An unmanned vehicle environment perception method and system under a low-speed scenario

Abstract

The application discloses a low-speed scene unmanned vehicle environment sensing method and system, the method comprises the following steps: collecting environment image data by using a monocular camera, collecting distance data by using a wireless device installed around the unmanned vehicle; carrying out gray scale conversion on the image data to obtain a gray scale image, using a fixed threshold method to carry out threshold segmentation on the gray scale image to obtain foreground pixels and background pixels; carrying out sliding average filtering processing on the distance data, establishing a mapping relationship between a device coordinate system and a vehicle body coordinate system based on the spatial coordinates of the wireless device, and converting the distance data in the vehicle body coordinate system; fusing the preprocessed image data and the distance data, extracting an obstacle image area existing within a preset safety threshold, obtaining a foreground proportion in the image area, and outputting a judgment result of a dynamic obstacle according to the foreground proportion. The application meets the basic sensing needs of low-speed driving of the unmanned vehicle by using a simplified sensor fusion logic and an obstacle judgment rule, and ensures driving safety.

Description

Technical Field

[0001] This invention relates to the field of environmental perception technology for unmanned vehicles, specifically to a method and system for environmental perception of unmanned vehicles in low-speed scenarios. Background Technology

[0002] With the development of autonomous driving technology, environmental perception serves as the foundation for autonomous vehicle decision-making and control, and its performance directly impacts the driving safety of autonomous vehicles. Current mainstream environmental perception solutions for autonomous vehicles mostly employ multi-sensor fusion technologies such as LiDAR, high-definition cameras, and millimeter-wave radar, combined with complex deep learning algorithms to achieve the detection, recognition, and tracking of environmental targets.

[0003] However, existing technologies have the following shortcomings: on the one hand, high-precision sensors such as lidar are expensive, which is not conducive to the large-scale application of low-speed unmanned vehicles (such as park shuttle buses and warehouse logistics vehicles); on the other hand, complex deep learning algorithms have high hardware computing power requirements and need to be equipped with high-performance computing platforms, which further increases system costs and energy consumption, and the algorithm debugging is difficult, which is not conducive to the implementation and promotion of the technology.

[0004] For autonomous vehicles operating in low-speed scenarios (such as parks, warehouses, and closed factory areas), where the environment is less complex, the driving speed is slow, and the types of obstacles are relatively simple (such as pedestrians, fixed facilities, and other low-speed vehicles), there is no need to adopt high-precision and high-complexity perception solutions. Summary of the Invention

[0005] Based on this, the purpose of this invention is to provide an environmental perception method and system for unmanned vehicles in low-speed scenarios, aiming to solve the current driving needs of unmanned vehicles in low-speed scenarios, where the environmental complexity is low, the driving speed is slow, the obstacle types are relatively simple, and there is no need to adopt high precision and high complexity.

[0006] To achieve the above objectives, this invention proposes a method for environmental perception of autonomous vehicles in low-speed scenarios, the method comprising: The vehicle uses monocular cameras at the front and sides to collect environmental image data and wireless devices installed around the vehicle to collect distance data. The image data is converted to grayscale to obtain a grayscale image, and the grayscale image is segmented using a fixed threshold method to obtain foreground pixels and background pixels; The distance data is processed by moving average filtering, and a mapping relationship between the device coordinate system and the vehicle body coordinate system is established based on the spatial coordinates of the wireless device, so that the distance data is transformed into the vehicle body coordinate system; By fusing the preprocessed image data and the distance data, the image region containing obstacles within a preset safety threshold is extracted, and the foreground ratio within the image region is obtained. Based on the foreground ratio, the judgment result of the dynamic obstacle is output.

[0007] According to one aspect of the above technical solution, in the step of converting the image data to grayscale, the color image is converted into a radian image, and the grayscale value is calculated using a weighted average method: Gray = 0.3R + 0.59G + 0.11B ± 2 Where R, G, and B are the red, green, and blue channel pixel values ​​of the corresponding pixel points in the color image, respectively (each ranging from 0 to 255). The ±2 in the formula is a compensation coefficient used to offset the color deviation of the sensor itself.

