KR102389466B1 - Method for detecting contamination of lidar sensor and apparatus using the same - Google Patents

Abstract

The present invention relates to a method of detecting the contamination of a LIDAR sensor. The method includes: a step (a) of acquriing, by a contamination detector, (L_1)^th monitoring data to (L_n)^th monitoring data through each of (L_1)^th layer to (L_n)^th layer determined as a layer interacting with the ground among a plurality of layers of the LIDAR sensor, and then, determining whether at least some of each of the (L_1)^th monitoring data to (L_n)^th monitoring data are ground data or non-ground data; and a step (b) of determining, by the contamination detector, whether at least some of accumulated counts of (L_1)^th non-ground data to (L_n)^th non-ground data included in each of the (L_1)^th monitoring data to (L_n)^th monitoring data exceed an (L_1)^th threshold count, by performing the step (a) multiple times with the passage of time, and then, determining that at least one (L_)^th specific layer corresponding to at least one piece of specific non-ground data determined as one exceeding the (L_1)^th threshold count, is contaminated. Therefore, the present invention is capable of accurately detecting the contamination of the LIDAR sensor within a short time.

Description

A method for detecting contamination of a lidar sensor and an apparatus using the same

The present invention relates to a method for detecting contamination of a lidar sensor and an apparatus using the same.

As the development of modern technology enables autonomous driving, people's interest in autonomous driving devices is rapidly increasing. However, due to the nature of autonomous driving technology, trust in the technology is absolutely required. Trust in technology here means that humans completely delegate the authority to drive to the machine. do.

However, the lidar sensor, which is a core technology for autonomous driving, is usually mounted outside due to the nature of using light. Therefore, they are exposed to various pollution situations such as rain, dust, and insects. Since such contamination is equivalent to blinding the autonomous driving device, trust in the above technology can be guaranteed only when the autonomous driving device is operating normally by accurately detecting and removing the contamination from the stomach.

In order to solve this problem, in the prior art, various techniques for cleaning the lidar sensor have been introduced. However, since it is only meaningful to accurately detect contamination before cleaning the lidar sensor, even if excellent cleaning technology is introduced, if the contamination is not accurately detected, the autonomous driving device may not operate normally. There has still been a problem in that trust in technology cannot be guaranteed.

In particular, in addition to contamination such as insects and paper that directly interferes with the interaction of the lidar sensor because light does not pass through, contamination is detected in real time even when the reception sensitivity is lowered, such as water droplets and stains, which causes performance degradation in perceived distance and accuracy. A method capable of detecting has been desired.

After all, in order to fully operate autonomous driving technology, there is a need for a method or device capable of accurately detecting contamination of a lidar sensor in real time in various contamination situations.

An object of the present invention is to solve all of the above problems.

Another object of the present invention is to provide a method and apparatus capable of accurately detecting contamination of a lidar sensor.

Another object of the present invention is to provide a method and apparatus capable of accurately detecting contamination of a lidar sensor within a short time by using ground data that has a very low reflection intensity and is always measured.

In addition, it is another object to provide an efficient detection method and apparatus by setting a lidar ground estimation area with reference to a part of the three-dimensional shape of the mounting body where the lidar sensor is installed, and the installation position, angle, and lidar sensor. do.

In addition, another object of the present invention is to provide an efficient detection method and apparatus by setting a region of a LIDAR object estimation candidate group using other sensor data.

In addition, it is another aspect to provide a method and apparatus for detecting contamination of LiDAR by setting an area in a two-dimensional array consisting of a layer and a viewing angle or a two-dimensional array consisting of a viewing angle and a separation distance and using the cell included therein. The purpose.

In order to achieve the object of the present invention as described above and to realize the characteristic effects of the present invention to be described later, the characteristic configuration of the present invention is as follows.

According to an aspect of the present invention, in a method for detecting contamination of a lidar sensor, (a) the contamination detection device L_1 th layer to L_n th layer determined to interact with the ground among a plurality of layers of the lidar sensor obtaining L_1 th monitoring data to L_n th monitoring data through each, and determining whether at least a portion of each of the L_1 th monitoring data to L_n th monitoring data is ground data or non-ground data; and (b) the pollution detection device performs step (a) a plurality of times according to time, so that the L_1th monitoring data to L_nth empty ground data included in each of the L_1th monitoring data to the L_nth monitoring data are obtained. It is determined whether at least a part of the accumulated count exceeds the L_1th threshold count, and at least one L_th layer corresponding to the at least one specific non-ground data determined to exceed the L_1th threshold count is contaminated with Disclosed is a method comprising the step of determining that this has occurred.

As an example, in step (b), step (a) is performed a plurality of times according to time, and the L_1th monitoring data to the L_nth empty page data included in each of the L_1th monitoring data to the L_nth monitoring data are performed. In determining whether at least a portion of the accumulated count of data exceeds the L_1 th threshold count, the pollution detection device determines that the ground data is included among the L_1 th monitoring data to the L_n th monitoring data Disclosed is a method characterized by initializing the accumulated count corresponding to at least some of the monitoring data.

As an example, in step (b), each of the L_1 th layer to the L_n th layer subdivides the observable viewing angle range from the angular range corresponding to the omnidirection in the space of the lidar sensor to a predetermined angle. Each of the plurality of generated cells is included, wherein the L_k-th layer, which is any one of the L_1th layer to the L_nth layer, includes L_k_1th cells to L_k_mk cells, and the pollution detection device performs the contamination detection device through the L_kth layer. The accumulated count of the L_kth busy data is calculated for each cell of the L_k_1th cell to the L_k_mk cell of the obtained L_kth monitoring data, and the accumulated count of at least one L_k_pth cell among the L_k_1th cell to the L_k_mk cell is Disclosed is a method comprising determining whether a threshold count for each cell is exceeded, and determining that the L_k_p-th cell determined to exceed the threshold count for each cell is contaminated.

As an example, in step (b), step (a) is performed a plurality of times according to time so that the cumulative count of at least one of the L_k_1 th to the L_k_mk cells exceeds the threshold count for each cell. In determining whether or not to do this, the pollution detection device initializes the accumulated count of at least one L_k_p-th cell determined to include the ground data among the L_k_1 th cell to the L_k_mk cell. This is initiated.

As an example, the non-ground data may include (i) L_1 th type object data including information about an object detected at less than a predetermined threshold distance by at least one of the L_1 th layer to the L_n th layer, ( ii) L_2 th type object data including information on an object sensed at a predetermined threshold distance or more by at least one of the L_1 th layer to the L_n th layer, and (iii) the L_1 th type object data or At least among non-object data that is not the L_2 type object data and that at least a part of the average and variance of the pattern of the reflection data included in the ground data is different by at least a part of the preset threshold average and the critical variance Disclosed is a method comprising a portion.

As an example, before the step (a), (a0) the contamination detection device, the installation position of the lidar sensor, the installation angle, and at least a part of the three-dimensional shape of the mounting body in which the lidar sensor is installed as a reference to set an observable viewing angle range of the lidar sensor and a lidar ground estimation area corresponding to the L_1 th layer to the L_n th layer , obtains the L_1 th monitoring data to the L_n th monitoring data through each of the L_1 th layer to the L_n th layer with reference to the LIDAR ground estimation area, and the L_1 th monitoring data to the L_n th monitoring data Disclosed is a method comprising determining whether at least a portion of each of the data is the ground data or the non-ground data.

