Methods and devices for detecting dirt in lidar optical windows, and methods and devices for cleaning them.

Abstract

This disclosure provides a method and apparatus for detecting and cleaning dirt on a lidar window, relating to the field of lidar detection technology. The specific implementation involves: emitting a laser beam; receiving the echo signal generated by the laser beam; and determining the dirt level corresponding to different areas of the lidar window based on the pulse width of the echo signal. This disclosure provides graded warnings for different degrees of dirt on the lidar window and can accurately locate dirty areas, improving detection accuracy, guiding vehicle cleaning systems to perform targeted cleaning, and enhancing the overall safety performance of autonomous driving systems.

Description

Technical Field

[0001] This disclosure relates to the field of lidar detection technology, and in particular to a method and apparatus for detecting dirt on lidar windows, a cleaning method and apparatus, and an autonomous vehicle. Background Technology

[0002] With the rapid development of autonomous driving technology, automotive LiDAR is widely used in automobiles. The functions of automotive LiDAR include environmental perception, assisting in route planning, enhancing driving safety, and providing detailed positioning. LiDAR is a sensor that can accurately detect the position of objects. It works by emitting a laser signal towards a target object, calculating the distance based on the time difference of the reflected signal, and then determining the angle between the object and the transmitter based on the laser emission angle, thus obtaining the relative relationship between the object and the transmitter.

[0003] In existing technologies, dust, rain, snow, insect remains, and other contaminants can accumulate on the lidar's optical window during vehicle operation, affecting its ranging performance. This dirt prevents emitted light from effectively penetrating the window, and the received signal attenuates upon returning to the window. Generally, dirt increases blind spots; severe dirt can create holes in the ground and target point cloud, resulting in missing vehicle perception point clouds, misjudgments, and even rendering the lidar unusable. Currently, there is a problem with accurately determining the degree and area of ​​dirt on the optical window. Summary of the Invention

[0004] This disclosure provides a method and apparatus for detecting dirt in a lidar window, a cleaning method and apparatus, an electronic device, a storage medium, and an autonomous vehicle.

[0005] According to a first aspect of this disclosure, a method for detecting dirt in a lidar optical window is provided, comprising:

[0006] Emit a laser beam;

[0007] Receive the echo signal generated by the laser beam;

[0008] The level of contamination corresponding to different areas of the lidar window is determined based on the pulse width of the echo signal.

[0009] According to a second aspect of this disclosure, a method for cleaning a lidar optical window is provided, comprising:

[0010] The dirt level of each area of ​​the lidar window is obtained; wherein the dirt level is obtained by the dirt detection method described in any one of the above technical solutions;

[0011] Determine the cleaning strategy for each area based on the level of dirtiness;

[0012] Perform corresponding cleaning actions for each area according to the cleaning strategy.

[0013] According to a third aspect of this disclosure, a dirt detection device for a lidar window is provided, comprising:

[0014] The transmitting module is configured to emit a laser beam;

[0015] The receiving module is configured to receive the echo signal generated by the laser beam;

[0016] The detection module is configured to determine the level of dirt corresponding to different areas of the lidar window based on the pulse width of the echo signal.

[0017] According to a fourth aspect of this disclosure, a cleaning device for a lidar optical window is provided, comprising:

[0018] The acquisition module is configured to acquire the dirt level corresponding to each area of ​​the lidar window; wherein the dirt level is acquired by the dirt detection method according to any one of claims 1-7;

[0019] The strategy determination module is configured to determine the cleaning strategy corresponding to each area based on the level of dirtiness.

[0020] The execution module is configured to perform corresponding cleanup actions for each region according to the cleanup strategy.

[0021] According to a fifth aspect of this disclosure, an electronic device is provided, comprising:

[0022] At least one processor; and

[0023] A memory communicatively connected to the at least one processor; wherein,

[0024] The memory stores instructions that can be executed by the at least one processor, which, when executed by the at least one processor, enables the at least one processor to perform the method described in any of the above technical solutions.

[0025] According to a sixth aspect of this disclosure, a non-transitory computer-readable storage medium is provided storing computer instructions, wherein the computer instructions are used to cause the computer to perform any one of the methods described above.

[0026] According to a seventh aspect of this disclosure, a computer program product is provided, comprising a computer program that, when executed by a processor, implements the method described in any one of the above technical solutions.

[0027] According to the eighth aspect of this disclosure, an autonomous driving vehicle is provided, including the electronic devices described in the above-described technical solutions.

[0028] This disclosure provides a method for detecting dirt in a lidar window. By classifying the dirt, different levels of dirt in the lidar window can be graded and given early warnings. It can also accurately locate dirty areas, improve detection accuracy, guide the vehicle cleaning system to perform targeted cleaning, and enhance the overall safety performance of the autonomous driving system.

[0029] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of this disclosure, nor is it intended to limit the scope of this disclosure. Other features of this disclosure will become readily apparent from the following description. Attached Figure Description

[0030] The accompanying drawings are provided to better understand this solution and do not constitute a limitation of this disclosure. Wherein:

[0031] Figure 1 This is a step diagram of the dirt detection method for the lidar optical window in the embodiments of this disclosure;

[0032] Figure 2 This is a diagram illustrating the specific steps for determining the level of dirtiness in an embodiment of this disclosure;

[0033] Figure 3 This is a schematic diagram illustrating the principle of dividing the lidar optical window into regions in this embodiment of the present disclosure;

[0034] Figure 4 This is a step diagram of the method for cleaning the lidar window in an embodiment of this disclosure;

[0035] Figure 5 This is a schematic block diagram of the dirt detection device for the lidar optical window in the embodiments of this disclosure;

[0036] Figure 6 This is a block diagram illustrating the principle of the detection module in an embodiment of this disclosure;

[0037] Figure 7 This is a schematic block diagram of the cleaning device for the lidar optical window in the embodiments of this disclosure;

[0038] Figure 8 This is a block diagram of an electronic device used to implement the method for detecting dirt and cleaning the lidar window in the embodiments of this disclosure. Detailed Implementation

[0039] The exemplary embodiments of this disclosure are described below with reference to the accompanying drawings, including various details of the embodiments to aid understanding, and should be considered merely exemplary. Therefore, those skilled in the art will recognize that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of this disclosure. Similarly, for clarity and brevity, descriptions of well-known functions and structures are omitted in the following description.