[0008] According to one aspect of the above technical solution, in the step of thresholding the grayscale image using a fixed threshold method, the fixed threshold adopts an illumination adaptive dynamic adjustment mechanism, and the threshold adjustment is achieved by combining scene illumination intensity feedback: The average grayscale value of each frame of grayscale image is extracted in real time, the average grayscale value of consecutive preset number of frames is calculated, and the average grayscale value is used as an evaluation index of the current scene illumination intensity. When the proportion of foreground pixels in consecutive preset frame images is greater than the first preset percentage, and the average grayscale value of the image is greater than the first average, the current scene is determined to be a strong lighting environment, and the fixed threshold is adjusted upward within the preset threshold range. When the proportion of foreground pixels in consecutive preset frame images is less than the second preset percentage, and the average grayscale value of the images is less than the second average, the current scene is determined to be a low-light environment, and the fixed threshold is adjusted downward within the preset threshold range. The adjusted fixed threshold is limited to a limit threshold range, which is greater than the preset threshold range.

[0009] According to one aspect of the above technical solution, the step of fusing the preprocessed image data and the distance data to extract the obstacle image region within a preset safety threshold is as follows: After fusing the image data and the distance data, the range of distances in front of the vehicle and / or the range of distances to the sides of the vehicle corresponding to different areas in the monocular camera image are determined based on the vehicle coordinate system. When a wireless device detects an obstacle within a preset safety threshold, it extracts the image region corresponding to the direction in which the obstacle exists.

[0010] According to one aspect of the above technical solution, after extracting the corresponding image region, the proportion of foreground pixels within the image region is determined. If the proportion of foreground pixels is greater than the foreground value, then there is an obstacle in the image area; if the proportion of foreground pixels is not greater than the foreground value, then the wireless device is determined to be a false detection and the image data is ignored. The spatial coordinates of the obstacle and a preliminary judgment on whether it is a dynamic obstacle are output to the autonomous vehicle decision system.

[0011] According to one aspect of the above technical solution, the specific steps for the preliminary determination of whether it is a dynamic obstacle are as follows: Linear interpolation smoothing is performed on the distance data of the same obstacle within a continuous preset acquisition period to eliminate the influence of random noise on the calculation of changes. Calculate the distance change and average rate of change between adjacent acquisition cycles, compare the calculation results with the preset judgment values ​​of the unmanned vehicle decision system, and output the obstacle dynamic attribute category. The confidence level of the dynamic attribute category judgment result is evaluated. If the confidence level is less than the confidence threshold, the judgment is re-evaluated after adding observation data for one more collection period.

[0012] This invention also proposes an environmental perception system for autonomous vehicles in low-speed scenarios. This system is used to implement the aforementioned environmental perception method for autonomous vehicles in low-speed scenarios. The system includes: The data acquisition module is used to acquire environmental image data using monocular cameras at the front and sides of the unmanned vehicle, and to acquire distance data through wireless devices installed around the unmanned vehicle. The segmentation module is used to convert the image data to grayscale to obtain a grayscale image, and to perform threshold segmentation on the grayscale image using a fixed threshold method to obtain foreground pixels and background pixels; The conversion module is used to perform moving average filtering on the distance data, and establish a mapping relationship between the device coordinate system and the vehicle body coordinate system based on the spatial coordinates of the wireless device, so as to convert the distance data into the vehicle body coordinate system. The judgment module is used to fuse the preprocessed image data and the distance data, extract the image region of the obstacle within the preset safety threshold, obtain the foreground ratio in the image region, and output the judgment result of the dynamic obstacle based on the foreground ratio.

[0013] The present invention also proposes a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the above-described method for environmental perception of unmanned vehicles in low-speed scenarios.

[0014] The present invention also proposes an electronic device, including a memory, a processor, and a computer program stored in the memory and running on the processor, wherein the processor executes the computer program to implement the above-described method for environmental perception of unmanned vehicles in low-speed scenarios.

[0015] In summary, the environmental perception method for autonomous vehicles in low-speed scenarios proposed in this invention significantly reduces costs compared to high-precision sensors such as LiDAR by employing a monocular camera and wireless devices as perception sensors. Simultaneously, the algorithm logic is simple, requiring low hardware computing power and eliminating the need for a high-performance computing platform, further reducing the overall system cost. The algorithm steps are concise, including data acquisition, simple preprocessing, data fusion, and obstacle detection, eliminating the need for complex deep learning model training and parameter tuning, thus reducing development difficulty and facilitating rapid deployment. This invention addresses the characteristics of low environmental complexity and limited obstacle types in low-speed scenarios, and through simplified sensor fusion logic and obstacle detection rules, it can meet the basic perception needs of autonomous vehicles driving at low speeds, ensuring driving safety.