As an example, in step (a), the pollution detection device uses the U_1th monitoring data to the U_th monitoring data through each of the U_1th layer to the U_sth layer which is determined not to interact with the ground among the plurality of layers of the lidar sensor. U_s monitoring data is additionally acquired, and at least a portion of each of the U_1th monitoring data to the U_sth monitoring data is detected by at least one of the U_1th layer to the U_sth layer at a predetermined threshold distance or more. It is additionally determined whether the data is U-th type object data including information on At least it is determined that at least some points of the U-th object data included in at least some of the U_1 th monitoring data to the U_s th monitoring data exceed the second threshold count among the accumulated counts corresponding to the number of missing points. Disclosed is a method comprising additionally determining that at least one U_th specific layer corresponding to one specific point is contaminated.

According to another aspect of the present invention, in the method for detecting contamination of a lidar sensor, (a) the contamination detection device is determined not to interact with the ground among the plurality of layers of the lidar sensor U_1 th to U_s th layer acquiring U_1 th monitoring data to U_s th monitoring data through each layer; and (b) the pollution detection device performs step (a) multiple times over time, thereby acquiring at least a portion of the U_1th monitoring data to the U_sth monitoring data multiple times over time to monitor the U_1th Whether or not it exceeds the U_1th threshold count among accumulated counts, which is the number of missed counts for each specific point that is at least some of the missing detection points among all points corresponding to the target object data included in at least some of the data to the U_s monitoring data determining that at least some specific points determined to have exceeded the U_1th threshold count as at least one problem point and determining that at least one U_th specific layer corresponding to the problem point is contaminated A method comprising

As an example, in the step (b), when the pollution detection device determines whether at least a portion of the accumulated count exceeds the U_1th threshold count, among the U_1th monitoring data to the U_sth monitoring data, the At least some specific points that are determined to have been detected again by at least some of the U_1 th monitoring data to the U_s th monitoring data for each specific point that is at least a partial point from which detection is missing among all the points corresponding to the target object data For points, a method is disclosed, characterized in that the accumulated count is initialized.

As an example, before the step (a), (a1) the contamination detection device, with respect to the U_k-th layer, which is any one of the U_1 th layer to the U_s th layer, omnidirectional in space of the lidar sensor (omnidirection) by subdividing the observable viewing angle range into a predetermined angle to form a first axis, and subdividing the separation distance from the lidar sensor into a predetermined unit distance to form a second axis, thereby forming a plurality of cells generating, and determining, from among the plurality of cells, at least one specific cell corresponding to the target object data, wherein the step (a) includes, in the step (a), the contamination detection device, the U_1 th monitoring data to the It is characterized in that at least a part of specific monitoring data corresponding to the specific cell is determined from among the U_sth monitoring data, and in step (b), the pollution detection device includes at least one corresponding to each of the specific cells for each specific cell. It is determined whether the accumulated count for the number of missed counts of a specific point for each cell, which is a point for which detection is missing among points for each cell, exceeds a threshold count for each cell, and contamination for some specific cells determined to exceed the threshold count for each cell Disclosed is a method characterized in that it is determined that this has occurred.

As an example, before step (a1), (a0) the contamination detection device additionally acquires other sensor data from another sensor (the other sensor includes at least some of an image sensor and a radar sensor) The method further includes a step, wherein the step (a1) is performed with respect to the U_k-th layer, which is any one of the U_1-th layer to the U_s-th layer, with reference to the other sensor data, the specific corresponding to the target object data A method comprising performing a process of determining a cell is disclosed.

As an example, in step (b), the accumulated count for the number of missed counts of the specific point for each cell, which is a point for which detection is missing, among at least one point for each cell corresponding to each of the specific cells for each specific cell is the threshold count for each cell. In determining whether to exceed Disclosed is a method characterized by initializing the accumulated count for each cell corresponding to .

As an example, in the step (b), with the specific cell as a target, the target point corresponding to the target cell among consecutive cells arranged on the first axis in a state where the separation distance on the second axis is constant a process of counting, by the contamination detection apparatus, the number of missed counts of the target point corresponding to the target cell, when detection of a neighboring point corresponding to an adjacent neighboring cell of the target cell is not detected; and When the specific cell is not detected for all target points corresponding to all consecutive cells arranged on the first axis while the separation distance on the second axis is constant, the contamination detection device is, a method characterized in that performing at least a part of the process of counting the number of missed times of the entire target point corresponding to the entire target cell is disclosed.

As an example, the U_1 th monitoring data to the U_s th monitoring data include (i) information about an object detected at less than a predetermined threshold distance by at least one of the U_1 th layer to the U_s th layer. U_1 th type object data, (ii) U_2 th type object data (the U_2 th type) including information on an object detected at a predetermined threshold distance or more by at least one of the U_1 th layer to the U_s th layer At least a portion of the object data is the target object data) and (iii) a non-object that affects the detection of a point that is not the U_1 th type object data or the U_2 th type object data and corresponds to the U_2 type object data A method comprising at least some of the data is disclosed.

According to another aspect of the present invention, there is provided an apparatus for detecting contamination of a lidar sensor, comprising: at least one memory storing instructions; and at least one processor configured to execute the instructions, wherein the processor includes (I) the L_1 th layer to the L_n th layer that is determined to interact with the ground among a plurality of layers of the lidar sensor. a process of obtaining L_1 monitoring data to L_n monitoring data, and determining whether at least a portion of each of the L_1 monitoring data to L_n monitoring data is ground data or non-ground data; and (II) at least a portion of the accumulated counts of the L_1th monitoring data to the L_nth busy data included in each of the L_1th monitoring data to the L_nth monitoring data by performing the process (I) a plurality of times over time. Process of determining whether the L_1th threshold count is exceeded, and determining that at least one L_th layer corresponding to the at least one specific non-ground data determined to exceed the L_1th threshold count is contaminated An apparatus for performing is disclosed.

As an example, in the process (II), the processor performs the process (I) a plurality of times according to time, so that the L_1th unoccupied data to the L_1th monitoring data included in each of the L_1th monitoring data to the L_nth monitoring data In determining whether at least a portion of the accumulated counts of the L_nth non-paper data exceeds the L_1th threshold count, at least a portion of the L_1th monitoring data to the L_nth monitoring data that is determined to include the ground data Disclosed is an apparatus characterized by initializing the accumulated count corresponding to the monitoring data of .

As an example, in the (II) process, each of the L_1 th layer to the L_n th layer subdivides the observable viewing angle range among the angle ranges corresponding to the omnidirection in the space of the lidar sensor into a predetermined angle. Each of the generated cells is included, wherein the L_k-th layer, which is any one of the L_1th layer to the L_nth layer, includes the L_k_1th cell to the L_k_mk cell, and the processor obtains the L_kth layer through the An accumulated count of L_k-th busy data is calculated for each L_k-th cell to the L_k_mk cell of the L_k-th monitoring data, and the accumulated count of at least one L_k_p cell among the L_k_1th cell to the L_k_mk cell is a threshold for each cell. Disclosed is an apparatus characterized in that it is determined whether the count is exceeded, and the L_k_p cell determined to exceed the threshold count for each cell is determined to be contaminated.

As an example, in the (II) process, the (I) process is performed a plurality of times according to time so that the accumulated count of at least one of the L_k_1 th to the L_k_mk cells exceeds the threshold count for each cell In determining whether or not to, the processor initializes the accumulated count of at least one L_k_p-th cell determined to include the ground data among the L_k_1 th cell to the L_k_mk cell do.