[0040] Existing methods for detecting dirt on lidar windows mainly include: The first involves adding an additional light-emitting device to detect dirt on the photomask. A beam transceiver transmits light through the window and receives it; when the window is dirty, it blocks the light path, preventing the reception of emitted light and triggering a dirt alarm. The second method involves monitoring the window with an external imaging device. The dirt level is determined by analyzing the image of the window surface using image recognition. However, this method cannot detect dirt on the lidar window if the external imaging device is dirty. Furthermore, current detection methods suffer from inaccurate judgment. Therefore, it is crucial to classify dirt levels, such as dirt and severe dirt, and to provide graded warnings for dirt and severe dirt to guide the vehicle cleaning system in cleaning.

[0041] To address the aforementioned technical problems, this disclosure provides a method for detecting dirt in a lidar optical window, such as... Figure 1 As shown, it includes:

[0042] Step S101: Emit a laser beam;

[0043] Step S102: Receive the echo signal generated by the laser beam;

[0044] Step S103: Determine the level of dirtiness corresponding to different areas of the lidar window based on the pulse width of the echo signal.

[0045] Specifically, before performing contamination detection, the degree of contamination of the optical window is classified into several levels: slight contamination, moderate contamination, relatively severe contamination, and severe contamination (e.g., large-area contamination causing the optical window to be in a severely contaminated state). A mapping relationship between the echo signal pulse width and the contamination level is established, and the corresponding contamination level is determined based on this mapping relationship during the detection process. This disclosure can utilize the laser beam transceiver and laser radar chip itself for contamination detection, eliminating the need for additional beam transceiver or imaging equipment for laser radar optical window contamination detection. Steps S101 and S102 are executed by the laser beam transceiver of the laser radar itself, and the contamination level corresponding to different areas of the laser radar optical window is determined by the laser radar chip based on the pulse width of the echo signal. Since the pulse width of the echo signal after the light window is dirty is greater than that of the echo signal when the light window is clean, multiple thresholds can be set according to the pulse width of the echo signal when the light window is clean. Each threshold corresponds to a level of dirtiness. By judging whether the pulse width of the echo signal reaches the corresponding threshold, the degree of dirtiness of the light window can be determined and graded for early warning. The light window can be cleaned in time, thereby improving the overall safety performance of the autonomous driving system.

[0046] As an optional implementation method, such as Figure 2As shown, step S103, determining the dirt level corresponding to different areas of the lidar window based on the pulse width of the echo signal, includes:

[0047] Step S1031: Determine whether each point is a dirt spot based on the pulse width of the echo signal returned by each point of the lidar window.

[0048] Step S1032: Count the number of dirt spots in each area of ​​the LiDAR window and determine the dirt status of each area.

[0049] Step S1033: Determine the dirt level of each area based on the duration of the dirt state of each area.

[0050] Specifically, in this embodiment, it first determines whether each point on the light window is a dirty spot, then counts the number of dirty spots in different areas to determine whether an area is dirty. If an area is dirty, the duration of the dirty state determines the corresponding dirt level of the current area. By detecting points and areas, the main dirty areas can be identified. Combining the time dimension can effectively reduce false positives. For example, if a flying insect or a speck of dust is on the light window, it may be blown away by the wind in a very short time, so the light window will not have a dirt problem in the next frame. If the time dimension is not considered, the system will immediately alarm as soon as it detects that the pulse width of the echo signal reaches the threshold, which will affect the user experience. This embodiment determines the dirt status of the light window through three dimensions: point, area, and time, and finally determines the dirt level of each area, which can achieve regional and graded detection, resulting in more accurate detection results and improved user experience.

[0051] As an optional implementation, the dirt detection method further includes: dirt spots include general dirt spots and severe dirt spots; step S1031, determining whether each point is a dirt spot based on the pulse width of the echo signal returned by each point of the lidar window includes:

[0052] The pulse width of each echo signal is compared with a preset reference pulse width; wherein the preset reference pulse width includes a first preset pulse width and a second preset pulse width.

[0053] If the pulse width of the echo signal is greater than or equal to the first preset pulse width and less than the second preset pulse width, the dirt state corresponding to the point is determined to be a general dirt point.

[0054] The system responds to the pulse width of the echo signal being greater than or equal to the second preset pulse width, and determines that the dirt state corresponding to that point is a severely dirty point.

[0055] If the pulse width of the echo signal is less than the first preset pulse width, it is determined that the point is not a dirt spot.

[0056] Specifically, ordinary dirt spots can increase radar blind spots; severe dirt spots can create holes in the ground and target point cloud, resulting in missing radar point cloud data, creating an obstruction effect, and leading to radar misjudgment. Therefore, further subdividing the degree of dirt spots allows for a more accurate assessment of the dirt status of an area.

[0057] For example, under conditions of no ambient light interference, no obstructions, and a clean light window, echo signals T0 from various points in different fields of view are pre-collected as reference signals, and the pulse width of the T0 signal is denoted as W0, which serves as the preset reference pulse width. During the detection process, echo signals T1 from various points in different fields of view are collected, and the pulse width of the echo signal T1 is denoted as W1. The signal pulse width W1 acquired during the detection process is compared with a preset reference pulse width W0. The first preset pulse width can be a threshold of x% exceeding the pulse width W0. For example, if the preset reference pulse width W0 is 5ns and x% is set to 40%, when the echo signal width W1 at a certain point on the optical window exceeds 40% of the preset reference pulse width W0, that is, when the echo signal width W1 reaches 7ns (the first preset pulse width), the point is determined to be a general dirt spot. The second preset pulse width can be a threshold of y% exceeding the pulse width W0. For example, if the preset reference pulse width W0 is 5ns and y% is set to 60%, when the echo signal width W1 at a certain point on the optical window exceeds 60% of the preset reference pulse width W0, that is, when the echo signal width W1 reaches 8ns (the second preset pulse width), the point is determined to be a severe dirt spot. Through the above scheme, not only can dirt spots be identified, but the degree of dirtiness of each dirt spot can also be determined. Different treatment measures can be taken for general dirt spots and severe dirt spots.

[0058] As an optional implementation, before determining whether each point is a dirt spot based on the pulse width of the echo signal returned from each point of the lidar optical window, the method further includes:

[0059] The lidar window is pre-divided into regions, namely a central region and a non-central region;

[0060] Different first preset pulse widths and second preset pulse widths are set for the central region and the non-central region; wherein, the first preset pulse width corresponding to the central region is smaller than the first preset pulse width corresponding to the non-central region, and the second preset pulse width corresponding to the central region is smaller than the second preset pulse width corresponding to the non-central region.