[0016] Additional aspects and advantages of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description

[0017] Figure 1 This is a flowchart of the environmental perception method for unmanned vehicles in low-speed scenarios in Embodiment 1 of the present invention; Figure 2 This is a schematic diagram of the structure of the unmanned vehicle environment perception system in a low-speed scenario in Embodiment 2 of the present invention. Detailed Implementation

[0018] To make the objectives, features, and advantages of the present invention more apparent and understandable, specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Several embodiments of the present invention are shown in the drawings. However, the present invention can be implemented in many different forms and is not limited to the embodiments described herein. Rather, these embodiments are provided so that the disclosure of the present invention will be more thorough and complete.

[0019] It should be noted that when an element is referred to as being "fixed to" another element, it can be directly on the other element or there may be an intervening element. When an element is considered to be "connected" to another element, it can be directly connected to the other element or there may be an intervening element. The terms "vertical," "horizontal," "left," "right," "upper," "lower," and similar expressions used herein are for illustrative purposes only and are not intended to indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as limiting the invention.

[0020] In this invention, unless otherwise explicitly specified and limited, the terms "installation," "connection," "linking," and "fixing," etc., should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal communication between two components. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances. The term "and / or" as used herein includes any and all combinations of one or more of the related listed items.

[0021] Example 1 like Figure 1 The diagram shows a flowchart of an environmental perception method for an autonomous vehicle in a low-speed scenario according to Embodiment 1 of the present invention. The environmental perception method for an autonomous vehicle in a low-speed scenario includes the following steps S01-S04, wherein: S01. Use monocular cameras at the front and sides of the driverless vehicle to collect environmental image data, and collect distance data through wireless devices installed around the driverless vehicle. S02. Convert the image data to grayscale to obtain a grayscale image, and use a fixed threshold method to perform threshold segmentation on the grayscale image to obtain foreground pixels and background pixels; S03. Perform a moving average filtering process on the distance data, and establish a mapping relationship between the device coordinate system and the vehicle body coordinate system based on the spatial coordinates of the wireless device, so as to convert the distance data into the vehicle body coordinate system. S04. Fuse the preprocessed image data and the distance data, extract the image region containing obstacles within a preset safety threshold, obtain the foreground ratio within the image region, and output the judgment result of dynamic obstacles based on the foreground ratio.

[0022] Environmental image data is collected by monocular cameras installed at the front and sides of the unmanned vehicle, and distance data is collected by ultrasonic radars installed around the unmanned vehicle. The wireless device selected in this embodiment can be ultrasonic radar. The monocular camera uses a CMOS image sensor with a fixed frame rate of 15-20fps, supports dynamic adjustment of exposure time according to the scene's light intensity (exposure time range of 10ms-100ms), and the image resolution is set to 640×480 with a pixel size of 1.4μm×1.4μm to ensure clear obstacle images can be obtained under different lighting conditions. The ultrasonic radar uses a piezoelectric ceramic transducer with a center frequency of 40kHz. The acquisition frequency of a single radar is set to 10-15Hz, the detection range covers 0.1-5m, the detection angle is 15°, and the ranging accuracy is ±1cm. Through the staggered deployment of 8 radars, 360° near-range obstacle detection without blind spots is achieved around the unmanned vehicle, and the overlap rate of the detection range of adjacent radars is ≥10%, avoiding detection blind spots.

[0023] Images captured by a monocular camera are converted to grayscale and segmented using a thresholding method. First, the color image is converted to grayscale, and then the grayscale value is calculated using a weighted average method. Gray = 0.3R + 0.59G + 0.11B ± 2 Where R, G, and B are the red, green, and blue channel pixel values ​​of the corresponding pixels in the color image, respectively (each ranging from 0 to 255). The ±2 in the formula is a compensation coefficient used to offset color deviations inherent in the sensor itself. This weighting coefficient is set based on the CIE 1931 standard colorimetric system, conforming to the sensitivity characteristics of the human eye to red, green, and blue light. It can maximally preserve the edge contours and texture features of obstacles in the original image, providing high-quality grayscale image data for subsequent thresholding and region of interest extraction.