As an example, the non-ground data may include (i) L_1 th type object data including information about an object detected at less than a predetermined threshold distance by at least one of the L_1 th layer to the L_n th layer, ( ii) L_2 th type object data including information on an object sensed at a predetermined threshold distance or more by at least one of the L_1 th layer to the L_n th layer, and (iii) the L_1 th type object data or At least among non-object data that is not the L_2 type object data and that at least a part of the average and variance of the pattern of the reflection data included in the ground data is different by at least a part of the preset threshold average and the critical variance Disclosed is an apparatus comprising a part.

As an example, before the (I) process, (I-0) the processor, with reference to at least a part of the installation position, the installation angle, and the three-dimensional shape of the mounting body in which the lidar sensor is installed Thus, a process of setting an observable viewing angle range of the lidar sensor and a lidar ground estimation area corresponding to the L_1 th layer to the L_n th layer is further performed; The L_1 th monitoring data to the L_n th monitoring data are obtained through each of the L_1 th layer to the L_n th layer with reference to the LIDAR ground estimation area, and the L_1 th monitoring data to the L_n th monitoring data, respectively and performing a process of determining whether at least a portion of the data is the ground data or the non-ground data.

As an example, in the (I) process, the processor monitors the U_1th monitoring data to the U_s through each of the U_1th layer to the U_sth layer determined not to interact with the ground among the plurality of layers of the lidar sensor. Data is further obtained, and at least a portion of each of the U_1 th monitoring data to the U_s th monitoring data is detected at a predetermined threshold distance or more by at least one of the U_1 th layer to the U_s layer. It is further determined whether the data is U-th type object data including information, and in the process (II), the processor acquires at least a portion of the U_1-th monitoring data to the U_s-th monitoring data multiple times over time to obtain the U_1th At least one specific point determined to exceed a second threshold count among accumulated counts corresponding to the number of times that at least some points of the U-th type object data included in each of at least some of the monitoring data to the U_s monitoring data have been omitted Disclosed is an apparatus characterized in that it is further determined that at least one U_th specific layer corresponding to U_ is contaminated.

According to another aspect of the present invention, there is provided an apparatus for detecting contamination of a lidar sensor, comprising: at least one memory storing instructions; and at least one processor configured to execute the instructions, wherein the processor includes (I) a U_1 th layer to a U_s th layer that is determined not to interact with the ground among a plurality of layers of the lidar sensor. a process of acquiring U_1th monitoring data to U_sth monitoring data; and (II) obtaining at least a portion of the U_1th monitoring data to the U_sth monitoring data multiple times over time by performing the (I) process multiple times over time to monitor the U_1th monitoring data to the U_sth monitoring data It is determined whether or not the U_1th threshold count is exceeded among accumulated counts, which is the number of missed counts for each specific point, which is at least some of the missing detection points among all points corresponding to the target object data included in at least a portion of the data, and the U_1th An apparatus for performing a process of determining that at least some specific points determined to have exceeded a threshold count as at least one problem point to be contaminated with at least one U_th specific layer corresponding to the problem point is disclosed do.

As an example, in the process (II), when the processor determines whether at least a portion of the accumulated count exceeds the U_1 th threshold count, the target object from among the U_1 th monitoring data to the U_s th monitoring data At least some specific points that are determined to have been detected again by at least some of the U_1 th monitoring data to the U_s th monitoring data for each specific point that is at least some points that are missing detection among all points corresponding to the data , an apparatus is disclosed for initializing the accumulated count.

As an example, before the (I) process, (I-1) the processor, with respect to the U_k-th layer, which is any one of the U_1 th layer to the U_s th layer, omnidirectional in space of the lidar sensor (omnidirection) by subdividing the observable viewing angle range into a predetermined angle to form a first axis, and subdividing the separation distance from the lidar sensor into a predetermined unit distance to form a second axis, thereby forming a plurality of cells generating, and further performing a process of determining at least one specific cell corresponding to the target object data from among the plurality of cells, wherein the (I) process includes, by the processor, the U_1 th monitoring data to the U_s th It is characterized in that at least a part of the specific monitoring data corresponding to the specific cell is determined from among the monitoring data, wherein, in the (II) process, the processor, for each specific cell, at least one point for each cell corresponding to each of the specific cells It is determined whether the accumulated count for the number of missed counts of a specific point per cell, which is a point for which detection is missing, exceeds a threshold count for each cell, and some specific cells determined to exceed the threshold count for each cell are judged to be contaminated Disclosed is an apparatus characterized in that

As an example, before the (I-1) process, (I-0) the processor further receives other sensor data from another sensor (the other sensor includes at least some of an image sensor and a radar sensor) Further performing a process of obtaining, the (I-1) process is, with respect to the U_k-th layer, which is any one of the U_1 th layer to the U_s th layer, with reference to the other sensor data, to the target object data Disclosed is an apparatus for performing a process for determining a corresponding specific cell.

As an example, in the (II) process, the accumulated count for the number of missed counts of the specific point for each cell, which is a point for which detection is missing, among at least one point for each cell corresponding to each of the specific cells for each specific cell is the threshold count for each cell. In determining whether to exceed Disclosed is an apparatus characterized by initializing the accumulated count for each cell.

As an example, in the process (II), the specific cell is targeted at a target point corresponding to the target cell among consecutive cells arranged on the first axis in a state where the separation distance on the second axis is constant. a process of counting, by the processor, the number of missed times of the target point corresponding to the target cell, when detection of a neighboring point corresponding to an adjacent neighboring cell of the target cell is not detected, and the specific When a cell is not detected for all target points corresponding to all consecutive cells arranged on the first axis while the separation distance on the second axis is constant, it corresponds to the entire target cell and performing at least a part of a process of counting the number of missed counts of the total target points being

As an example, the U_1 th monitoring data to the U_s th monitoring data include (i) information about an object detected at less than a predetermined threshold distance by at least one of the U_1 th layer to the U_s th layer. U_1 th type object data, (ii) U_2 th type object data (the U_2 th type) including information on an object detected at a predetermined threshold distance or more by at least one of the U_1 th layer to the U_s th layer At least a portion of the object data is the target object data) and (iii) a non-object that affects the detection of a point that is not the U_1 th type object data or the U_2 th type object data and corresponds to the U_2 type object data An apparatus comprising at least a portion of data is disclosed.

The present invention has the effect of providing a method and apparatus capable of accurately detecting contamination of a lidar sensor.

In addition, the present invention has an effect of providing a method and apparatus capable of accurately detecting contamination of a lidar sensor within a short time by using ground data that has a very low reflection intensity and is always measured.

In addition, the present invention has the effect of providing an efficient detection method and apparatus by setting a lidar ground estimation area with reference to a part of the three-dimensional shape of the mounting body in which the lidar installation position, angle, and lidar sensor is installed. there is.

In addition, the present invention has the effect of providing an efficient detection method and apparatus by setting the region of the LIDAR object estimation candidate group using other sensor data.

