A 3D ranging system

Abstract

The application discloses a 3D ranging system, comprising: a line light source assembly for generating and collimating a line light beam in a first direction to irradiate a target scene; a scanning assembly for scanning the line light beam in the first direction to cover the target scene; a receiving assembly comprising a receiving lens and a planar array detector, the planar array detector being a SPAD planar array detector with a counting function, for receiving a light spot of the line light beam reflected by the target scene and counting photons during exposure; and a signal processing unit connected with the receiving assembly, for: after each frame exposure, extracting pixel position information and integral energy information of the light spot in the first direction based on counting data of the SPAD planar array detector; calculating distance data of each point of the target scene in a second direction according to the pixel position information by using a triangulation principle; obtaining a current scanning angle of the scanning assembly in the first direction; and combining the current scanning angle with the distance data to generate three-dimensional point cloud data of the target scene.

Description

Technical Field

[0001] This invention relates to the field of three-dimensional measurement equipment, and more specifically to a 3D ranging system based on a line light source and a SPAD area array detector. Background Technology

[0002] With the rapid development of 3D perception technology, lidar and 3D sensors based on the triangulation principle have been widely used in fields such as autonomous driving, robot navigation, and industrial inspection. These systems illuminate target objects with laser beams, capture the reflected light signals using receivers, calculate the target distance based on the changes in the position of the light spot on the detector, and then generate 3D point cloud data.

[0003] However, existing lidar or sensors based on the triangulation principle have significant technical limitations when performing 3D scanning. First, to acquire true 3D information, the system needs to capture images at extremely high frame rates, but ordinary CIS (CMOS Image Sensor) image sensors struggle to meet these requirements. For example, a 3D sensor with a vertical resolution of 480, a horizontal resolution of 640, and a frame rate of 10Hz requires a capture rate of 6.4kHz, resulting in an actual data rate as high as 6Gbps. This leads to extremely high power consumption and limited dynamic performance. Second, the pixel position accuracy of traditional area array detectors is limited, making it difficult to achieve high-precision ranging. Third, existing systems often fail to accurately identify and separate different target information when dealing with complex scenes involving multipath effects or multiple targets being illuminated simultaneously. Summary of the Invention

[0004] This application provides a 3D ranging system that combines a line light source with a specific divergence angle constraint with a SPAD area array detector with counting function, and integrates the collaborative processing of scanning components and signal processing units. This system can achieve high-precision three-dimensional perception while significantly reducing system data rate and power consumption, and improving real-time dynamic performance.

[0005] This application discloses a 3D ranging system, comprising: a line light source assembly for generating and collimating a line beam in a first direction to illuminate a target scene, wherein the far-field divergence angle of the line beam in the first direction is less than or equal to 1°, and the far-field divergence angle in a second direction is greater than or equal to 30°, and the first direction and the second direction are perpendicular to each other; a scanning assembly for scanning the line beam in the first direction to cover the target scene; a receiving assembly including a receiving lens and an area array detector, wherein the area array detector is a SPAD area array detector with counting function, for receiving the light spot of the line beam reflected back from the target scene and performing photon counting during exposure; and a signal processing unit connected to the receiving assembly for: after each frame exposure, extracting pixel position information and integrated energy information of the light spot in the first direction based on the counting data of the SPAD area array detector; calculating distance data of each point in the target scene in the second direction based on the pixel position information using the principle of triangulation; obtaining the current scanning angle of the scanning assembly in the first direction; and generating three-dimensional point cloud data of the target scene by combining the current scanning angle and the distance data.

[0006] According to the 3D ranging system of this application, by limiting the far-field divergence angle in the first direction (e.g., horizontal direction) to ≤1° and the divergence angle in the second direction (e.g., vertical direction) to ≥30° in the online light source component, the line beam has high directivity in the scanning dimension to ensure angular resolution, while having wide coverage in another dimension to illuminate multiple target points within the field of view of the target scene at once. The receiver uses a SPAD area array detector with counting function, which can directly count photon events, avoiding the extremely high data rate and power consumption problems caused by the output of full-frame data by traditional CIS image sensors in high-speed application scenarios. The signal processing unit only extracts the effective spot position and energy information, which can significantly reduce the data rate. By combining the triangulation principle with the scanning angle, the system can efficiently and accurately reconstruct the three-dimensional point cloud of the target scene.