[0061] Specifically, the light window can be divided into central and non-central areas, with further subdivisions within the non-central area. The level of dirtiness in both the central and non-central areas is categorized into several levels: slightly dirty, moderately dirty, fairly dirty, and severely dirty. Since vehicle operation demands stricter standards for dirtiness in the central area of ​​the light window, even a small amount of dirt will trigger an alarm indicating a severe situation. Therefore, by dividing the field of view into central and non-central areas, a relatively low threshold can be set for the central area to enhance dirt monitoring. Even a small amount of dirt will trigger an alarm, ensuring the central area remains relatively clean and guaranteeing safe driving.

[0062] like Figure 3 As shown, this embodiment can be divided into three location points: A, B, and C. A is the central region, B and C are non-central regions, B is the secondary central region, and C is the edge region. Different judgment thresholds can also be set for B and C. Specific implementation includes: assuming the echo signal pulse width W0 collected under clean conditions is 5ns, when x% is 40%, if the echo signal pulse width W1 collected at a point in region A is greater than W0 reaching 40%, it is judged as a general dirt spot, i.e., when W1 ≥ 7ns, it is judged as a general dirt spot; when y% is 60%, if the echo signal pulse width W1 collected at a point in region A is greater than W0 reaching 60%, it is judged as a severely dirty spot, i.e., W1 > 8ns, it is judged as a severely dirty spot. Since vehicle driving requirements are more lenient regarding the degree of dirt in the non-central region B, the threshold for judging dirt in region B can be set higher, such as setting the echo signal pulse width W0 in region B to 7ns and the echo signal pulse width W0 in region C to 10ns.

[0063] As an optional implementation, step S1032, counting the number of dirt spots in each area of ​​the lidar window and determining the dirt status of each area includes:

[0064] Count the number of dirt spots in each area of ​​the lidar window and determine the total number of spots in each area of ​​the lidar window;

[0065] Based on the number of dirt spots in each area and the total number of spots, calculate the percentage of dirt spots and severely dirty spots in each area;

[0066] When the proportion of severely soiled spots in each area reaches a first preset ratio, the soiling status of that area is determined to be severely soiled.

[0067] If the proportion of severely dirty spots in each area does not reach the first preset proportion, and the total proportion of general dirty spots and severely dirty spots reaches the second preset proportion, the dirt status of the area is determined to be general dirty status.

[0068] Specifically, assuming each area of ​​the light window theoretically has N points, if the number of severely soiled points exceeds m% of the total number N, then the soiling state of this area is marked as severely soiled; if the total number of general soiled points plus severely soiled points exceeds n% of the total number of points N, then this small area is marked as generally soiled. Specific example: Assuming the total number of points N = 300 in area A, when the ratio m% of the number of severely soiled points M to the total number of points N is > 10%, that is, when the number of severely soiled points exceeds 30, the soiling state of area A is severely soiled; when there are 45 soiled points L, 20 severely soiled points M, and the total number of M + L is 65, and the ratio n% of the total number of M + L to the total number of points N is > 20%, the soiling state of this area is generally soiled.

[0069] As an optional implementation, A can be further divided into multiple regions, for example, A can be further divided into Z regions. The dirt status of each small region can be determined, which can more accurately determine the location of dirt and the degree of dirt at that location, thus achieving more accurate dirt detection.

[0070] As an optional implementation, before step S1032, which counts the number of dirt spots in each area of ​​the lidar window and determines the dirt status of each area, the method further includes:

[0071] The lidar window is pre-divided into regions, namely a central region and a non-central region;

[0072] Different first and second preset ratios are set for the central and non-central areas; wherein the first preset ratio for the central area is smaller than the first preset ratio for the non-central area, and the second preset ratio for the central area is smaller than the second preset ratio for the non-central area.

[0073] Specifically, before detecting the dirt status of a region, the light window can be divided into regions, specifically a central region and a non-central region. The non-central region can be further subdivided. Dirt levels can be categorized for both the central and non-central regions, with each region classified as slightly dirty, moderately dirty, moderately dirty, and severely dirty. Since vehicle operation requires stricter control over the dirt level in the central region of the light window, even a small amount of dirt will trigger an alarm as a severe situation. Therefore, by dividing the field of view into a central and non-central region, a relatively low threshold can be set for the central region to enhance dirt monitoring. Once even a small amount of dirt is detected, an alarm will sound, ensuring the central region remains relatively clean and guaranteeing safe driving.

[0074] As an optional implementation, step S1033, determining the dirt level for each area based on the duration of the dirt state for each area, includes:

[0075] Multiple time thresholds are preset, and a mapping relationship is established between the multiple time thresholds and multiple dirt levels;

[0076] Based on the duration and mapping relationship of the dirt status of each region within the current statistical period, the current dirt level of each region is determined.

[0077] Specifically, determining the current dirt level for each region based on the duration and mapping relationship of the dirt state within the current statistical period can include:

[0078] When the area remains in a state of general dirtiness for a first preset time, the area is determined to be at the level of slight dirtiness.

[0079] When the area remains in a state of general dirtiness for a second preset time, the area is determined to be at a moderate level of dirtiness.

[0080] When the area remains in a state of general dirtiness for a third preset time, the area is determined to be at a relatively severe level of dirtiness.

[0081] When the area reaches the fourth preset time for a sustained state of severe contamination, the area is determined to be at the severe contamination level.

[0082] In this embodiment, each region is marked with a status indicator every 100ms, and the dirt status of all regions (no dirt, moderate dirt, or severe dirt) is statistically analyzed. Moderate dirt can be further classified into several levels based on the duration of the moderate dirt state: slight dirt, moderate dirt, and severe dirt. If the duration of dirt in a region exceeds p1% within time t, the region is classified as slightly dirty; if it exceeds p2%, it is classified as moderately dirty; if it exceeds p3%, it is classified as severely dirty; and if the duration of severe dirt reaches p4%, it is classified as severely dirty and reported. In this embodiment, p1% < p2% < p3%, and p4% can be greater than p2%.