In addition, the present invention provides a method and apparatus for setting an area in a two-dimensional array consisting of a layer and a viewing angle or a two-dimensional array consisting of a viewing angle and a separation distance and accurately detecting contamination of the lidar using the cell included therein. has the effect of

The accompanying drawings for use in the description of the embodiments of the present invention are only a part of the embodiments of the present invention, and for those of ordinary skill in the art to which the present invention pertains (hereinafter, "those of ordinary skill in the art"), the invention Other drawings may be obtained based on these drawings without any work being done.
1 is a diagram illustrating an apparatus for detecting contamination of a lidar sensor according to an embodiment of the present invention.
2A is an L_1 th layer to L_n th layer determined to interact with the ground among a plurality of layers of a lidar sensor according to an embodiment of the present invention, and a U_1 th layer to U_s layer determined not to interact with the ground. 2B is a diagram illustrating L_1 th monitoring data to L_n th monitoring data obtained through each of the L_1 th layer to L_n th layer.
3 is a diagram illustrating a flowchart of a process for detecting contamination of a lidar sensor according to an embodiment of the present invention.
4 is a diagram illustrating in more detail a flowchart of a process for detecting contamination of a lidar sensor according to an embodiment of the present invention.
5 is a diagram illustrating concentric circles of a lidar sensor according to various positions of an installed vehicle in setting a ground estimation area with reference to an installation angle, height, and mounting body of the lidar according to an embodiment of the present invention; It is a drawing.
6A to 6C are diagrams specifically illustrating a process of accumulating the number of non-detection counts of ground data for each cell for a ground estimation area set in a two-dimensional array including layers and viewing angles according to an embodiment of the present invention.
7 is a diagram illustrating a flowchart of a process for detecting contamination of a lidar sensor according to another embodiment of the present invention.
8 is a diagram illustrating in more detail a flowchart of a process for detecting contamination of a lidar sensor according to another embodiment of the present invention.
9 is a diagram illustrating setting of an object estimation candidate group region using other sensor data according to another embodiment of the present invention.
10 is a diagram illustrating that a specific point is not detected according to another embodiment of the present invention.
11A and 11B illustrate a process of receiving other sensor-based data, setting an object estimation candidate group region in a two-dimensional array composed of a viewing angle and a separation distance, and accumulating the number of times of non-detection of a specific point for each cell, according to another embodiment of the present invention; It is a drawing expressing more specifically.

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS [0012] DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS [0010] DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS [0010] Reference is made to the accompanying drawings, which show by way of illustration specific embodiments in which the present invention may be practiced. These embodiments are described in sufficient detail to enable those skilled in the art to practice the present invention. It should be understood that the various embodiments of the present invention are different but need not be mutually exclusive. For example, certain shapes, structures, and characteristics described herein with respect to one embodiment may be implemented in other embodiments without departing from the spirit and scope of the invention. In addition, it should be understood that the location or arrangement of individual components within each disclosed embodiment may be changed without departing from the spirit and scope of the present invention. Accordingly, the detailed description set forth below is not intended to be taken in a limiting sense, and the scope of the present invention, if properly described, is limited only by the appended claims, along with all scope equivalents as those claimed. Like reference numerals in the drawings refer to the same or similar functions throughout the various aspects.

Hereinafter, in order to enable those skilled in the art to easily practice the present invention, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings.

1 is a view showing an apparatus for detecting contamination of a lidar sensor according to the present invention.

The memory 110 of the contamination detection device 100 may store instructions to be executed by the processor 120 , specifically, the instructions are generated for the purpose of causing the contamination detection device 100 to function in a particular manner. code that may be stored in a computer-usable or computer-readable memory that may point to a computer or other programmable data processing equipment. The instructions may perform processes for performing the functions described herein.

In addition, the processor 120 of the pollution detection apparatus 100 may include a hardware configuration such as a micro processing unit (MPU) or a central processing unit (CPU), a cache memory, and a data bus. . In addition, the operating system may further include a software configuration of an application for performing a specific purpose.

Also, the pollution detection apparatus 100 may be linked with a database (not shown). Here, the database (not shown) includes a flash memory type, a hard disk type, a multimedia card micro type, and a card type memory (eg, SD or XD memory). , Random Access Memory (RAM), Static Random Access Memory (SRAM), ReadOnly Memory, ROM, Electrically Erasable Programmable ReadOnly Memory (EEPROM), Programmable ReadOnly Memory (PROM), Magnetic Memory, Magnetic Disk, Optical Disk It may include at least one type of storage medium, but is not limited thereto, and may include any medium capable of storing data. In addition, the database (not shown) may be installed separately from the contamination detection device 100 , or alternatively, it may be installed inside the contamination detection device 100 to transmit data or record received data. It may be implemented separately from two or more, which may vary depending on the operating conditions of the invention.

In addition, the pollution detection apparatus 100 may be included in the lidar sensor. However, the present invention is not limited thereto and may exist outside the lidar sensor. As a specific example, a computer installed in a vehicle separately from a lidar sensor mounted in the vehicle may be assumed.

2A is an L_1 th layer to L_n th layer determined to interact with the ground among a plurality of layers of a lidar sensor according to an embodiment of the present invention, and a U_1 th layer to U_s layer determined not to interact with the ground. 2B is a diagram illustrating L_1 th monitoring data to L_n th monitoring data obtained through each of the L_1 th layer to L_n th layer.

Referring to FIG. 2A , from among the plurality of layers of the lidar sensor, L_1 th to L_n th layers 210 determined to interact with the ground and U_1 th to U_s layers 220 determined not to interact with the ground can be checked. Here, “interaction with the ground” refers to a state in which the light emitted through the lidar is reflected on the ground and received reflected light exists. Even if not, if there is reflected light received through a neighboring cell or a neighboring layer, it may be regarded as a layer that “interacts with the ground”. In some cases, a layer that emits light toward the ground (including those directed diagonally as well as perpendicular to the ground), with or without reflected light, may be treated as a layer that “interacts with the ground”. will be.

Referring to FIG. 2B , it is possible to check L_1 th monitoring data to L_n th monitoring data obtained through each of the L_1 th layer to L_n th layer,

As will be described in detail later, this is object data 240 sensed at a predetermined threshold distance or more by at least one of the L_1 th ground data to L_n th ground data 230 and the L_1 th layer to the L_n th layer. can be checked Here, the object data 240 may be L_2 type object data, which will be described later.

3 is a diagram illustrating a flowchart of a process for detecting contamination of a lidar sensor according to an embodiment of the present invention.

Referring to FIG. 3 , the pollution detection apparatus 100 monitors L_1 th monitoring data to L_n th monitoring data through each of the L_1 th layer to L_n th layer 210 determined to interact with the ground among the plurality of layers of the lidar sensor. Acquire data, determine whether at least a portion of each of the L_1 th monitoring data to the L_n th monitoring data is ground data or non-ground data (S310), and the pollution detection device performs the step S310 according to time It is performed a plurality of times to determine whether at least some of the accumulated counts of the L_1th monitoring data to the L_nth busy data included in each of the L_1th monitoring data to the L_nth monitoring data exceed the L_1th threshold count, and the It may be determined that at least one L_th layer corresponding to the at least one specific non-ground data determined to exceed the L_1th threshold count is contaminated ( S320 ).

Here, the non-ground data includes (i) L_1 th type object data including information on an object detected at less than a predetermined threshold distance by at least one of the L_1 th layer to the L_n th layer, (ii) ) L_2 th type object data including information on an object detected at a predetermined threshold distance or more by at least one of the L_1 th layer to the L_n th layer, and (iii) the L_1 th type object data or the At least a part of non-object data that is not the L_2 type object data and that at least a part of the average and variance of the pattern of reflection data included in the ground data is different by at least some of the preset threshold average and critical variance may include

As a specific example, the L_1 type object data includes bird droppings, leaflets, or dust on the surface of the lidar sensor, and the L_2 type object data corresponds to the object 240 of FIG. 2B. Examples include stones, carts on the road, or the wheels of vehicles traveling together. In addition, at least a part of the average and variance of the pattern of the reflection data included in the ground data that is not the L_1 type object data or the L_2 type object data is at least a part of the preset threshold average and the critical variance. Non-object data that is judged to be present is affected by a pollutant that lowers reception sensitivity, such as water droplets and stains, and is detected as contaminated if there is a difference between the reflected data pattern on the ground and the reflected data pattern over a certain level. Here, the reflection data pattern of the ground may be different depending on the type of the ground, and for example, a predetermined reflection data pattern may be managed in a predetermined database depending on whether the ground is a gravel field, a loess road, or asphalt.