[0007] In the 3D ranging system disclosed in this application, the SPAD array detector includes multiple detection units arranged in an array, and each detection unit includes a SPAD unit and a counter corresponding to the SPAD unit.

[0008] The 3D ranging system according to this application achieves pixel-level photoelectric conversion and on-site digitization by integrating SPAD and counter in each pixel unit, which significantly suppresses circuit noise in analog readout, improves the signal-to-noise ratio, and provides a high-quality, low-redundancy data foundation for subsequent row-level or on-chip signal processing.

[0009] In the 3D ranging system disclosed in this application, the counter is a two-stage counter structure, including: a first-stage counter for counting avalanche events generated by the SPAD unit during the exposure period; and a second-stage counter for receiving and storing the counting results from the first-stage counter during the exposure period, and being read out by the signal processing unit in turn during the exposure period.

[0010] According to the 3D ranging system of this application, a high degree of parallelism between exposure and readout is achieved through a two-stage counter design: the first-stage counter focuses on real-time counting, while the second-stage counter serves as a data buffer, allowing polling and readout of the previous batch of data to begin during exposure. This mechanism can effectively improve the sensor's frame rate (achieving ≥1kHz) and photon utilization, providing hardware support for real-time dynamic 3D perception.

[0011] In the 3D ranging system disclosed in this application, when the signal processing unit extracts the pixel position information of the light spot, it obtains the sub-pixel level center position of the light spot by fitting counting data spanning at least two pixels.

[0012] According to the 3D ranging system of this application, the light spot moderately covers adjacent pixels, and the intensity distribution data of multiple pixels is fitted by algorithms (such as centroid method and Gaussian fitting), which can break through the limitation of physical pixel size on resolution and achieve sub-pixel level spot center positioning, thereby greatly improving the accuracy of triangulation ranging.

[0013] In the 3D ranging system disclosed in this application, the signal processing unit is further configured to calculate the reflectivity information of a corresponding point in the target scene based on the integrated energy information and the distance data.

[0014] According to the 3D ranging system of this application, while acquiring distance information, the integrated energy data output by the same SPAD array is combined with distance for compensation and correction, enabling the simultaneous acquisition of a reflectivity map of the target scene. This ensures that the output 3D point cloud includes not only spatial coordinates but also material property information.

[0015] In the 3D ranging system disclosed in this application, the signal processing unit is further used to identify, based on the position and energy information of multiple light spots output by the SPAD array detector in the second direction, the situation where the same pulse of the line beam illuminates multiple targets at different distances.

[0016] The 3D ranging system according to this application utilizes the high temporal resolution and multi-pulse extraction capability of the SPAD detector to resolve multiple echo signals at the same angle, effectively solving the multipath effect problem in complex scenes, and can detect the contour information of partially occluded targets, thereby improving the system's perception robustness in complex environments.

[0017] In the 3D ranging system disclosed in this application, the scanning component is one of a rotating mirror, a tilting mirror, or a mechanical rotating scanning device that carries the line light source component and the receiving component.

[0018] According to the 3D ranging system of this application, the rotating mirror / swinging mirror scheme has a small moment of inertia and high reliability, making it suitable for high-speed scanning, while the mechanical rotation scheme can achieve 360° omnidirectional coverage.

[0019] In the 3D ranging system disclosed in this application, the receiving component further includes a narrowband filter whose transmission wavelength matches the wavelength of the line light source component.

[0020] According to the 3D ranging system of this application, the narrowband filter can effectively filter out most of the noise photons in ambient light such as sunlight and lamplight, significantly reducing the probability of false triggering of the SPAD detector in outdoor or strong background light environments, and improving the detection signal-to-noise ratio and dynamic range of the system under complex lighting conditions.

[0021] In the 3D ranging system disclosed in this application, the SPAD unit and the counter of the SPAD array detector are respectively fabricated on two wafers, and the two wafers are electrically connected by hybrid bonding.