[0024] After being converted to a grayscale image, the grayscale image is segmented using a fixed threshold method to obtain foreground and background pixels. The fixed threshold employs an adaptive dynamic adjustment mechanism based on illumination, combined with scene illumination intensity feedback to achieve threshold adjustment. The average grayscale value of each frame of grayscale image is extracted in real time, and the average grayscale value of a consecutive preset number of frames is calculated. The average grayscale value is used as an evaluation index of the current scene illumination intensity. In this embodiment, the consecutive preset number of frames can be 10 frames. When the proportion of foreground pixels in consecutive preset frame images is greater than the first preset percentage, and the average gray value of the image is greater than the first average, the current scene is determined to be a strong lighting environment. At this time, the fixed threshold will fluctuate upward by 5-10 units within the preset threshold range. Typically, the first preset percentage is set to 60%, the first average is set to 180, and the preset threshold range is (120-150). When the proportion of foreground pixels in a series of preset frames (10 frames) is less than the second preset percentage (10%), and the average gray value of the image is less than the second average (80), the current scene is determined to be a low-light environment. At this time, the fixed threshold will fluctuate downward by 5-10 units within the preset threshold range. The adjusted fixed threshold is limited to the extreme threshold range (110-160) to avoid excessive threshold adjustment leading to segmentation failure. The threshold adjustment adopts a smooth transition strategy, with each adjustment not exceeding 5 units to ensure the stability of the segmentation results. The extreme threshold range is greater than the preset threshold range.

[0025] The distance data collected by the ultrasonic radar is processed by moving average filtering, and the average value is calculated by taking the distance data of 3-5 consecutive collection cycles. At the same time, according to the installation position of the ultrasonic radar, a mapping relationship between the radar coordinate system and the vehicle body coordinate system is established, and the distance data collected by the radar is transformed into the vehicle body coordinate system.

[0026] The preprocessed image data and ultrasonic radar data are fused together. Based on the vehicle coordinate system, the distance range in front of / to the side of the vehicle corresponding to different areas in the monocular camera image is determined. When the ultrasonic radar detects an obstacle in a certain direction with a distance less than a preset safety threshold, the image area corresponding to that direction is extracted.

[0027] It should be noted that the preset safety threshold adopts a two-factor adaptive adjustment strategy based on driving speed and scene type, specifically as follows: Based on the vehicle-mounted map and positioning information of the unmanned vehicle, the low-speed scene is divided into three types: open road section, narrow passage and intersection. Real-time driving speed is obtained by the vehicle's onboard wheel speed sensors, and the speed is divided into three levels: low speed (≤3km / h), medium-low speed (3km / h < speed ≤5km / h), and medium speed (5km / h < speed ≤10km / h). For different combinations of scene types and speed levels, corresponding safety thresholds are preset: in open road sections, the safety threshold for low speed is 0.8m, for medium-low speed is 1m, and for medium speed is 2.5m; in narrow passages, the safety threshold for low speed is 1m, for medium-low speed is 1.2m, and for medium speed is 3m; at intersections, the safety threshold for low speed is 1.2m, for medium-low speed is 1.5m, and for medium speed is 3m. The safety thresholds are dynamically updated in real time according to changes in driving speed and scene type.

[0028] If the proportion of foreground pixels in the image region is greater than the foreground value (30%), then there is an obstacle in the image region. If the proportion of foreground pixels is not greater than the foreground value, then the wireless device is judged to be a false detection and the image data is ignored. The spatial coordinates of the obstacle and the preliminary judgment of whether it is a dynamic obstacle are output to the unmanned vehicle decision system.

[0029] The specific process for the preliminary assessment of dynamic obstacles is as follows: The distance data of the same obstacle within a continuous preset acquisition period is smoothed by linear interpolation to eliminate the influence of random noise on the calculation of the change. In this embodiment, three acquisition periods are used as the preset acquisition period for illustration.

[0030] Calculate the distance change Δd1 (difference between the 2nd and 1st cycles) and Δd2 (difference between the 3rd and 2nd cycles) between adjacent acquisition cycles, and the average rate of change v = (|Δd1| + |Δd2|) / (2T), where T is the acquisition cycle; Based on the calculated distance change and average rate of change, determine the dynamic attributes of the current obstacle. Level 1 dynamic obstacle (high-speed movement): cumulative distance change |Δd1 + Δd2| > 0.8m, or average rate of change v > 0.3m / s; Level 2 dynamic obstacle (medium-speed movement): 0.5m < |Δd1 + Δd2| ≤ 0.8m, and 0.15m / s < v ≤ 0.3m / s; Static obstacle: |Δd1 + Δd2| ≤ 0.5m, and v ≤ 0.15m / s. The confidence level of the dynamic attribute judgment result is evaluated. The confidence level is calculated as 1 - (distance measurement error / cumulative distance change). When the confidence level is ≥0.7, the judgment result is output. When the confidence level is <0.7, the judgment is re-evaluated after adding one more collection cycle of observation data.