This will be described in more detail with reference to FIG. 4 as follows.

4 is a diagram illustrating in more detail a flowchart of a process for detecting contamination of a lidar sensor according to an embodiment of the present invention.

Referring to FIG. 4 , the L_1 th layer to L_n th layer 210 determined to interact with the ground acquires L_1 th monitoring data to L_n th monitoring data (S410), and L_1 th monitoring data to L_n th monitoring data It may be determined whether each corresponds to the ground data (S420). At this time, if there is no object at a distance shorter than the distance from which the ground is measured, a pattern identical or similar to the reflection pattern of the ground data will be detected, and accordingly, it is determined that the ground data is the ground data (S421), and a distance shorter than the distance from which the ground is measured If there is an object in , a different pattern that is different from the average and dispersion of the reflection pattern of the ground data by more than a predetermined threshold average and critical dispersion will be detected, and accordingly, it can be determined as non-ground data (S422) .

On the other hand, in step S430 of FIG. 4 , the contamination detection device performs the lidar with reference to at least a part of the installation position, the installation angle, and the three-dimensional shape of the mounting body in which the lidar sensor is installed. It is possible to set the observable viewing angle range of the sensor and the LiDAR ground estimation area corresponding to the L_1 th layer to the L_n th layer.

Specifically, depending on the environment in which the lidar sensor is installed, the observable viewing angle of the lidar sensor changes and the layer interacting with the ground changes, so this is taken into account. In addition, a vehicle or a power pole may be mentioned as an example of the mounting body in which the lidar sensor is installed, but the 3D shape part of the mounting body, which is always obtained by the lidar sensor, is not recognized as a foreign material. It may be assumed to set the estimation area. In addition, this may be applied to other embodiments of the present invention, which will be described with reference to FIGS. 7 and below, which will be described later.

Specifically, FIG. 5 shows the lidar sensor according to various installation positions on the vehicle in setting the ground estimation area with reference to the installation angle, height, and mounting body of the lidar according to an embodiment of the present invention. It is a diagram expressing a virtual concentric circle with It can be seen that the LIDAR ground estimation area corresponding to the L_nth layer is changed. A yellow area 610 of FIG. 6A , which will be described later, is an example of such a ground estimation area.

Referring back to FIG. 4 , the non-detection state of the ground data is detected for each cell included in the two-dimensional array composed of layers and viewing angles with the ground estimation area set in step S430 as a target (S440), and the ground The number of times of data non-detection may be counted and accumulated (S450). Next, if it is determined that the accumulated number of ground data non-detection determinations of a specific cell is greater than a predetermined threshold count (S460), it is determined that the specific cell is contaminated (S461), and the accumulated number of ground data non-detection determinations of the specific cell is If it is determined that it is not greater than a predetermined threshold count (S460), it may be determined that the specific cell is not contaminated (S462).

This will be described in more detail with reference to FIGS. 6A to 6C.

6A to 6C are diagrams specifically expressing the process of accumulating the number of times of non-detection of ground data for each cell with respect to a ground estimation area on a two-dimensional array composed of a layer of a lidar sensor and a viewing angle according to an embodiment of the present invention.

As mentioned above, the yellow area visible at reference numeral 610 in FIG. 6A may correspond to the LiDAR ground estimation area. The L_1 th monitoring data to the L_n th monitoring data are obtained through each of the L_1 th layer to the L_n th layer with reference to the LIDAR ground estimation area determined as described above, and the L_1 th monitoring data to the L_n th monitoring data are obtained. It will be determined whether at least a portion of each of the monitoring data is the ground data or the non-ground data.

Referring to FIG. 6B , it can be seen that the area 620 in which object data is detected is displayed in red by targeting the ground estimation area 610 .

Also, referring to FIG. 6C , when non-ground data is acquired at predetermined time intervals in the L_1 th layer to the L_n th layer from which the ground data is to be acquired, the accumulated number of times is counted and the threshold count is exceeded. _ Determining that a specific layer is contaminated.

As an example of determining whether the threshold count is exceeded for each layer, L_5th monitoring data is acquired once per second from the L_5th layer. When the threshold count corresponding to the L_5th layer is 5, L_5th ungrounded data When is accumulated 6 times, it is determined that the L_5th layer is contaminated. Here, the criterion for determining whether the accumulated count exceeds the threshold count may mean, but is not limited to, determining whether the continuously accumulated count exceeds the threshold count, and in some cases, during a preset time period It may mean determining whether a ratio of the number of times of acquiring unoccupied data among the number of acquisitions of the acquired monitoring data exceeds a predetermined threshold ratio.

On the other hand, referring back to FIG. 6A to explain the case of determining whether the threshold count is exceeded for each cell, each of the L_1 th layer to the L_n th layer is omnidirection in space of the lidar sensor. Each of the plurality of cells generated by subdividing the observable viewing angle range into a predetermined angle among the corresponding viewing angle range may be included, but the omnidirectional direction is not necessarily limited to 360 degrees.

Specifically, the L_kth layer, which is any one of the L_1th layer to the L_nth layer covering the reference number 610 , includes the L_k_1th cell to the L_k_mk cell (according to FIG. 6A , the portion marked with Layer 2 is the first Corresponding to the L_1 layer), referring to FIG. 6B , cells in which non-ground data is sensed in the ground estimation area 610 as shown in FIG. 6A are indicated by reference numerals 620 .

In addition, referring to FIG. 6C , the pollution detection apparatus calculates an accumulated count for each cell of L_k th empty surface data for each L_k_1 cell to the L_k_mk cell of the L_k th monitoring data acquired through the L_k th layer ( (refer to the number written for each cell), it is determined whether the accumulated count of at least one L_k_p cell among the L_k_1 th cell to the L_k_mk cell exceeds a threshold count for each cell, and the th cell determined to exceed the threshold count for each cell It may be determined that the L_k_p cell is contaminated (eg, the cell corresponding to reference number 613).

As a specific example, monitoring data corresponding to the L_k_p cell is acquired once per second from the L_k_p cell, and when the non-ground data corresponding to the L_k_p cell is accumulated 6 times, when the threshold count for each cell is 5, L_k_p The cell is judged as a contaminated cell.

On the other hand, in determining whether at least a portion of the accumulated counts of the L_1th monitoring data to the L_nth busy data included in each of the L_1th monitoring data to the L_nth monitoring data exceeds the L_1th threshold count In this case, the pollution detection apparatus may initialize the accumulated count corresponding to at least a portion of the monitoring data determined to include the ground data among the L_1 th monitoring data to the L_n th monitoring data.

For example, to initialize the accumulated count for each layer, when the L_5th monitoring data is acquired once per second from the L_5th layer, when the L_5th non-ground data is accumulated 6 times and the L_5th ground data is acquired, foreign substances The accumulated count is initialized because it may have disappeared or may correspond to another object on the ground (eg, a box on the ground) that is not a foreign substance in the first place. At this time, as a condition for initializing the accumulated count, the accumulated count may be initialized only when it is determined that the ground data is detected once, but is not limited thereto, and among the number of acquisitions of the monitoring data obtained during the preset time period. It may be initialized by determining whether the ratio of the number of times the ground data is acquired exceeds a predetermined threshold ratio.