[0083] For example, suppose a small area within region A is region Q, and one frame lasts 100ms. Assume t = 500ms (i.e., one statistical period is 5 frames). If the time spent in a dirty state exceeds 20%, meaning a dirty alarm occurs in 1 out of every 5 frames, it is classified as a light dirty level. If the time spent in a dirty state exceeds 40%, meaning a dirty alarm occurs in 2 out of every 5 frames, it is classified as a moderate dirty level. If the time spent in a dirty state exceeds 80%, meaning a dirty alarm occurs in 4 out of every 5 frames, it is classified as a severe dirty level. For areas already marked as severely dirty, if the time spent in a severely dirty state exceeds 60%, meaning a severe dirty alarm occurs in 3 out of every 5 frames, it is classified as a severely dirty level, and there is no need to set multiple thresholds for judgment.

[0084] In this embodiment, combining the time dimension can effectively reduce misjudgments. For example, if a flying insect or a speck of dust stays on the light window, the wind may blow it away in a very short time. In the next frame, the light window will not have a dirt problem. If the time dimension is not combined for judgment, the system will immediately alarm once it detects that the pulse width of the echo signal reaches the threshold, which will affect the user experience.

[0085] It should be noted that the thresholds W0, x, y, n, m, t, p1, p2, and p3 in this embodiment need to be determined through actual testing and requirements. They are first defined based on the level of dirtiness in different light windows, and then the values ​​of W0, x, y, n, m, t, p1, p2, and p3 are determined through testing. Their values ​​will vary under different light window types and different testing scenarios, and will also differ for different regions (central or non-central regions).

[0086] This disclosure also provides a method for cleaning the lidar optical window, such as... Figure 4 As shown, it includes:

[0087] Step S401: Obtain the dirt level corresponding to each area of ​​the lidar window; wherein, the dirt level is obtained by the dirt detection method described in any of the above embodiments;

[0088] Step S402: Determine the cleaning strategy for each area based on the level of dirt.

[0089] Step S403: Perform the corresponding cleaning action for each area according to the cleaning strategy.

[0090] Specifically, the cleaning strategies in this embodiment include, but are not limited to, at least one of the following: air jetting; water spraying; and disengaging the autonomous driving system. After the dirt level of each area of ​​each light window is obtained by dirt detection, for example, if a certain area is determined to be slightly dirty, moderately dirty, seriously dirty, or severely dirty, it can be reported to the vehicle control system. The vehicle control system determines the corresponding cleaning strategy based on the current dirt level, so as to take reasonable cleaning measures for different dirt conditions. This avoids unreasonable cleaning measures that result in incomplete cleaning of the light windows, causing radar misjudgment and affecting driving safety. At the same time, it avoids over-cleaning of the light windows. For example, if the light windows are only slightly dirty and do not affect driving, automatically disengaging the driving system to clean them would affect the user experience.

[0091] For example, when the dirt is determined to be slightly dirty, the vehicle control system activates the self-cleaning system to spray air onto the area. After spraying, the system can reassess the level of dirt to determine whether the alarm needs to be deactivated. When the dirt is determined to be moderately dirty, the vehicle control system activates the self-cleaning system to spray water, and then determines whether the alarm needs to be deactivated. When the dirt is determined to be severely dirty, the vehicle needs to exit the autonomous driving system before spraying water, and then determines whether the alarm needs to be deactivated. When the dirt is determined to be severely dirty (such as obstructing radar point cloud data, which may seriously affect the safety performance of autonomous driving), for such large-area, long-term severe dirt, the autonomous driving system needs to be exited immediately. Otherwise, a serious safety accident may occur due to system misjudgment. After exiting the autonomous driving system, water spraying and air spraying are performed again, and finally, the system determines whether the alarm needs to be deactivated after cleaning. Through the above method, different cleaning strategies can be selected based on the accurate detected level of dirt and the dirty area, and the light window can be cleaned in a timely manner to avoid the radar misjudging pedestrians or crossing objects due to dirty light windows, causing the autonomous vehicle to brake suddenly or decelerate incorrectly, affecting driving safety.

[0092] This disclosure provides a dirt detection device 500 for a lidar optical window, such as... Figure 5 As shown, it includes:

[0093] The transmitting module 501 is configured to emit a laser beam;

[0094] The receiving module 502 is configured to receive the echo signal generated by the laser beam;

[0095] The detection module 503 is configured to determine the level of dirt corresponding to different areas of the lidar window based on the pulse width of the echo signal.

[0096] Specifically, before performing contamination detection, the degree of contamination of the optical window is classified into several levels: slight contamination, moderate contamination, relatively severe contamination, and severe contamination. A mapping relationship between the echo signal pulse width and the contamination level is established, and the corresponding contamination level is determined based on this mapping relationship during the detection process. This disclosure can utilize the laser beam transceiver and laser radar chip itself for contamination detection, eliminating the need for additional beam transceiver or imaging equipment for laser radar optical window contamination detection. The transmitting module 501 and receiving module 502 can be the laser beam transceiver of the laser radar itself, and the detection module 503 can be the laser radar chip, without the need for additional equipment. Steps S101 and S102 are executed by the transmitting module 501 and receiving module 502, and then the detection module 503 determines the contamination level corresponding to different areas of the laser radar optical window based on the pulse width of the echo signal. Since the pulse width of the echo signal after the light window is dirty is greater than that of the echo signal when the light window is clean, multiple thresholds can be set according to the pulse width of the echo signal when the light window is clean. Each threshold corresponds to a level of dirtiness. By judging whether the pulse width of the echo signal reaches the corresponding threshold, the degree of dirtiness of the light window can be determined and graded for early warning. The light window can be cleaned in time, thereby improving the overall safety performance of the autonomous driving system.

[0097] As an optional implementation method, such as Figure 6 As shown, the detection module 503 includes:

[0098] The point detection unit 5031 is configured to determine whether each point is a dirt spot based on the pulse width of the echo signal returned by each point of the lidar window.

[0099] The area detection unit 5032 is configured to count the number of dirt points in each area of ​​the lidar window and determine the dirt status of each area.

[0100] The dirt level determination unit 5033 is configured to determine the dirt level corresponding to each area based on the duration of the dirt state corresponding to each area.

[0101] Specifically, in this embodiment, it first determines whether each point on the light window is a dirty spot, then counts the number of dirty spots in different areas to determine whether an area is dirty. If an area is dirty, the duration of the dirty state determines the corresponding dirt level of the current area. By detecting points and areas, the main dirty areas can be identified. Combining the time dimension can effectively reduce false positives. For example, if a flying insect or a speck of dust is on the light window, it may be blown away by the wind in a very short time, so the light window will not have a dirt problem in the next frame. If the time dimension is not considered, the system will immediately alarm as soon as it detects that the pulse width of the echo signal reaches the threshold, which will affect the user experience. This embodiment determines the dirt status of the light window through three dimensions: point, area, and time, and finally determines the dirt level of each area, which can achieve regional and graded detection, resulting in more accurate detection results and improved user experience.