[0031] In summary, the environmental perception method for autonomous vehicles in low-speed scenarios proposed in this invention significantly reduces costs compared to high-precision sensors such as LiDAR by employing a monocular camera and wireless devices as perception sensors. Simultaneously, the algorithm's simple logic and low hardware computing power requirements eliminate the need for a high-performance computing platform, further reducing the overall system cost. The algorithm's steps are concise, including data acquisition, simple preprocessing, data fusion, and obstacle detection, requiring no complex deep learning model training and parameter tuning, thus simplifying development and facilitating rapid deployment. Addressing the low environmental complexity and limited obstacle types in low-speed scenarios, the simplified sensor fusion logic and obstacle detection rules effectively meet the basic perception needs of autonomous vehicles at low speeds, ensuring driving safety.

[0032] Example 2 Another aspect of this invention provides an environmental perception system for unmanned vehicles in low-speed scenarios; please refer to [link / reference needed]. Figure 2 The diagram shown is a structural schematic of the unmanned vehicle environmental perception system in a low-speed scenario according to Embodiment 2 of the present invention. The unmanned vehicle environmental perception system in a low-speed scenario includes: The acquisition module 11 is used to acquire environmental image data using monocular cameras at the front and sides of the unmanned vehicle, and to acquire distance data through wireless devices installed around the unmanned vehicle. The segmentation module 12 is used to convert the image data to grayscale to obtain a grayscale image, and to perform threshold segmentation on the grayscale image using a fixed threshold method to obtain foreground pixels and background pixels; The conversion module 13 is used to perform moving average filtering on the distance data, and establish a mapping relationship between the device coordinate system and the vehicle body coordinate system based on the spatial coordinates of the wireless device, so as to convert the distance data into the vehicle body coordinate system. The judgment module 14 is used to fuse the preprocessed image data and the distance data, extract the image region of the obstacle within the preset safety threshold, obtain the foreground ratio in the image region, and output the judgment result of the dynamic obstacle based on the foreground ratio.

[0033] Environmental image data is collected by monocular cameras installed at the front and sides of the unmanned vehicle, and distance data is collected by ultrasonic radars installed around the unmanned vehicle. The wireless device selected in this embodiment can be ultrasonic radar. The monocular camera uses a CMOS image sensor with a fixed frame rate of 15-20fps, supports dynamic adjustment of exposure time according to the scene's light intensity (exposure time range of 10ms-100ms), and the image resolution is set to 640×480 with a pixel size of 1.4μm×1.4μm to ensure clear obstacle images can be obtained under different lighting conditions. The ultrasonic radar uses a piezoelectric ceramic transducer with a center frequency of 40kHz. The acquisition frequency of a single radar is set to 10-15Hz, the detection range covers 0.1-5m, the detection angle is 15°, and the ranging accuracy is ±1cm. Through the staggered deployment of 8 radars, 360° near-range obstacle detection without blind spots is achieved around the unmanned vehicle, and the overlap rate of the detection range of adjacent radars is ≥10%, avoiding detection blind spots.

[0034] Images captured by a monocular camera are converted to grayscale and segmented using a thresholding method. First, the color image is converted to grayscale, and then the grayscale value is calculated using a weighted average method. Gray = 0.3R + 0.59G + 0.11B ± 2 Where R, G, and B are the red, green, and blue channel pixel values ​​of the corresponding pixels in the color image, respectively (each ranging from 0 to 255). The ±2 in the formula is a compensation coefficient used to offset color deviations inherent in the sensor itself. This weighting coefficient is set based on the CIE 1931 standard colorimetric system, conforming to the sensitivity characteristics of the human eye to red, green, and blue light. It can maximally preserve the edge contours and texture features of obstacles in the original image, providing high-quality grayscale image data for subsequent thresholding and region of interest extraction.

[0035] After being converted to a grayscale image, the grayscale image is segmented using a fixed threshold method to obtain foreground and background pixels. The fixed threshold employs an adaptive dynamic adjustment mechanism based on illumination, combined with scene illumination intensity feedback to achieve threshold adjustment. The average grayscale value of each frame of grayscale image is extracted in real time, and the average grayscale value of a consecutive preset number of frames is calculated. The average grayscale value is used as an evaluation index of the current scene illumination intensity. In this embodiment, the consecutive preset number of frames can be 10 frames. When the proportion of foreground pixels in consecutive preset frame images is greater than the first preset percentage, and the average gray value of the image is greater than the first average, the current scene is determined to be a strong lighting environment. At this time, the fixed threshold will fluctuate upward by 5-10 units within the preset threshold range. Typically, the first preset percentage is set to 60%, the first average is set to 180, and the preset threshold range is (120-150). When the proportion of foreground pixels in a series of preset frames (10 frames) is less than the second preset percentage (10%), and the average gray value of the image is less than the second average (80), the current scene is determined to be a low-light environment. At this time, the fixed threshold will fluctuate downward by 5-10 units within the preset threshold range. The adjusted fixed threshold is limited to the extreme threshold range (110-160) to avoid excessive threshold adjustment leading to segmentation failure. The threshold adjustment adopts a smooth transition strategy, with each adjustment not exceeding 5 units to ensure the stability of the segmentation results. The extreme threshold range is greater than the preset threshold range.