As another example, in determining whether the accumulated count of at least one of the L_k_1 th cell to the L_k_mk cell for initializing the accumulated count for each cell exceeds the threshold count for each cell, The pollution detection apparatus may initialize the accumulated count of at least one L_k_p-th cell determined to include the ground data among the L_k_1 th cell to the L_k_mk cell.

As a specific example, monitoring data corresponding to the L_k_p cell is acquired once per second from the L_k_p cell. When the threshold count for each cell is 5, the L_5th busy data is accumulated 6 times and corresponds to the L_k_p cell. When ground data is acquired, since the foreign material may have disappeared or may correspond to an object on the ground that is not a foreign material in the first place, the accumulated count for each cell is initialized.

7 is a diagram illustrating a flowchart of a process for detecting contamination of a lidar sensor according to another embodiment of the present invention.

Referring to FIG. 7 , the pollution detection device acquires U_1 th monitoring data to U_s th monitoring data through each of the U_1 th layer to U_s layer determined not to interact with the ground among the plurality of layers of the lidar sensor ( S710), the pollution detection device acquires at least a portion of the U_1th monitoring data to the U_sth monitoring data multiple times over time by performing the step S710 multiple times according to time to obtain the U_1th monitoring data to the U_1th monitoring data It is determined whether the U_1th threshold count is exceeded among the accumulated counts, which is the number of missed counts for each specific point, which is at least some of the missing detection points among all the points corresponding to the target object data included in at least some of the U_s monitoring data, At least some specific points determined to have exceeded the U_1th threshold count may be regarded as at least one problem point, and it may be determined that at least one U_th specific layer corresponding to the problem point is contaminated (S720). ).

Here, as an example of a layer that does not interact with the ground, a layer that interacts with an object such as a building without interaction with the ground or a layer that faces the sky may be mentioned.

In addition, the U_1th monitoring data to the U_sth monitoring data (corresponding to reference number 220 in FIG. 2 ) are (i) less than a predetermined threshold distance by at least one of the U_1th layer to the U_sth layer. U_1 th type object data including information on the sensed object, (ii) the U_1 th layer to the U_s th layer U_2 type object data (at least a part of the U_2 type object data is the target object data), and (iii) the U_2 type object data that is not the U_1 type object data or the U_2 type object data but corresponds to the U_2 type object data It may include at least a part of non-object data affecting the detection of the point.

As a specific example, the U_1 type object data may include bird droppings, leaflets, or dust on the surface of the lidar sensor, and the U_2 type object data may include vehicles traveling together, surrounding buildings, etc. . Non-object data that is not the U_1 type object data or the U_2 type object data and that affects the detection of a point corresponding to the U_2 type object data means It means non-object data that has an effect such as preventing a point corresponding to U_2 type object data from being detected when it should be detected.

On the other hand, before step S710, the contamination detection device, with respect to the U_k-th layer that is any one of the U_1 th layer to the U_s th layer, an angle corresponding to the omnidirection in space of the lidar sensor A plurality of cells is generated by subdividing the observable viewing angle range into a predetermined angle to constitute a first axis, and subdividing the separation distance from the lidar sensor into a predetermined unit distance to constitute a second axis, and the plurality of cells Among them, at least one specific cell corresponding to the target object data may be determined (S705: not shown). This will be described later with reference to FIGS. 11A and 11B . In this case, in step S710, the pollution detection device may determine at least some specific monitoring data corresponding to the specific cell from among the U_1th monitoring data to the U_sth monitoring data.

Also, before step S705, the pollution detection device may additionally acquire other sensor data from another sensor (the other sensor includes at least a portion of an image sensor and a radar sensor) (S700: not shown) city). Thereafter, the step S705 includes a process of determining the specific cell corresponding to the target object data with respect to the U_k-th layer that is any one of the U_1 th layer to the U_s th layer with reference to the other sensor data. can be done

Specifically, referring to FIG. 8 , when other sensor-based data is obtained (S810), an object estimation candidate group region corresponding to target object data is set on a two-dimensional array composed of a viewing angle and a separation distance (S820), and an object The number of times that points corresponding to the target object data are not detected may be accumulated for each of a plurality of cells corresponding to the estimated candidate group region (S830). Thereafter, as a result of repeating this process, if it is determined that the accumulated number of non-detection determinations of a specific point in a specific cell is greater than the threshold count (S840), the contamination detection apparatus determines that the specific cell is contaminated (S841), If it is determined that the accumulated number of non-detection determinations of a specific point in a specific cell is not greater than the threshold count (S840), the contamination detection apparatus may determine that the specific cell is not contaminated (S842).

Referring to FIG. 9 , an area 900 estimated to be an object may be identified using another sensor such as a camera or a radar sensor to set an object estimation candidate group area. In the case of a camera or a radar sensor, since information about the distance cannot be accurately acquired, a process of enlarging the area by a predetermined ratio to capture a rather large area may be included. This area 900 may be reference number 1100 of FIG. 11A, which will be described later, which will be described later.

Referring to FIG. 10 , among all points 1010 corresponding to target object data, there may be at least some of the points 1020 for which detection is omitted, or all of the points 1030 may be omitted.

This may be due to foreign matter when the relevant point is missing even though the target object data is expected to be obtained in the object estimation candidate group area. It is determined that a specific layer is contaminated.

As an example of determining whether the number of omissions for each layer exceeds the threshold count, when the threshold count is 5, the U_5th monitoring data is acquired once per second from the U_5th layer. Points corresponding to the target object data If the number of missed detections of a specific point is counted as 6, it is determined that the U_5th layer corresponding to the problem point is contaminated. Here, the criterion for determining whether the accumulated count exceeds the threshold count may mean, but is not limited to, determining whether the continuously accumulated count exceeds the threshold count, and in some cases, during a preset time period It may mean determining whether the ratio of the number of missed detections among the number of acquisitions of the acquired monitoring data exceeds a predetermined threshold ratio.

Meanwhile, referring to FIGS. 11A and 11B in order to explain the case of determining whether the number of omissions for each cell exceeds the threshold count, the interval of a predetermined angle is 0.5° and the interval of the predetermined separation distance is 0.5m. A plurality of cells may be assumed. Here, the concept of a cell is different from the concept of a cell described in an embodiment of the present invention, but is not limited thereto. However, in FIGS. 11A and 11B , it is assumed that cells included in a two-dimensional array including a viewing angle and a separation distance are assumed.

Specifically, referring to FIG. 11B , a target point corresponding to the target cell is detected among consecutive cells arranged on the first axis in a state where the separation distance on the second axis is constant with the specific cell as a target If this does not occur and a neighboring point corresponding to an adjacent neighboring cell of the target cell is detected (1120), the contamination detection apparatus includes a process of counting the number of missed times of the target point corresponding to the target cell; and when the specific cell is not detected for all target points corresponding to all consecutive cells arranged on the first axis while the separation distance on the second axis is constant (1140), the The contamination detection apparatus may perform at least a part of a process of counting the number of missed counts of all target points corresponding to all target cells.

In addition, specifically, referring to FIG. 11B , the continuous cells arranged on the first axis in a state where the separation distance is constant means the same relationship as the reference number 1120 arranged in the upper and lower parts of 1.5 m. In this case, the undetected portions of the target point corresponding to the target cell are yellow cells of reference number 1120 , and those detected with respect to the neighboring points corresponding to the adjacent neighboring cells of the target cell are green of reference number 1120 . means cell. Therefore, in this case, the yellow cells of reference number 1120 are counted. This is, if there are cells in which the target point is detected with the yellow cells of the reference number 1120 in between, such as the green cells of the reference number 1120, in the case of the yellow cells of the reference number 1120, the point is This counting is done because the probability of omission is high.