[0102] As an optional implementation, the point detection unit 5031 determines whether each point is a dirt spot based on the pulse width of the echo signal returned from each point of the lidar optical window, including:

[0103] The pulse width of each echo signal is compared with a preset reference pulse width; wherein the preset reference pulse width includes a first preset pulse width and a second preset pulse width.

[0104] If the pulse width of the echo signal is greater than or equal to the first preset pulse width and less than the second preset pulse width, the point is determined to be a general dirt spot.

[0105] If the pulse width of the echo signal is greater than or equal to the second preset pulse width, the point is determined to be a severely dirty spot.

[0106] If the pulse width of the echo signal is less than the first preset pulse width, it is determined that the point is not a dirt spot.

[0107] For example, dirt spots can be roughly divided into two levels in advance: general dirt spots and severe dirt spots. Under conditions of no ambient light interference, no obstructions, and a clean light window, echo signals T0 from various points in different fields of view are pre-collected as reference signals, and the pulse width of the T0 signal is denoted as W0, which serves as the preset reference pulse width. During the detection process, echo signals T1 from various points in different fields of view are collected, and the pulse width of the echo signal T1 is denoted as W1. The signal pulse width W1 acquired during the detection process is compared with a preset reference pulse width W0. The first preset pulse width can be a threshold that exceeds the pulse width W0 by x%. For example, if the preset reference pulse width W0 is 5ns and x% is set to 40%, when the echo signal width W1 at a certain point on the optical window exceeds 40% of the preset reference pulse width W0, that is, when the echo signal width W1 reaches 7ns (the first preset pulse width), then that point is determined to be a general dirt spot. The second preset pulse width can be a threshold that exceeds the pulse width W0 by y%. For example, if the preset reference pulse width W0 is 5ns and y% is set to 60%, when the echo signal width W1 at a certain point on the optical window exceeds 60% of the preset reference pulse width W0, that is, when the echo signal width W1 reaches 8ns (the second preset pulse width), then that point is determined to be a severe dirt spot. Through the above scheme, it is possible to determine whether each point is a dirt spot and the degree of dirtiness, and different treatment measures can be taken for general dirtiness and severe dirtiness.

[0108] As an optional implementation, before the point detection unit 5031 determines whether each point is a dirt spot based on the pulse width of the echo signal returned by each point of the lidar optical window, it further includes:

[0109] The lidar window is pre-divided into regions, namely a central region and a non-central region;

[0110] Different first preset pulse widths and second preset pulse widths are set for the central region and the non-central region; wherein, the first preset pulse width corresponding to the central region is smaller than the first preset pulse width corresponding to the non-central region, and the second preset pulse width corresponding to the central region is smaller than the second preset pulse width corresponding to the non-central region.

[0111] Specifically, the light window can be divided into central and non-central areas, with further subdivisions within the non-central area. The level of dirtiness in both the central and non-central areas is categorized into several levels: slightly dirty, moderately dirty, fairly dirty, and severely dirty. Since vehicle operation demands stricter standards for dirtiness in the central area of ​​the light window, even a small amount of dirt will trigger an alarm indicating a severe situation. Therefore, by dividing the field of view into central and non-central areas, a relatively low threshold can be set for the central area to enhance dirt monitoring. Even a small amount of dirt will trigger an alarm, ensuring the central area remains relatively clean and guaranteeing safe driving.

[0112] like Figure 3 As shown, this embodiment can be divided into three location points: A, B, and C. A is the central region, B and C are non-central regions, B is the secondary central region, and C is the edge region. Different judgment thresholds can also be set for B and C. Specific implementation includes: assuming the echo signal pulse width W0 collected under clean conditions is 5ns, when x% is 40%, if the echo signal pulse width W1 collected at a point in region A is greater than W0 reaching 40%, it is judged as a dirty spot; that is, when W1 ≥ 7ns, it is judged as a general dirty spot. When y% is 60%, if the echo signal pulse width W1 collected at a point in region A is greater than W0 reaching 60%, it is judged as a severely dirty spot; that is, W1 > 8ns, it is judged as a severely dirty spot. Since vehicle driving requirements are more lenient on the degree of dirtiness in the non-central region B, the threshold for judging dirtiness in region B can be set higher, such as setting the echo signal pulse width W0 in region B to 7ns and the echo signal pulse width W0 in region C to 10ns.

[0113] As an optional implementation, the area detection unit 5032 counts the number of dirt spots in each area of ​​the lidar window and determines the dirt status of each area, including:

[0114] Count the number of dirt spots in each area of ​​the lidar window and determine the total number of spots in each area of ​​the lidar window;

[0115] Calculate the percentage of dirt points in each area based on the number of dirt points in each area and the total number of dirt points.

[0116] When the proportion of severely soiled spots in each area reaches a first preset ratio, the soiling status of that area is determined to be severely soiled.

[0117] If the proportion of severely dirty spots in each area does not reach the first preset proportion, and the total proportion of general dirty spots and severely dirty spots reaches the second preset proportion, the dirt status of the area is determined to be general dirty status.

[0118] Specifically, assuming each area of ​​the light window theoretically has N points, if the number of severely soiled points exceeds m% of the total number N, then the soiling status of this area is marked as severely soiled; if the total number of general soiled points plus severely soiled points exceeds n% of the total number of points N, then this small area is marked as generally soiled. Specific example: Assuming the total number of points N = 300 in area A, when the ratio m% of severely soiled points M to the total number of points N is greater than 10%, that is, when the number of severely soiled points exceeds 30, the soiling status of area A is severely soiled; when there are 45 soiled points L and 20 severely soiled points M, which is less than 10%, and the total number of M + L is 65, and the ratio n% of the total number of M + L to the total number of points N is greater than 20%, the soiling status of this area is generally soiled.

[0119] As an optional implementation, A can be further divided into multiple regions, for example, A can be further divided into Z regions. The dirt status of each small region can be determined, which can more accurately determine the location of dirt and the degree of dirt at that location, thus achieving more accurate dirt detection.