[0022] According to the 3D ranging system of this application, by optimizing the fabrication of the photosensitive layer and the digital circuit layer at different process nodes, and then achieving high-density vertical interconnection through hybrid bonding, the fill factor and photosensitivity of the pixels can be significantly improved. At the same time, it provides space for integrating more complex digital processing functions, which is the basic process guarantee for realizing large-area array, high-frame-rate SPAD sensors.

[0023] In the 3D ranging system disclosed in this application, the frame rate of the SPAD area array detector is greater than or equal to 1kHz, and the frequency at which the scanning component completes the scanning of the entire field of view is greater than or equal to 3Hz.

[0024] According to the 3D ranging system of this application, the real-time dynamic performance of the system is limited by quantitative indicators: a detector frame rate of ≥1kHz ensures that a sufficient density of ranging points can be acquired during high-speed scanning to avoid motion artifacts; a full-field scanning frequency of ≥3Hz means that the system can output a three-dimensional point cloud flow of dynamic scenes in real time that is perceptible to the human eye, meeting the stringent requirements of data update rate for applications such as autonomous driving and real-time robot navigation.

[0025] This application proposes a 3D ranging system that deeply integrates a line light source constrained by a specific divergence angle, a SPAD area array detector with counting capabilities, a scanning component, and a signal processing unit to construct a high-efficiency, high-precision, and low-data-redundancy real-time 3D perception solution. According to the technical solution of this application, the full-frame data readout bottleneck caused by the use of a CIS image sensor in traditional triangulation methods is eliminated. Only effective sparse light spot information is extracted, maintaining high accuracy (thanks to sub-pixel positioning) while reducing the data rate by more than two orders of magnitude. The system achieves high frame rate exposure and readout parallelism through a two-stage counter architecture, ensures high sensitivity and large-area array integration through a hybrid bonding process, and enriches the dimensions of perceived information through reflectivity calculation and multi-echo recognition. This technology provides a compact, efficient, and reliable hardware solution for applications requiring high-precision real-time 3D perception, such as autonomous driving, robot navigation, and industrial 3D inspection. Attached Figure Description

[0026] Figure 1 This is a schematic diagram of the structure of one embodiment of the 3D ranging system involved in this application;

[0027] Figure 2 This is a circuit diagram of one embodiment of a single detection unit in the SPAD array detector involved in this application;

[0028] Figure 3 This is a schematic diagram of one embodiment of the SPAD area array detector array architecture involved in this application. Detailed Implementation

[0029] The present application will be further described below with reference to specific embodiments and accompanying drawings. It is to be understood that the illustrative embodiments of this disclosure are merely for explaining the present application and not for limiting it. Furthermore, for ease of description, the accompanying drawings show only the parts relevant to the present application, and not all of the structures or processes.

[0030] The following specific embodiments illustrate the implementation of this application. Those skilled in the art can easily understand other advantages and effects of this application from the content disclosed in this specification. Although the description of this application is presented in conjunction with preferred embodiments, this does not mean that the features of this invention are limited to this embodiment. On the contrary, the purpose of describing the invention in conjunction with embodiments is to cover other options or modifications that may be derived based on the claims of this application. To provide a thorough understanding of this application, many specific details will be included in the following description. This application may also be implemented without using these details. Furthermore, to avoid confusion or obscuring the focus of this application, some specific details will be omitted in the description. It should be noted that, unless otherwise specified, the embodiments and features in the embodiments of this application can be combined with each other.

[0031] Unless the context otherwise specifies, the terms “contains,” “has,” and “includes” are synonyms. The phrase “A / B” means “A or B.” The phrase “A and / or B” means “(A and B) or (A or B).”

[0032] It should be understood that although terms such as "first," "second," etc., may be used herein to describe various components, units, or data, these components, units, or data should not be limited by these terms. These terms are used merely to distinguish one feature from another. For example, without departing from the scope of the exemplary embodiments, a first feature may be referred to as a second feature, and similarly, a second feature may be referred to as a first feature.

[0033] It should be understood that although directional terms such as "up," "down," "left," and "right" may be used here to describe the positional relationship between the various components, these directional terms are only for the convenience of understanding and are not intended to limit the scope of protection of this application.