[0036] The distance data collected by the ultrasonic radar is processed by moving average filtering, and the average value is calculated by taking the distance data of 3-5 consecutive collection cycles. At the same time, according to the installation position of the ultrasonic radar, a mapping relationship between the radar coordinate system and the vehicle body coordinate system is established, and the distance data collected by the radar is transformed into the vehicle body coordinate system.

[0037] The preprocessed image data and ultrasonic radar data are fused together. Based on the vehicle coordinate system, the distance range in front of / to the side of the vehicle corresponding to different areas in the monocular camera image is determined. When the ultrasonic radar detects an obstacle in a certain direction with a distance less than a preset safety threshold, the image area corresponding to that direction is extracted.

[0038] It should be noted that the preset safety threshold adopts a two-factor adaptive adjustment strategy based on driving speed and scene type, specifically as follows: Based on the vehicle-mounted map and positioning information of the unmanned vehicle, the low-speed scene is divided into three types: open road section, narrow passage and intersection. Real-time driving speed is obtained by the vehicle's onboard wheel speed sensors, and the speed is divided into three levels: low speed (≤3km / h), medium-low speed (3km / h < speed ≤5km / h), and medium speed (5km / h < speed ≤10km / h). For different combinations of scene types and speed levels, corresponding safety thresholds are preset: in open road sections, the safety threshold for low speed is 0.8m, for medium-low speed is 1m, and for medium speed is 2.5m; in narrow passages, the safety threshold for low speed is 1m, for medium-low speed is 1.2m, and for medium speed is 3m; at intersections, the safety threshold for low speed is 1.2m, for medium-low speed is 1.5m, and for medium speed is 3m. The safety thresholds are dynamically updated in real time according to changes in driving speed and scene type.

[0039] If the proportion of foreground pixels in the image region is greater than the foreground value (30%), then there is an obstacle in the image region. If the proportion of foreground pixels is not greater than the foreground value, then the wireless device is judged to be a false detection and the image data is ignored. The spatial coordinates of the obstacle and the preliminary judgment of whether it is a dynamic obstacle are output to the unmanned vehicle decision system.

[0040] The specific process for the preliminary assessment of dynamic obstacles is as follows: The distance data of the same obstacle within a continuous preset acquisition period is smoothed by linear interpolation to eliminate the influence of random noise on the calculation of the change. In this embodiment, three acquisition periods are used as the preset acquisition period for illustration.

[0041] Calculate the distance change Δd1 (difference between the 2nd and 1st cycles) and Δd2 (difference between the 3rd and 2nd cycles) between adjacent acquisition cycles, and the average rate of change v = (|Δd1| + |Δd2|) / (2T), where T is the acquisition cycle; Based on the calculated distance change and average rate of change, determine the dynamic attributes of the current obstacle. Level 1 dynamic obstacle (high-speed movement): cumulative distance change |Δd1 + Δd2| > 0.8m, or average rate of change v > 0.3m / s; Level 2 dynamic obstacle (medium-speed movement): 0.5m < |Δd1 + Δd2| ≤ 0.8m, and 0.15m / s < v ≤ 0.3m / s; Static obstacle: |Δd1 + Δd2| ≤ 0.5m, and v ≤ 0.15m / s. The confidence level of the dynamic attribute judgment result is evaluated. The confidence level is calculated as 1 - (distance measurement error / cumulative distance change). When the confidence level is ≥0.7, the judgment result is output. When the confidence level is <0.7, the judgment is re-evaluated after adding one more collection cycle of observation data.