[0120] As an optional implementation, before the area detection unit 5032 counts the number of dirt spots in each area of ​​the lidar window and determines the dirt status corresponding to each area, it further includes:

[0121] The lidar window is pre-divided into regions, namely a central region and a non-central region;

[0122] Different first and second preset ratios are set for the central and non-central areas; wherein the first preset ratio for the central area is smaller than the first preset ratio for the non-central area, and the second preset ratio for the central area is smaller than the second preset ratio for the non-central area.

[0123] Specifically, before detecting the dirt status of a region, the light window can be divided into regions, specifically a central region and a non-central region. The non-central region can be further subdivided. Dirt levels can be categorized for both the central and non-central regions, with each region classified as slightly dirty, moderately dirty, moderately dirty, and severely dirty. Since vehicle operation requires stricter control over the dirt level in the central region of the light window, even a small amount of dirt will trigger an alarm as a severe situation. Therefore, by dividing the field of view into a central and non-central region, a relatively low threshold can be set for the central region to enhance dirt monitoring. Once even a small amount of dirt is detected, an alarm will sound, ensuring the central region remains relatively clean and guaranteeing safe driving.

[0124] As an optional implementation, the dirt level determination unit 5033 determines the dirt level of each area based on the duration of the dirt state corresponding to each area, including:

[0125] Multiple time thresholds are preset, and a mapping relationship is established between the multiple time thresholds and multiple dirt levels;

[0126] Based on the duration and mapping relationship of the dirt status of each region within the current statistical period, the current dirt level of each region is determined.

[0127] Specifically, the dirt level determination unit 5033 determines the current dirt level of each area based on the duration and mapping relationship of the dirt state of each area within the current statistical period, which may include:

[0128] When the area remains in a state of general dirtiness for a first preset time, the area is determined to be at the level of slight dirtiness.

[0129] When the area remains in a state of general dirtiness for a second preset time, the area is determined to be at a moderate level of dirtiness.

[0130] When the area remains in a state of general dirtiness for a third preset time, the area is determined to be at a relatively severe level of dirtiness.

[0131] When the area reaches the fourth preset time for a sustained state of severe contamination, the area is determined to be at the severe contamination level.

[0132] In this embodiment, each region is marked with a status indicator every 100ms, and the dirt status of all regions is statistically analyzed (no dirt, moderate dirt, or severe dirt). Based on the duration of the dirt status, moderate dirt can be further divided into several dirt levels: slight dirt, moderate dirt, and severe dirt. If the duration of dirt in a region exceeds p1% within time t, the region is classified as slightly dirty; if it exceeds p2%, it is classified as moderately dirty; if it exceeds p3%, it is classified as severely dirty; and if the duration of severe dirt reaches p4%, it is classified as severely dirty and reported. In this embodiment, p1% < p2% < p3%, and p4% can be greater than p2%.

[0133] For example, suppose a small area within region A is region Q, and one frame lasts 100ms. Assume t = 500ms (i.e., one statistical period is 5 frames). If the time spent in a dirty state exceeds 20%, meaning a dirty alarm occurs in 1 out of every 5 frames, it is classified as a light dirty level. If the time spent in a dirty state exceeds 40%, meaning a dirty alarm occurs in 2 out of every 5 frames, it is classified as a moderate dirty level. If the time spent in a dirty state exceeds 80%, meaning a dirty alarm occurs in 4 out of every 5 frames, it is classified as a severe dirty level. For areas already marked as severely dirty, if the time spent in a severely dirty state exceeds 60%, meaning a dirty alarm occurs in 3 out of every 5 frames, it is classified as a severely dirty level, and there is no need to set multiple thresholds for judgment.

[0134] In this embodiment, combining the time dimension can effectively reduce misjudgments. For example, if a flying insect or a speck of dust stays on the light window, the wind may blow it away in a very short time. In the next frame, the light window will not have a dirt problem. If the time dimension is not combined for judgment, the system will immediately alarm once it detects that the pulse width of the echo signal reaches the threshold, which will affect the user experience.

[0135] It should be noted that the thresholds W0, x, y, n, m, t, p1, p2, and p3 in this embodiment need to be determined through actual testing and requirements. They are first defined based on the level of dirtiness in different light windows, and then the values ​​of W0, x, y, n, m, t, p1, p2, and p3 are determined through testing. Their values ​​will vary under different light window types and different testing scenarios, and will also differ for different regions (central or non-central regions).

[0136] This disclosure also provides a cleaning device 700 for a lidar optical window, such as... Figure 7 As shown, it includes:

[0137] The acquisition module 701 is configured to acquire the dirt level corresponding to each area of ​​the lidar window; wherein the dirt level is acquired by the dirt detection method described in any of the above embodiments;

[0138] The strategy determination module 702 is configured to determine the cleaning strategy corresponding to each area based on the level of dirtiness. The cleaning strategy includes, but is not limited to, one or more of the following strategies: jetting, water spraying, and exiting the autopilot system.

[0139] Execution module 703 is configured to perform corresponding cleanup actions for each region according to the cleanup strategy.

[0140] Specifically, the cleaning strategies in this embodiment include, but are not limited to, at least one of the following: air jetting; water spraying; and disengaging the autonomous driving system. After the dirt level of each area of ​​each light window is obtained by dirt detection, for example, if a certain area is determined to be slightly dirty, moderately dirty, seriously dirty, or severely dirty, it can be reported to the vehicle control system. The strategy determination module 702 of the vehicle control system determines the corresponding cleaning strategy based on the current dirt level, so as to take reasonable cleaning measures for different dirt conditions, avoid unreasonable cleaning measures that result in incomplete cleaning of the light window, causing radar misjudgment and affecting driving safety, and at the same time avoid over-cleaning of the light window. For example, if the light window is only slightly dirty and does not affect driving, automatically disengaging the driving system before cleaning would affect the user experience.

[0141] For example, when the dirt is determined to be slightly dirty, the vehicle control system activates the self-cleaning system to spray air onto the area. After spraying, the system can reassess the level of dirt to determine whether the alarm needs to be deactivated. When the dirt is determined to be moderately dirty, the vehicle control system activates the self-cleaning system to spray water, and then determines whether the alarm needs to be deactivated. When the dirt is determined to be severely dirty, the vehicle needs to exit the autonomous driving system before spraying water, and then determines whether the alarm needs to be deactivated. When the dirt is determined to be severely dirty (such as obstructing radar point cloud data, which may seriously affect the safety performance of autonomous driving), for such large-area, long-term dirt, the vehicle needs to exit the autonomous driving system immediately; otherwise, a serious safety accident may occur due to system misjudgment. After exiting the autonomous driving system, water spraying and air spraying are performed, and finally, the system determines whether the alarm needs to be deactivated after cleaning. Through the above method, different cleaning strategies can be selected based on the accurate detected level of dirt and the dirty area, and the light window can be cleaned in a timely manner to avoid the radar misjudging pedestrians or crossing objects due to dirty light windows, causing the autonomous vehicle to brake suddenly or decelerate incorrectly, affecting driving safety.