[0034] It should be noted that in this specification, similar reference numerals and letters in the accompanying drawings indicate similar items. Therefore, once an item is defined in one drawing, it does not need to be further defined and explained in subsequent drawings.

[0035] To make the objectives, technical solutions, and advantages of this application clearer, the embodiments of this application will be described in further detail below with reference to the accompanying drawings.

[0036] Figure 1 This is a schematic diagram of the structure of one embodiment of the 3D ranging system involved in this application. Figure 1 In the embodiment shown, the 3D ranging system includes a line light source assembly 100, a scanning assembly 200, a receiving assembly 300, and a signal processing unit 400.

[0037] A line light source assembly 100 is used to generate and collimate a line beam in a first direction (e.g., the horizontal direction) to illuminate a target scene. In this embodiment, the line light source assembly 100 includes a laser 110 and a collimating lens 120. The laser 110 may be a laser diode (LD) or a vertical-cavity surface-emitting laser (VCSEL). The collimating lens 120 is configured to shape the laser emitted by the laser 110 into a line beam such that the far-field divergence angle of the line beam in the first direction is less than or equal to 1°, and the far-field divergence angle in the second direction (e.g., the vertical direction) is greater than or equal to 30°, wherein the first direction and the second direction are perpendicular to each other. The collimating lens 120 may be a cylindrical lens. This beam design enables the line beam to have high directivity in the scanning dimension (horizontal direction) to ensure angular resolution, while having wide coverage in another dimension perpendicular to the scanning dimension (vertical direction) to illuminate multiple target points within the field of view of the target scene at once.

[0038] The scanning component 200 is used to scan the line beam in a first direction to cover the entire target scene. In this embodiment, the scanning component 200 is a rotating mirror. Specifically, the rotating mirror has multiple vertically arranged reflective surfaces, such as a three-sided mirror, a four-sided mirror, a five-sided mirror, a six-sided mirror, etc. The multiple reflective surfaces are driven by a motor to rotate around a rotation axis perpendicular to the horizontal direction, thereby realizing the scanning of the line beam in the horizontal direction. However, the scanning component of this application is not limited to a rotating mirror; the scanning component can also be a swing mirror, or a mechanical rotating scanning device that carries the line light source component 100 and the receiving component 300. Through the periodic movement of the scanning component 200, the line beam can traverse the entire field of view in the first direction, achieving complete coverage of the target scene.

[0039] The receiving component 300 is used to receive the light spot of the line beam reflected back from the target scene, and includes a receiving lens 310, a narrowband filter 320, and an area array detector 330. The transmission wavelength of the narrowband filter 320 is matched with the wavelength of the laser emitted by the laser 110 of the line light source component 100, so as to effectively filter out ambient light noise. The area array detector 330 is a single-photon avalanche diode (SPAD) area array detector with counting function, used to count the received photons during exposure.

[0040] The signal processing unit 400 is connected to the receiving component 300 and is used to process the counting data output by the SPAD area array detector 330. After each frame exposure, the signal processing unit 400 extracts the pixel position information and integrated energy information of the light spot in the first direction based on the counting data of the SPAD area array detector 330; according to the pixel position information, it calculates the distance data of each point in the target scene in the second direction using the principle of triangulation; at the same time, it obtains the current scanning angle of the scanning component 200 in the first direction; and combines the current scanning angle and distance data to generate three-dimensional point cloud data of the target scene.