However, unless a neighbor point corresponding to a neighboring cell of the target cell is detected, it is not counted as indicated by reference number 1130 and the determination is withheld. This is because, when the neighboring point corresponding to the neighboring cell is not detected, the probability that it is not contaminated is higher because the object estimation candidate group region has been enlarged due to a predetermined correction.

However, if the specific cell is not detected, all target points corresponding to all consecutive cells arranged on the first axis in a state where the separation distance on the second axis is constant, reference number It means the yellow cells of (1140), so all three cells are counted.

Meanwhile, when initializing the accumulated count for each layer, for example, when the pollution detection apparatus determines whether at least a portion of the accumulated count exceeds the U_1th threshold count, the U_1th monitoring data to the U_sth monitoring data It is determined that detection has been performed again by at least a portion of the U_1 th monitoring data to the U_s th monitoring data for each specific point that is at least a partial point for which detection is omitted among all the points corresponding to the target object data among the data For at least some specific points, the accumulated count may be initialized.

This is a specific example, and when the threshold count is 5, the U_5th layer acquires the U_5th monitoring data once per second, and the number of missing detections of a specific point whose detection is missing among all points corresponding to the target object data When a specific point corresponding to the target object data is detected again after being counted as 6, the accumulated count is initialized because it may correspond to a situation such as the disappearance of a foreign substance. At this time, as a condition for initializing the accumulated count, the accumulated count may be initialized only by determining that a specific point is detected once, but is not limited thereto, and among the number of acquisitions of the monitoring data acquired during the preset time period. It may be initialized by determining whether the ratio of the number of times a specific point is acquired exceeds a predetermined threshold ratio.

As another example, an example of initializing the accumulated count for each cell is described with reference to FIG. 11B , for at least one specific cell 1100 corresponding to specific target object data, at least corresponding to each of the specific cells 1100 In determining whether the accumulated count for the number of missed counts of the specific point for each cell, which is a point for which detection is omitted among points for each cell, exceeds the threshold count for each cell, the contamination detection device includes: at least a portion of the specific points for each cell , when it is determined that detection has been performed again by at least a part of the U_k th monitoring data, the accumulated count for each cell corresponding to at least a part of the specific point for each cell may be initialized.

Specifically, referring to the green cell 1110 of the specific cells 1100 of FIG. 11B , it is determined that the detection has been re-detected by at least a part of the U_k th monitoring data for a part of the specific point for each cell, and a specific point for each cell It can be confirmed that the accumulated count for each cell corresponding to some of them is initialized and becomes 0.

Meanwhile, although one embodiment and another embodiment of the present invention have been separately described so far, one embodiment and another embodiment of the present invention may be performed at the same time, and in some cases, an embodiment of the present invention and the present invention Various modifications may be envisaged, such as giving a change to the temporal relationship of another embodiment of .

In the above, the present invention has been described with specific matters such as specific components and limited embodiments and drawings, but these are provided to help a more general understanding of the present invention, and the present invention is not limited to the above embodiments. , various modifications and variations can be devised from these descriptions by those of ordinary skill in the art to which the present invention pertains.

Accordingly, the spirit of the present invention should not be limited to the above-described embodiments, and not only the claims described below, but also all modifications equivalently or equivalently to the claims described below belong to the scope of the spirit of the present invention. will do it

Claims (28)