[0142] The acquisition, storage, and application of user personal information involved in the technical solution disclosed herein comply with the provisions of relevant laws and regulations and do not violate public order and good morals.

[0143] According to embodiments of this disclosure, this disclosure also provides an electronic device, a readable storage medium, a computer program product, and an autonomous vehicle.

[0144] Specifically, electronic devices are intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. Electronic devices can also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the present disclosure described and / or claimed herein.

[0145] The device includes a computing unit that can perform various appropriate actions and processes based on a computer program stored in read-only memory (ROM) or loaded from a storage unit into random access memory (RAM). The RAM can also store various programs and data required for device operation. The computing unit, ROM, and RAM are interconnected via a bus. Input / output (I / O) interfaces are also connected to the bus.

[0146] Multiple components in the device are connected to the I / O interface, including: input units such as keyboards and mice; output units such as various types of displays and speakers; storage units such as disks and optical discs; and communication units such as network interface cards (NICs), modems, and wireless transceivers. The communication unit allows the device to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0147] The computing unit can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of computing units include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. The computing unit performs the various methods and processes described above, such as the method for detecting or cleaning a LiDAR window in the above embodiments. For example, in some embodiments, the method for detecting or cleaning a LiDAR window may be implemented as a computer software program tangibly contained in a machine-readable medium, such as a storage unit. In some embodiments, part or all of the computer program may be loaded into and / or installed on the device via ROM and / or a communication unit. When the computer program is loaded into RAM and executed by the computing unit, one or more steps of the method for detecting or cleaning a LiDAR window described above may be performed. Alternatively, in other embodiments, the computing unit may be configured to perform the method for detecting or cleaning a LiDAR window by any other suitable means (e.g., by means of firmware).

[0148] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-a-chip (SoCs), payload-programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.

[0149] The program code for implementing the dirt detection or cleaning method of the lidar window of this disclosure can be written in any combination of one or more programming languages. This program code can be provided to a processor or controller of a general-purpose computer, special-purpose computer, or other programmable data processing device, such that when executed by the processor or controller, the program code causes the functions / operations specified in the flowcharts and / or block diagrams to be implemented. The program code can be executed entirely on a machine, partially on a machine, as a standalone software package partially on a machine and partially on a remote machine, or entirely on a remote machine or server.

[0150] In the context of this disclosure, a machine-readable medium can be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium can be, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0151] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device for displaying information to the user (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor); and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the computer. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).

[0152] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as a data server), or computing systems that include middleware components (e.g., an application server), or computing systems that include frontend components (e.g., a user computer with a graphical user interface or web browser through which a user can interact with embodiments of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., a communication network). Examples of communication networks include local area networks (LANs), wide area networks (WANs), and the Internet.

[0153] Computer systems can include clients and servers. Clients and servers are generally located far apart and typically interact via communication networks. Client-server relationships are created by computer programs running on the respective computers and having a client-server relationship with each other. Servers can be cloud servers, servers in distributed systems, or servers incorporating blockchain technology.

[0154] It should be understood that the various forms of processes shown above can be used to rearrange, add, or delete steps. For example, the steps described in this disclosure can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution disclosed in this disclosure can be achieved, and this is not limited herein.

[0155] The specific embodiments described above do not constitute a limitation on the scope of protection of this disclosure. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this disclosure should be included within the scope of protection of this disclosure.

Claims

1. A method for detecting dirt in a lidar optical window, comprising: Emit a laser beam; Receive the echo signal generated by the laser beam; Based on the pulse width of the echo signal returned by each point of the lidar window, determine whether each point is a dirty spot. Specifically, this includes: comparing the pulse width of each echo signal with a preset reference pulse width, which includes a first preset pulse width and a second preset pulse width. Based on the comparison results, the level of dirtiness corresponding to different areas of the lidar window is determined; The method also includes: The lidar window is pre-divided into regions, namely a central region and a non-central region; Different first preset pulse widths and second preset pulse widths are set for the central region and the non-central region; wherein, the first preset pulse width corresponding to the central region is smaller than the first preset pulse width corresponding to the non-central region, and the second preset pulse width corresponding to the central region is smaller than the second preset pulse width corresponding to the non-central region.

2. The method of claim 1, wherein, The method further includes: The number of dirt spots in each area of ​​the lidar window is counted to determine the dirt status of each area; The level of dirtiness for each area is determined based on the duration of the dirtiness associated with that area.

3. The method of claim 2, the dirt points comprising general dirt points and severe dirt points; wherein, The method further includes: If the pulse width of the echo signal is greater than or equal to the first preset pulse width and less than the second preset pulse width, the point is determined to be the general dirt spot. In response to the pulse width of the echo signal being greater than or equal to the second preset pulse width, the point is determined to be the severely dirty spot; If the pulse width of the echo signal is less than the first preset pulse width, it is determined that the point is not a dirt spot.

4. The method according to claim 3, wherein, The step of counting the number of dirt spots in each area of ​​the lidar window and determining the dirt status of each area includes: The number of dirt spots in each area of ​​the lidar window is counted, and the total number of spots in each area of ​​the lidar window is determined. Calculate the percentage of dirt points in each area based on the number of dirt points in each area and the total number of dirt points. In response to the proportion of severely soiled spots in each area reaching a first preset ratio, the soiling status of that area is determined to be severely soiled. In response to the fact that the proportion of severely soiled spots in each area does not reach a first preset proportion, and the total proportion of general soiled spots and severely soiled spots reaches a second preset proportion, the soiling status of the area is determined to be a general soiled status.

5. The method according to claim 4, wherein, Before counting the number of dirt spots in each area of ​​the lidar window and determining the dirt status of each area, the method further includes: The lidar window is pre-divided into regions, namely a central region and a non-central region; Different first preset ratios and second preset ratios are set for the central region and the non-central region; wherein, the first preset ratio corresponding to the central region is smaller than the first preset ratio corresponding to the non-central region, and the second preset ratio corresponding to the central region is smaller than the second preset ratio corresponding to the non-central region.