[0041] Specifically, after each frame exposure, the signal processing unit 400 first reads the count data of each pixel unit in the SPAD area array detector 330. This data records the number of photons received by each pixel during the exposure, reflecting the intensity distribution of the echo spot formed by the line beam reflected back from the target scene. Since the optical system design makes the blur spot of the receiving lens 310 slightly larger than the size of the pixel unit of the SPAD area array detector, an echo spot usually covers multiple adjacent pixels. By fitting these count data spanning at least two pixels (e.g., using the centroid method, Gaussian fitting method, or linear interpolation algorithm), the signal processing unit 400 can accurately calculate the center position of the spot in the first direction, with a positioning accuracy reaching the sub-pixel level (i.e., subdivided by at least 10 times under the SPAD pixel size). At the same time, by accumulating the count values ​​of all pixels within the area covered by the spot, the integrated energy information corresponding to the spot is obtained. Subsequently, the signal processing unit 400 calculates the distance using the principle of triangulation based on the extracted pixel position information of the light spot in the first direction: based on the known baseline length (the fixed distance between the line light source component 100 and the receiving component 300) and the focal length of the receiving lens 310, combined with the offset of the light spot center in the first direction, it calculates the distance data of each corresponding point of the target scene in the second direction through similar triangle geometric relationships. Since the line beam has a divergence angle of greater than or equal to 30° in the second direction, one frame exposure can simultaneously cover multiple target points in the second direction. Therefore, the signal processing unit 400 can calculate the distance values ​​corresponding to all effective echoes within the profile in parallel. At the same time, the signal processing unit 400 also acquires the current scanning angle of the scanning component 200 in the first direction in real time. This angle information comes from the encoder or controller of the scanning component and accurately indicates the horizontal azimuth angle pointed to when the line beam is emitted in the current frame. Finally, the signal processing unit 400 combines the calculated distance data of each point in the second direction with the current scanning angle in the first direction, and generates 3D point cloud data of the target scene at the scanning position through coordinate transformation (that is, converting the distance and angle in the polar coordinate system to X, Y, Z coordinates in the Cartesian coordinate system). As the scanning component 200 moves continuously in the first direction, the system repeats the above process frame by frame, eventually accumulating to cover the entire field of view and forming a complete 3D point cloud image of the target scene.

[0042] The SPAD array detector 330 includes multiple detector elements arranged in an array. Figure 2 This is a circuit diagram illustrating one embodiment of a single detection unit in the SPAD array detector involved in this application. Figure 2As shown, a single detection unit in a SPAD area array detector includes a SPAD unit 331 and a counter 332. The cathode of the SPAD unit 331 is connected to the first port (Cathode) and receives the bias voltage, while its anode is connected to the second port (Anode). The two ends of the SPAD unit 331 are also coupled through a coupling network formed by a resistor Rq and a capacitor Cd, respectively, to couple the pulse signal generated by the avalanche event to the input terminal CntIn of the counter 332. The counter 332 is a 16-bit counter used to count the avalanche events generated by the SPAD unit 331 during exposure and outputs the counting result to adjacent units or readout circuits through the CntOut port. By integrating the SPAD and counter within each pixel unit, pixel-level photoelectric conversion and on-site digitization are achieved, providing a high-quality, low-redundancy data foundation for subsequent row-level or on-chip signal processing.

[0043] In this application, the counter can be a single-stage counter structure, meaning each unit corresponds to one counter. However, this application is not limited to this. In a preferred embodiment, the counter can also adopt a two-stage counter structure, including a first-stage counter and a second-stage counter. The first-stage counter is used to count avalanche events generated by the SPAD unit in real time during exposure. The second-stage counter is used to receive and store the counting results from the first-stage counter during exposure and is read out by the signal processing unit 400 in turn during exposure. This two-stage counter design achieves a high degree of parallelism between exposure and readout: the first-stage counter focuses on real-time counting, while the second-stage counter acts as a data buffer, allowing polling and reading out of the previous batch of data to begin during exposure, thereby effectively improving the sensor's frame rate and photon utilization.

[0044] In practical applications, taking a 3D sensor with a vertical resolution of 480, a horizontal resolution of 640, and a frame rate of 10Hz as an example, a traditional CIS image sensor needs to capture images at a rate of 6.4kHz, resulting in an actual data rate as high as 6Gbps. However, using the SPAD area array detector of this application, since the location of each bright spot requires a maximum of 2 bytes of storage, the total data rate is only 6.4K × 480 × 2 × 8 = 49Mbps. This significant reduction in data rate also means a substantial reduction in power consumption and a significant improvement in the dynamic performance of the 3D sensor.