[0042] In summary, the environmental perception system for unmanned vehicles in low-speed scenarios proposed in this invention significantly reduces costs compared to high-precision sensors such as LiDAR by employing a monocular camera and wireless devices as perception sensors. Simultaneously, the algorithm logic is simple, requiring low hardware computing power and eliminating the need for a high-performance computing platform, further reducing the overall system cost. The algorithm steps are concise, including data acquisition, simple preprocessing, data fusion, and obstacle detection, eliminating the need for complex deep learning model training and parameter tuning, thus reducing development difficulty and facilitating rapid deployment. Addressing the characteristics of low-speed scenarios—low environmental complexity and limited obstacle types—the simplified sensor fusion logic and obstacle detection rules can meet the basic perception needs of unmanned vehicles driving at low speeds, ensuring driving safety.

[0043] Example 3 In another aspect, the present invention also proposes a computer-readable storage medium having stored thereon one or more computer programs that, when executed by a processor, implement the above-described method for environmental perception of unmanned vehicles in low-speed scenarios.

[0044] Those skilled in the art will understand that the logic or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable storage medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-included system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable storage medium" can mean any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device.

[0045] More specific examples (a non-exhaustive list) of computer-readable storage media include: electrical connections (electronic devices) having one or more wires, portable computer disk drives (magnetic devices), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Furthermore, computer-readable storage media can even be paper or other suitable media on which the program can be printed, since the program can be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, interpreting, or otherwise processing as necessary, and then stored in computer memory.

[0046] Example 4 This embodiment provides a structural block diagram of an electronic device. The electronic device includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, it implements the environmental perception method for unmanned vehicles in low-speed scenarios described in the above embodiment. The electronic device is merely an example and should not impose any limitations on the functionality and scope of use of the embodiments of the present invention.

[0047] Electronic devices can take the form of general-purpose computing devices, such as server devices. Components of an electronic device may include, but are not limited to: at least one processor, at least one memory, and buses connecting different system components (including memory and processor).

[0048] The bus includes a data bus, an address bus, and a control bus.

[0049] The memory may include volatile memory, such as RAM321 (random access memory), and / or cache memory, and may further include ROM (read-only memory).

[0050] The memory may also include program tools having a set (at least one) of program modules, including but not limited to: an operating system, one or more application programs, other program modules, and program data, each or some combination of these examples may include an implementation of a network environment.

[0051] The processor executes various functional applications and data processing by running computer programs stored in memory, such as the environmental perception method for unmanned vehicles in low-speed scenarios described above.

[0052] The electronic device can also communicate with one or more external devices (such as keyboards, pointing devices, etc.). This communication can be achieved through I / O interfaces (input / output interfaces). Furthermore, the electronic device can communicate with one or more networks (such as local area networks (LANs), wide area networks (WANs), and / or public networks, such as the Internet) via a network adapter. The network adapter communicates with other modules of the model-generated electronic device via a bus. It should be understood that, although not shown in the figures, other hardware and / or software modules can be used in conjunction with the model-generated electronic device, including but not limited to: microcode, device drivers, redundant processors, disk drive arrays, RAID (disk array) systems, tape drives, and data backup storage systems.

[0053] It should be noted that although several units / modules or sub-units / modules of the electronic device have been mentioned in the detailed description above, this division is merely exemplary and not mandatory. In fact, according to embodiments of the present invention, the features and functions of two or more units / modules described above can be embodied in one unit / module. Conversely, the features and functions of one unit / module described above can be further divided and embodied by multiple units / modules.

[0054] In the description of this specification, references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0055] The embodiments described above are merely illustrative of several implementations of the present invention, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the present invention. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these all fall within the protection scope of the present invention. Therefore, the protection scope of this patent should be determined by the appended claims.

Claims

1. A method for environmental perception of unmanned vehicles in low-speed scenarios, characterized in that, The method for autonomous vehicle environmental perception in low-speed scenarios includes: The vehicle uses monocular cameras at the front and sides to collect environmental image data and wireless devices installed around the vehicle to collect distance data. The image data is converted to grayscale to obtain a grayscale image, and the grayscale image is segmented using a fixed threshold method to obtain foreground pixels and background pixels; The distance data is processed by moving average filtering, and a mapping relationship between the device coordinate system and the vehicle body coordinate system is established based on the spatial coordinates of the wireless device, so that the distance data is transformed into the vehicle body coordinate system; By fusing the preprocessed image data and the distance data, the image region containing obstacles within a preset safety threshold is extracted, and the foreground ratio within the image region is obtained. Based on the foreground ratio, the judgment result of the dynamic obstacle is output.

