe, there is external light intrusion into sampling blocks A, B, C, D, and E. At this time, as shown in Table 2, the distance values of the second and sixth subframes of sampling block A are lower than the lower threshold value, so they should be considered as abnormal values and filtered out. The distance values of the sixth subframes of sampling blocks B, C, D, and E are all lower than the individual lower threshold values, so they should be considered as abnormal values and filtered out.

In the example shown in Table 3, in the fourth subframe, there is external light intrusion in sampling block A, and the measured distance is very close to the normal value, while in the sixth subframe, there is external light intrusion in sampling blocks A, B, C, D, and E. At this time, as shown in Table 3, the distance value of the fourth subframe of sampling block A is higher than the upper threshold value, and the distance value of the sixth subframe is lower than the lower threshold value, so it should be considered as an abnormal value and filtered out. The distance values of the sixth subframes of sampling blocks B, C, D, and E are all lower than the individual lower threshold values, so they should be considered as abnormal values and filtered out. Therefore, even though the distances measured by sampling block A in the fourth and sixth subframes are very close to normal values, the two subframes can still be correctly identified as abnormal values and filtered out.

In the example shown in Table 4, in the sixth subframe, external light intrudes into sampling blocks B, C, D, and E. At this time, as shown in Table 4, the distance values of the sixth subframe of sampling blocks B, C, D, and E are all lower than the individual lower threshold values, so they should be considered as abnormal values and filtered out.

The above is only used to explain the preferred embodiment of the present invention, and is not intended to limit the present invention in any form. Therefore, any modification or change of the present invention made under the same spirit of the invention should still be included in the scope of protection intended by the present invention.

100: LiDAR system

101:Microcontroller

102:Laser light source

104: Laser light

106: Lens module

108: Receiver lens module

110: Laser spectrometer module

112: Receiver

120: Target

122: Image field

124: Field of view

126: Reflected light

Claims (12)

A lidar system includes: a microcontroller; a laser light source coupled to the microcontroller; a lens module, and a receiver coupled to the microcontroller, wherein: the laser light source emits a plurality of laser lights of different wavelengths, and includes an optical coupler and an optical fiber, the optical coupler optically couples the laser lights into a collimated light signal, and transmits the collimated light signal through the optical fiber; the lens module includes a laser spectroscope module and a receiver lens module, the laser spectroscope module receives the laser light emitted from the laser light source, and diverts the laser light into a plurality of diverted lights, the diverted lights are emitted toward a target; the laser spectroscope module includes a diverting optical element and a collimating lens assembly; the receiver lens ... A reflected light signal is generated after the diffracted light contacts the target, and the reflected light signal is sent to the receiver; the laser light source emits a pulse signal in a cycle time; the microcontroller controls the receiver to open in a sensor shutter time and close in a reset time in each cycle time; within a sensor shutter time of a subframe of a frame, a plurality of pixels of the receiver receive at least one reflected light signal of the laser lights of different wavelengths, obtain environmental images of a plurality of subframes, and use the distance values represented by the reflected light signals as the distance values of the pixels of the subframe; the microcontroller obtains a final distance value of the frame according to the distance values of the pixels of the environmental images of the plurality of subframes. The lidar system as described in claim 1 further includes: in an environment image including a plurality of sampling blocks, comparing the average distance values of the plurality of sampling blocks of the subframes; Based on the result of the comparison, the microcontroller eliminates abnormal subframes and merges normal subframes into the final distance value of the frame. The lidar system as described in claim 1, wherein the diffraction optical element has the function of rotation or oscillation. The lidar system as described in claim 1, wherein the receiver lens module includes a focal length adjustable lens module composed of at least one concave lens and at least one convex lens, and the field of view size is adjusted according to the ranging range. The lidar system as described in claim 1, wherein the receiver lens module includes a plurality of lens modules with fixed focal lengths, each lens module includes at least one concave lens and at least one convex lens, and the lens modules are switched according to the ranging range to adjust the field of view. The lidar system as described in claim 1, wherein the laser spectrometer module includes the diffraction optical element and a plurality of collimating lens sets with fixed focal lengths, and the collimating lens sets are switched according to the ranging range to adjust the image field range. The lidar system as described in claim 1, wherein the laser spectrometer module includes the diffraction optical element and a collimator lens set with a variable focal length, and the collimator lens sets are switched according to a distance measurement range to adjust the image field range. As described in claim 6 or 7, the diffraction optical element diffracts the laser light into the diffraction lights, the collimator set is arranged directly in front of the diffraction optical element, and the mirror surface of the collimator set is perpendicular to the incident direction of the laser light, so as to converge the diffraction lights to be substantially parallel to each other. The lidar system as described in claim 6 or 7 further includes a concave mirror, the diffraction optical element diffracts the laser light into the diffracted light, the concave mirror collects the diffracted light, and the collimator set is arranged in front of the concave mirror to converge the diffracted light to be substantially parallel to each other. A lidar system as described in claim 1, wherein the sensor shutter time and the reset time are determined according to the ranging range. The lidar system as described in claim 10 further includes a start time and an end time, and the microcontroller controls the receiver to turn on between the start time and the end time in each cycle time, and turn off the receiver at other times; the start time is determined according to the lower limit of the ranging range, and the end time is determined according to the upper limit of the ranging range. A method for improving the resolution of a lidar system as claimed in claim 1, the method comprising: setting the diffraction optical element as a movable part with the function of rotation and/or reciprocating movement; taking a plurality of sub-frame environmental images under a plurality of rotation angles or reciprocating position conditions; each pixel of the reflected light signals in the environmental image represents a sub-range value, and each sub-frame environmental image is composed of a plurality of sub-range values. The distance values are combined to form a three-dimensional image with depth information; after the microcontroller eliminates abnormal sub-frames, the remaining multiple sub-frame environmental images are fused. If the same pixel has multiple sub-distance values, the values are averaged or selected. If the same pixel has only one sub-distance value, the sub-distance value is selected. If the pixel does not have a sub-distance value, the maximum value in the distance measurement range is selected to calculate the final distance value of the three-dimensional image of the frame.

Images