[0045] Next, the spot position extraction algorithm will be explained. When the optical system design makes the blur spot of the receiving lens slightly larger than the size of a pixel (SPAD unit) of the SPAD array detector, a single spot will cover multiple adjacent pixels of the SPAD array detector. When extracting the pixel position information of the spot, the signal processing unit can obtain the sub-pixel level spot center position by fitting count data spanning at least two pixels. Specifically, the centroid method, Gaussian fitting method, or linear interpolation algorithm can be used to process the intensity distribution data of multiple pixels. For example, when the spot covers at least 2 pixels, a linear interpolation algorithm can be used to calculate the spot center; when higher precision is required, a multi-term fitting algorithm of ADC sampling points can be used to calculate the leading and trailing edges of the pulse, and the average value can be taken as the effective data. This sub-pixel positioning technology allows the pixel position accuracy to be subdivided by at least 10 times under the SPAD pixels, thereby significantly improving the accuracy of triangulation.

[0046] In the embodiments of this application, the signal processing unit is further configured to calculate the reflectivity information of a corresponding point in the target scene based on the integrated energy information and distance data. Specifically, the process by which the signal processing unit calculates the target reflectivity based on the integrated energy information and distance data is a physical inversion process that deduces the material properties of the target surface from the echo intensity. The integrated energy information output by the SPAD area array detector represents the cumulative photon count during exposure, and its value depends not only on the reflectivity of the target surface but also is significantly affected by the target distance. According to the inverse square law of distance in optics, under the same conditions, the farther the target distance, the more severe the attenuation of the echo energy with the square of the distance. In order to decouple the true reflectivity information from the original energy data, the signal processing unit uses the already calculated distance data to compensate and correct the integrated energy: specifically, the measured integrated energy value is multiplied by the square of the target distance (or more precise nonlinear compensation is performed according to the system calibration parameters), thereby eliminating the attenuation effect caused by the distance factor and obtaining a normalized energy value that is proportional to the target reflectivity. The reflectivity data is quantized and output with a bit width that can be set to 4 bits or more. This means that the reflectivity of each point can be distinguished into more than 16 gray levels. Through this processing, the final 3D point cloud output by the system not only contains the spatial coordinate information (X, Y, Z) of each point, but also includes reflectivity information that characterizes the surface material properties of the target. This upgrades the point cloud data from a simple geometric model to a "colored" point cloud containing semantic attributes, providing a rich data foundation for subsequent advanced applications such as target recognition, scene understanding, and material classification.

[0047] Furthermore, the signal processing unit can also identify situations where the same pulse of a line beam illuminates multiple targets at different distances based on the multiple spot positions and energy information output by the SPAD area array detector in the second direction. Specifically, this function effectively solves the multipath effect and partial occlusion problems in complex scenes. In practical applications, because the line beam has a large divergence angle (greater than or equal to 30°) in the second direction, a single laser pulse may simultaneously illuminate multiple targets at different distances along its propagation path. For example, part of the line beam may illuminate nearby leaves, while another part penetrates gaps to illuminate distant tree trunks, or the laser may simultaneously hit the front and back surfaces of a transparent object. Although the echo signals reflected from these targets at different distances arrive at different times, the SPAD area array detector, with its high temporal resolution and multi-pulse extraction capability, can capture these echo signals separately within the same frame exposure. By analyzing the multiple spot positions and corresponding integrated energy information output by the SPAD area array detector in the second direction, combined with time-correlated single-photon counting technology, the signal processing unit can extract multiple independent pulse peaks from the counting data of each pixel. Each peak corresponds to a different target distance, thus simultaneously outputting multiple distance values ​​at the same scanning angle. This multi-echo detection capability enables the system to penetrate some obstructions to obtain background information, effectively suppressing ranging errors caused by multipath effects, and significantly improving the system's perception robustness and data integrity in complex environments.

[0048] In one embodiment of this application, the SPAD cells and counter of the SPAD area array detector are fabricated on two different wafers, which are electrically connected by hybrid bonding technology. This heterogeneous integration method separates the high-sensitivity SPAD photosensitive layer from the high-density, high-performance digital circuit layer, fabricating them on an optimized process node. Hybrid bonding then achieves high-density, low-latency vertical interconnects, significantly improving the pixel fill factor and photosensitivity, and also providing space for integrating more complex DSP functions into the counter layer.