2. The method for environmental perception of unmanned vehicles in low-speed scenarios according to claim 1, characterized in that, In the step of converting the image data to grayscale, the color image is converted into a radian image, and the grayscale value is calculated using a weighted average method. Gray = 0.3R + 0.59G + 0.11B ± 2 Where R, G, and B are the red, green, and blue channel pixel values ​​of the corresponding pixel points in the color image, respectively (each ranging from 0 to 255). The ±2 in the formula is a compensation coefficient used to offset the color deviation of the sensor itself.

3. The method for environmental perception of unmanned vehicles in low-speed scenarios according to claim 2, characterized in that, In the step of thresholding the grayscale image using a fixed threshold method, the fixed threshold employs an illumination-adaptive dynamic adjustment mechanism, combining scene illumination intensity feedback to achieve threshold adjustment: The average grayscale value of each frame of grayscale image is extracted in real time, the average grayscale value of consecutive preset number of frames is calculated, and the average grayscale value is used as an evaluation index of the current scene illumination intensity. When the proportion of foreground pixels in consecutive preset frame images is greater than the first preset percentage, and the average grayscale value of the images is greater than the first average, the current scene is determined to be a strong lighting environment, and the fixed threshold is adjusted upward within the preset threshold range. When the proportion of foreground pixels in consecutive preset frame images is less than the second preset percentage, and the average grayscale value of the images is less than the second average, the current scene is determined to be a low-light environment, and the fixed threshold is adjusted downward within the preset threshold range. The adjusted fixed threshold is limited to a limit threshold range, which is greater than the preset threshold range.

4. The method for environmental perception of unmanned vehicles in low-speed scenarios according to claim 1, characterized in that, The step of extracting the obstacle image region within a preset safety threshold by fusing the preprocessed image data and the distance data is as follows: After fusing the image data and the distance data, the range of distances in front of the vehicle and / or the range of distances to the sides of the vehicle corresponding to different areas in the monocular camera image are determined based on the vehicle coordinate system. When a wireless device detects an obstacle within a preset safety threshold, it extracts the image region corresponding to the direction in which the obstacle exists.

5. The method for environmental perception of unmanned vehicles in low-speed scenarios according to claim 4, characterized in that, After extracting the corresponding image region, determine the proportion of foreground pixels within the image region. If the proportion of foreground pixels is greater than the foreground value, then there is an obstacle in the image area; if the proportion of foreground pixels is not greater than the foreground value, then the wireless device is determined to be a false detection and the image data is ignored. The spatial coordinates of the obstacle and a preliminary judgment on whether it is a dynamic obstacle are output to the autonomous vehicle decision system.

6. The method for environmental perception of unmanned vehicles in low-speed scenarios according to claim 5, characterized in that, The specific steps for the preliminary determination of whether it is a dynamic obstacle are as follows: Linear interpolation smoothing is performed on the distance data of the same obstacle within a continuous preset acquisition period to eliminate the influence of random noise on the calculation of changes. Calculate the distance change and average rate of change between adjacent acquisition cycles, compare the calculation results with the preset judgment values ​​of the unmanned vehicle decision system, and output the obstacle dynamic attribute category. The confidence level of the dynamic attribute category judgment result is evaluated. If the confidence level is less than the confidence threshold, the judgment is re-evaluated after adding observation data for one more collection period.

7. An environmental perception system for unmanned vehicles in low-speed scenarios, characterized in that, The low-speed scene autonomous vehicle environment perception system is used to implement the low-speed scene autonomous vehicle environment perception method according to any one of claims 1-6, the system comprising: The data acquisition module is used to acquire environmental image data using monocular cameras at the front and sides of the unmanned vehicle, and to acquire distance data through wireless devices installed around the unmanned vehicle. The segmentation module is used to convert the image data to grayscale to obtain a grayscale image, and to perform threshold segmentation on the grayscale image using a fixed threshold method to obtain foreground pixels and background pixels; The conversion module is used to perform moving average filtering on the distance data, and establish a mapping relationship between the device coordinate system and the vehicle body coordinate system based on the spatial coordinates of the wireless device, so as to convert the distance data into the vehicle body coordinate system. The judgment module is used to fuse the preprocessed image data and the distance data, extract the image region of the obstacle within the preset safety threshold, obtain the foreground ratio in the image region, and output the judgment result of the dynamic obstacle based on the foreground ratio.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by the processor, the program implements the environmental perception method for unmanned vehicles in low-speed scenarios as described in any one of claims 1-6.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, When the processor executes the computer program, it implements the method for environmental perception of unmanned vehicles in low-speed scenarios as described in any one of claims 1-6.

Images