6. The method according to claim 2, wherein, The step of determining the dirt level for each area based on the duration of the dirt state for each area includes: Multiple time thresholds are preset, and a mapping relationship is established between the multiple time thresholds and the multiple dirt levels; Based on the duration of the dirt state corresponding to each region within the current statistical period and the mapping relationship, the current dirt level of each region is determined.

7. The method according to claim 6, wherein, The step of determining the current dirt level of each region based on the duration of the dirt state of each region within the current statistical period and the mapping relationship includes: In response to the area being in a state of general dirtiness for a first preset time, the area is determined to be at a slightly dirty level. In response to the area remaining in the general state of dirt for a second preset time, the area is determined to be at a moderate level of dirt. In response to the area remaining in the general state of dirt for a third preset time, the area is determined to be at a relatively severe level of dirt. In response to the area being in a state of continuous severe dirtiness for a fourth preset time, the area is determined to be at a severe dirtiness level.

8. A method for cleaning a lidar optical window, comprising: The dirt level of each area of ​​the lidar window is obtained; wherein the dirt level is obtained by the dirt detection method according to any one of claims 1-7; Determine the cleaning strategy for each area based on the level of dirtiness; Perform corresponding cleaning actions for each area according to the cleaning strategy.

9. The cleaning method according to claim 8, wherein, The cleaning strategy includes at least one of the following: jetting air; spraying water; disengaging the autopilot system.

10. A dirt detection device for a lidar optical window, comprising: The transmitting module is configured to emit a laser beam; The receiving module is configured to receive the echo signal generated by the laser beam; The point detection unit is configured to determine whether each point is a dirty point based on the pulse width of the echo signal returned by each point of the lidar window. Specifically, it includes comparing the pulse width of each echo signal with a preset reference pulse width, which includes a first preset pulse width and a second preset pulse width. Based on the comparison results, the level of dirtiness corresponding to different areas of the lidar window is determined; The point detection unit is also used to pre-divide the lidar window into regions, dividing the lidar window into a central region and a non-central region. Different first preset pulse widths and second preset pulse widths are set for the central region and the non-central region; wherein, the first preset pulse width corresponding to the central region is smaller than the first preset pulse width corresponding to the non-central region, and the second preset pulse width corresponding to the central region is smaller than the second preset pulse width corresponding to the non-central region.

11. The apparatus according to claim 10, wherein, The device includes: The area detection unit is configured to count the number of dirt spots in each area of ​​the lidar window and determine the dirt status of each area. The dirt level determination unit is configured to determine the dirt level of each area based on the duration of the dirt state corresponding to each area.

12. The apparatus according to claim 11, wherein the dirt spots include general dirt spots and severe dirt spots; wherein, The point detection unit is also used for: If the pulse width of the echo signal is greater than or equal to the first preset pulse width and less than the second preset pulse width, the point is determined to be the general dirt spot. In response to the pulse width of the echo signal being greater than or equal to the second preset pulse width, the point is determined to be the severely dirty spot; If the pulse width of the echo signal is less than the first preset pulse width, it is determined that the point is not a dirt spot.

13. The apparatus according to claim 12, wherein, The area detection unit counts the number of dirt spots in each area of ​​the lidar window and determines the dirt status of each area, including: The number of dirt spots in each area of ​​the lidar window is counted, and the total number of spots in each area of ​​the lidar window is determined. Calculate the percentage of dirt points in each area based on the number of dirt points in each area and the total number of dirt points. In response to the proportion of severely soiled spots in each area reaching a first preset ratio, the soiling status of that area is determined to be severely soiled. In response to the fact that the proportion of severely soiled spots in each area does not reach a first preset proportion, and the total proportion of general soiled spots and severely soiled spots reaches a second preset proportion, the soiling status of the area is determined to be a general soiled status.

14. The apparatus according to claim 13, wherein, Before the area detection unit counts the number of dirt points in each area of ​​the lidar window and determines the dirt status of each area, it also includes: The lidar window is pre-divided into regions, namely a central region and a non-central region; Different first preset ratios and second preset ratios are set for the central region and the non-central region; wherein, the first preset ratio corresponding to the central region is smaller than the first preset ratio corresponding to the non-central region, and the second preset ratio corresponding to the central region is smaller than the second preset ratio corresponding to the non-central region.

15. The apparatus according to claim 11, wherein, The dirt level determination unit determines the dirt level for each area based on the duration of the dirt state in each area, including: Multiple time thresholds are preset, and a mapping relationship is established between the multiple time thresholds and the multiple dirt levels; Based on the duration of the dirt state corresponding to each region within the current statistical period and the mapping relationship, the current dirt level of each region is determined.

16. The apparatus according to claim 15, wherein, The dirt level determination unit determines the current dirt level of each region based on the duration of the dirt state corresponding to each region within the current statistical period and the mapping relationship, including: In response to the area being in a state of general dirtiness for a first preset time, the area is determined to be at a slightly dirty level. In response to the area remaining in the general state of dirt for a second preset time, the area is determined to be at a moderate level of dirt. In response to the area remaining in the general state of dirt for a third preset time, the area is determined to be at a relatively severe level of dirt. In response to the area being in a state of continuous severe dirtiness for a fourth preset time, the area is determined to be at a severe dirtiness level.

17. A cleaning device for a lidar optical window, comprising: The acquisition module is configured to acquire the dirt level corresponding to each area of ​​the lidar window; wherein the dirt level is acquired by the dirt detection method according to any one of claims 1-7; The strategy determination module is configured to determine the cleaning strategy corresponding to each area based on the level of dirtiness. The execution module is configured to perform corresponding cleanup actions for each region according to the cleanup strategy.

18. The cleaning apparatus according to claim 17, wherein, The cleaning strategy includes at least one of the following: jetting air; spraying water; disengaging the autopilot system.

19. An electronic device comprising: At least one processor; as well as A memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor to enable the at least one processor to perform the method of any one of claims 1-9.

20. A non-transitory computer-readable storage medium storing computer instructions, wherein, The computer instructions are used to cause the computer to perform the method according to any one of claims 1-9.

21. A computer program product comprising a computer program that, when executed by a processor, implements the method according to any one of claims 1-9.

22. An autonomous vehicle comprising the electronic device of claim 19.

Images