[0049] Figure 3 This is a schematic diagram of one embodiment of the SPAD area array detector array architecture involved in this application. Figure 3 As shown, the SPAD area array detector consists of a large number of detector cells, each of which is a pixel. The figure exemplifies the array size from row 0 to row 767, with each row containing 1024 detector cells (from (0,0) to (0,1023)), forming a high-resolution pixel array of 768×1024. This large-scale array design enables the system to achieve high-density spatial sampling, thereby acquiring fine distance distribution information of the target scene in a single exposure.

[0050] Regarding the data processing architecture, one embodiment of this application employs a row-level parallel processing strategy, where multiple rows of probe units share a single digital signal processing unit (DSP). For example, every four rows of probe units share one DSP. Specifically, the probe units in rows 0 to 3 output their counting data to DSP (0), rows 4 to 7 output to DSP (1), and so on, until rows 764 to 767 output to DSP (255), for a total of 256 DSP units. This grouped sharing design effectively controls the consumption of hardware resources while ensuring processing power.

[0051] Because the integration time of the SPAD area array detector is extremely short, the DSP unit is fully capable of completing the serial processing of all data within the frame interval. For example, within 100 microseconds, a single DSP unit can sequentially process the counting data of its four rows, totaling 4096 pixels, performing noise filtering, peak detection, and waveform fitting for each pixel. Each DSP unit packages the extracted effective spot position information and integrated energy information according to a predetermined data format, and then, through unified scheduling by the Timing Ctrl, outputs it to the signal processing unit via a high-speed serial interface of MIPI (Mobile Industry Processor Interface) or LVDS (Low-Voltage Differential Signaling). This pipeline architecture of "pixel-level acquisition, row-level preprocessing, and global serial output" enables the system to maintain a high frame rate while outputting only sparse effective data, avoiding the bandwidth pressure and power consumption issues caused by full-frame data readout.

[0052] In one embodiment of this application, the frame rate of the SPAD area array detector is greater than or equal to 1 kHz, and the scanning component completes the entire field of view scan at a frequency greater than or equal to 3 Hz. A detector frame rate greater than or equal to 1 kHz ensures that a sufficient density of ranging points can be acquired during high-speed scanning, thus avoiding motion artifacts; a full-field scanning frequency greater than or equal to 3 Hz means that the system can output a 3D point cloud stream of a dynamic scene in real time, perceptible to the human eye, meeting the stringent requirements for data update rates in applications such as autonomous driving and real-time robot navigation.

[0053] It should be noted that in the embodiments described above, the first direction is the horizontal direction and the second direction is the vertical direction, but this application is not limited to this. In the technical solution involved in this application, the first direction can be defined as the direction parallel to the line connecting the center of the light source (laser) and the center of the area array detector. The collimation direction of the line beam, the scanning direction of the scanning component, the direction of extracting the pixel position information and integrated energy information of the light spot are all in the first direction, which is also the direction of the counter displacement of the area array detector. The second direction is the direction of line beam divergence. The DSP in the area array detector first processes the data in the first direction, and then sequentially extracts and processes all the independent data in the second direction.

[0054] It should be further pointed out that, in Figure 2 In the illustrated embodiment, each SPAD unit is connected to a counter, meaning there is a one-to-one correspondence between units and counters. However, this application is not limited to this. In other embodiments of this application, multiple SPAD units may share a single counter. Therefore, the correspondence between SPAD units and counters in this application can be either one-to-one or many-to-one. Figure 3 Taking the SPAD area array detector architecture shown as an example, each detector cell can include a dedicated counter, but it can also be that two or more adjacent detector cells in the second direction share a single counter. For example, Cell (0,0) and Cell (1,0) share a counter, Cell (0,1) and Cell (1,1) share a counter, Cell (0,2) and Cell (1,2) share a counter, and so on, that is, two rows of detector cells share one row of counters. Of course, it is also possible for three or more rows of detector cells to share one row of counters.

[0055] In actual operation, while the line light source component emits light, the SPAD area array detector begins exposure and counting. After exposure, the signal processing unit calculates the spot position and the integrated energy within the spot, and transmits all position and energy data in the first direction. During each frame of pulse emission and exposure, the signal processing unit can acquire the angle of the current line light source component in the first direction, and combine it with the calculated distance data of each angle distributed in the second direction to generate three-dimensional point cloud data in polar coordinates. The counter array of the SPAD area array detector can be time-division multiplexed in the first and / or second directions. After multiple exposures, the counts of all SPAD units can be polled in multiple exposures within each frame.

[0056] This application proposes a 3D ranging system that deeply integrates a line light source constrained by a specific divergence angle, a SPAD area array detector with counting capabilities, a scanning component, and a signal processing unit to construct a high-efficiency, high-precision, and low-data-redundancy real-time 3D perception solution. According to the technical solution of this application, the full-frame data readout bottleneck caused by the use of a CIS image sensor in traditional triangulation methods is eliminated. Only effective sparse light spot information is extracted, maintaining high accuracy (thanks to sub-pixel positioning) while reducing the data rate by more than two orders of magnitude. The system achieves high frame rate exposure and readout parallelism through a two-stage counter architecture, ensures high sensitivity and large-area array integration through a hybrid bonding process, and enriches the dimensions of perceived information through reflectivity calculation and multi-echo recognition. This technology provides a compact, efficient, and reliable hardware solution for applications requiring high-precision real-time 3D perception, such as autonomous driving, robot navigation, and industrial 3D inspection.

[0057] The above are merely specific embodiments of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Where there is no conflict, the embodiments and features described in the embodiments of this application can be combined with each other. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A 3D ranging system, characterized in that, include: A line light source assembly for generating and collimating a line beam in a first direction to illuminate a target scene, the line beam having a far-field divergence angle of less than or equal to 1° in the first direction and a far-field divergence angle of greater than or equal to 30° in a second direction, the first direction and the second direction being perpendicular to each other. A scanning component is used to scan the line beam in a first direction to cover the target scene; The receiving component includes a receiving lens and an area array detector, wherein the area array detector is a SPAD area array detector with counting function, used to receive the light spot of the line beam reflected back by the target scene, and to count photons during exposure; The signal processing unit, connected to the receiving component, is used for: After each frame exposure, based on the counting data of the SPAD area array detector, the pixel position information and integrated energy information of the light spot in the first direction are extracted; Based on the pixel position information, the distance data of each point in the target scene in the second direction is calculated using the principle of triangulation. Obtain the current scanning angle of the scanning component in the first direction; By combining the current scanning angle and the distance data, three-dimensional point cloud data of the target scene is generated.

2. The 3D ranging system according to claim 1, characterized in that, The SPAD array detector includes multiple detection units arranged in an array, and each detection unit includes a SPAD unit and a counter corresponding to the SPAD unit.

3. The 3D ranging system according to claim 2, characterized in that, The counter is a two-stage counter structure, including: A first-stage counter is used to count avalanche events generated by the SPAD unit during the exposure period; The second-stage counter is used to receive and store the counting results from the first-stage counter during the exposure period, and is read out by the signal processing unit in turn during the exposure period.

4. The 3D ranging system according to claim 1, characterized in that, When the signal processing unit extracts the pixel position information of the light spot, it obtains the sub-pixel level center position of the light spot by fitting count data spanning at least two pixels.

5. The 3D ranging system according to claim 1, characterized in that, The signal processing unit is also used to calculate the reflectivity information of the corresponding point in the target scene based on the integrated energy information and the distance data.

6. The 3D ranging system according to claim 1, characterized in that, The signal processing unit is also used to identify, based on the position and energy information of multiple light spots output by the SPAD area array detector in the direction, the situation where the same pulse of the line beam illuminates multiple targets at different distances.

7. The 3D ranging system according to claim 1, characterized in that, The scanning component is one of a rotating mirror, a tilting mirror, or a mechanical rotating scanning device that carries the line light source component and the receiving component.

8. The 3D ranging system according to claim 1, characterized in that, The receiving component also includes a narrowband filter whose transmitted wavelength matches the wavelength of the line light source component.

9. The 3D ranging system according to claim 2, characterized in that, The SPAD unit and the counter of the SPAD array detector are respectively fabricated on two wafers, and the two wafers are electrically connected by hybrid bonding.

10. The 3D ranging system according to any one of claims 1 to 9, characterized in that, The frame rate of the SPAD array detector is greater than or equal to 1 kHz, and the frequency at which the scanning component completes the entire field of view scanning is greater than or equal to 3 Hz.

Images