A method for detecting contamination of a lidar sensor, the method comprising:
(a) The pollution detection device obtains L_1 th monitoring data to L_n th monitoring data through each of L_1 th layer to L_n th layer determined to interact with the ground among a plurality of layers of the lidar sensor, and the L_1 th monitoring determining whether at least a portion of each of data to the L_n-th monitoring data is ground data or non-ground data; and
(b) the pollution detection apparatus performs step (a) a plurality of times according to time to accumulate the L_1th monitoring data to L_nth empty ground data included in each of the L_1th monitoring data to the L_nth monitoring data It is determined whether at least some of the counts exceed the L_1th threshold count, and the at least one L_th layer corresponding to the at least one specific non-ground data determined to exceed the L_1th threshold count is contaminated with step to determine
including,
The non-ground data includes (i) L_1 th type object data including information on an object detected at less than a predetermined threshold distance by at least one of the L_1 th layer to the L_n th layer, (ii) the L_2 th type object data including information on an object detected at a predetermined threshold distance or more by at least one of the L_1 th layer to the L_n th layer, and (iii) the L_1 th type object data or the L_2 th Non-object data, which is not type object data, and includes non-object data in which at least a portion of the average and variance of the pattern of reflection data included in the ground data is determined to be different by at least a portion of a preset threshold average and critical variance how to do it with According to claim 1,
In step (b),
By performing step (a) a plurality of times according to time, at least a portion of the accumulated counts of the L_1th monitoring data to the L_nth busy data included in each of the L_1th monitoring data to the L_nth monitoring data is In determining whether the L_1 th threshold count is exceeded, the pollution detection device corresponds to at least a portion of the monitoring data determined to include the ground data among the L_1 th monitoring data to the L_n th monitoring data. A method characterized in that the accumulated count is initialized. According to claim 1,
In step (b),
Each of the L_1 th layer to the L_n th layer includes a plurality of cells generated by subdividing the observable viewing angle range into a predetermined angle among the angular ranges corresponding to the omnidirection in space of the lidar sensor, The L_k-th layer, which is any one of the L_1 th layer to the L_n th layer, includes L_k_1 th cells to L_k_mk cells,
The pollution detection apparatus calculates a cell-by-cell cumulative count of L_k-th empty surface data for each L_k_1 th cell to the L_k_mk cell of the L_k th monitoring data acquired through the L_k th layer, and the L_k_1 th cell to the L_k_mk th cell A method characterized in that it is determined whether the accumulated count of at least one L_k_p cell among the cells exceeds a threshold count for each cell, and the L_k_p cell determined to exceed the threshold count for each cell is determined to be contaminated. . 4. The method of claim 3,
In step (b),
In determining whether the accumulated count of the L_k_p cell of at least one of the L_k_1 cell to the L_k_mk cell exceeds the threshold count for each cell by performing the step (a) a plurality of times according to time, the contamination detection The method of claim 1, wherein the apparatus initializes the accumulated count of at least one L_k_p-th cell determined to include the ground data among the L_k_1th cell to the L_k_mk cell. delete According to claim 1,
Before step (a),
(a0) the contamination detection device, with reference to at least a part of the installation position, the installation angle, and the three-dimensional shape of the mounting body in which the lidar sensor is installed, the observable viewing angle range of the lidar sensor and setting a LiDAR ground estimation area corresponding to the L_1 th layer to the L_n th layer;
The step (a) is,
The pollution detection apparatus obtains the L_1 th monitoring data to the L_n th monitoring data through each of the L_1 th layer to the L_n th layer with reference to the LiDAR ground estimation area, and the L_1 th monitoring data to Method characterized in that it is determined whether at least a portion of each of the L_n-th monitoring data is the ground data or the non-ground data. According to claim 1,
In step (a),
The pollution detection device additionally acquires U_1 th monitoring data to U_s th monitoring data through each of the U_1 th layer to U_s th layer determined not to interact with the ground among the plurality of layers of the lidar sensor, At least a portion of each of the U_1 th monitoring data to the U_s th monitoring data includes information on an object detected at a predetermined threshold distance or more by at least one of the U_1 th layer to the U_s th layer U-th object It is additionally determined whether the data is
In step (b),
The contamination detection device acquires at least a portion of the U_1th monitoring data to the U_sth monitoring data multiple times according to time, and the U-th type included in each of at least a portion of the U_1th monitoring data to the U_sth monitoring data It is additionally determined that at least one U_th layer corresponding to at least one specific point that is determined to exceed the second threshold count among the accumulated counts corresponding to the number of times that at least some points of the object data is omitted is contaminated A method characterized in that A method for detecting contamination of a lidar sensor, the method comprising:
(a) obtaining, by the pollution detection apparatus, U_1 th monitoring data to U_s th monitoring data through each of the U_1 th layer to U_s th layer determined not to interact with the ground among a plurality of layers of the lidar sensor; and
(b) the contamination detection device obtains at least a portion of the U_1th monitoring data to the U_sth monitoring data multiple times over time by performing step (a) multiple times over time to obtain the U_1th monitoring data to determine whether the U_1th threshold count is exceeded among the accumulated counts, which is the number of missed counts for each specific point that is at least some of the missing detection points among all the points corresponding to the target object data included in at least a portion of the U_s monitoring data and determining that at least some specific points that are determined to have exceeded the U_1th threshold count as at least one problem point are contaminated with at least one U_th specific layer corresponding to the problem point
including,
In step (b),
When the pollution detection apparatus determines whether at least a part of the accumulated count exceeds the U_1th threshold count, among all the points corresponding to the target object data from among the U_1th monitoring data to the U_sth monitoring data For each of the specific points that are at least some points that are missing detection, for at least some specific points that are determined to have been detected again by at least some of the U_1th monitoring data to the U_sth monitoring data, the accumulated count is initialized characterized by,
Before step (a),
(a1) The contamination detection device can observe the U_k-th layer, which is any one of the U_1th layer to the U_sth layer, in an angular range corresponding to omnidirection in space of the lidar sensor A plurality of cells are generated by subdividing the viewing angle range into a predetermined angle to constitute a first axis, and subdividing the separation distance from the lidar sensor into a predetermined unit distance to constitute a second axis, and among the plurality of cells, the target Further comprising the step of determining at least one specific cell corresponding to the object data,
The step (a) is,
characterized in that the pollution detection device determines at least some specific monitoring data corresponding to the specific cell from among the U_1th monitoring data to the U_sth monitoring data,
The step (b) is,
The contamination detection device determines whether the accumulated count for the number of missed counts of the specific point for each cell, which is the missing point among the at least one point for each cell corresponding to each of the specific cell, for each specific cell exceeds a threshold count for each cell and determining that some specific cells determined to exceed the threshold count for each cell are contaminated,
Before step (a1),
(a0) further comprising, by the contamination detection device, additionally acquiring other sensor data from another sensor, wherein the other sensor includes at least some of an image sensor and a radar sensor;
The step (a1) is,
A process for determining the specific cell corresponding to the target object data is performed with respect to the U_k-th layer, which is any one of the U_1 th layer to the U_s th layer, with reference to the other sensor data. . delete delete delete 9. The method of claim 8,
In step (b),
In determining whether the accumulated count for the number of missed counts of the specific point for each cell, which is a point for which detection is missing, among at least one point for each cell corresponding to each of the specific cells for each specific cell exceeds the threshold count for each cell, the When the pollution detection apparatus determines that at least some of the specific points for each cell have been detected again by at least a part of the U_k th monitoring data, the accumulated count for each cell corresponding to at least some of the specific points for each cell is initialized A method characterized in that. 9. The method of claim 8,
In step (b),
Targeting the specific cell, in a state where the separation distance on the second axis is constant, a target point corresponding to the target cell among consecutive cells arranged on the first axis is not detected. When a neighboring point corresponding to an adjacent neighboring cell is detected, the contamination detection apparatus includes a process of counting the number of missed times of the target point corresponding to the target cell, and targeting the specific cell, When the detection is not performed for all target points corresponding to all consecutive cells arranged on the first axis in a state where the separation distance on the two axes is constant, the contamination detection device is configured to: and performing at least a part of the process of counting the number of missed counts of all target points. A method for detecting contamination of a lidar sensor, the method comprising:
(a) obtaining, by the pollution detection apparatus, U_1 th monitoring data to U_s th monitoring data through each of the U_1 th layer to U_s th layer determined not to interact with the ground among a plurality of layers of the lidar sensor; and
(b) the contamination detection device obtains at least a portion of the U_1th monitoring data to the U_sth monitoring data multiple times over time by performing step (a) multiple times over time to obtain the U_1th monitoring data to determine whether the U_1th threshold count is exceeded among the accumulated counts, which is the number of missed counts for each specific point that is at least some of the missing detection points among all the points corresponding to the target object data included in at least a portion of the U_s monitoring data and determining that at least some specific points that are determined to have exceeded the U_1th threshold count as at least one problem point are contaminated with at least one U_th specific layer corresponding to the problem point
including,
In step (b),
When the pollution detection apparatus determines whether at least a part of the accumulated count exceeds the U_1th threshold count, among all the points corresponding to the target object data from among the U_1th monitoring data to the U_sth monitoring data For each of the specific points that are at least some points that are missing detection, for at least some specific points that are determined to have been detected again by at least some of the U_1th monitoring data to the U_sth monitoring data, the accumulated count is initialized characterized by,
The U_1 th monitoring data to the U_s th monitoring data include: (i) a U_1 th type including information on an object detected at less than a predetermined threshold distance by at least one of the U_1 th layer to the U_s th layer Object data, (ii) U_2 th type object data including information on an object detected at a predetermined threshold distance or more by at least one of the U_1 th layer to the U_s th layer - Among the U_2 th type object data At least a portion is the target object data; A method characterized in that An apparatus for detecting contamination of a lidar sensor, comprising:
at least one memory storing instructions; and
at least one processor configured to execute the instructions;
The processor (I) obtains L_1 th monitoring data to L_n th monitoring data through each of L_1 th layer to L_n th layer determined to interact with the ground among a plurality of layers of the lidar sensor, and the L_1 th monitoring data to a process of determining whether at least a portion of each of the L_n-th monitoring data is ground data or non-ground data; and (II) at least a portion of the accumulated counts of the L_1th monitoring data to the L_nth busy data included in each of the L_1th monitoring data to the L_nth monitoring data by performing the process (I) a plurality of times over time A process of determining whether the L_1th threshold count is exceeded, and determining that at least one L_th specific layer corresponding to the at least one specific non-ground data determined to exceed the L_1th threshold count is contaminated
do, but
The non-ground data includes: (i) L_1 th type object data including information on an object detected at less than a predetermined threshold distance by at least one of the L_1 th layer to the L_n th layer, (ii) the L_2 th type object data including information on an object detected at a predetermined threshold distance or more by at least one of the L_1 th layer to the L_n th layer, and (iii) the L_1 th type object data or the L_2 th Non-object data that is not type object data, and includes non-object data in which at least a portion of the average and variance of the pattern of reflection data included in the ground data is determined to be different by at least a portion of a preset critical average and critical variance device to do. 16. The method of claim 15,
In the process (II) above,
the processor performs the (I) process a plurality of times according to time, and among the accumulated counts of the L_1th monitoring data to the L_nth busy data included in each of the L_1th monitoring data to the L_nth monitoring data. In determining whether at least a portion exceeds the L_1 th threshold count, the cumulative count corresponding to at least a portion of the monitoring data determined to include the ground data among the L_1 th monitoring data to the L_n th monitoring data Device characterized in that it initializes. 16. The method of claim 15,
In the process (II) above,
Each of the L_1 th layer to the L_n th layer includes a plurality of cells generated by subdividing the observable viewing angle range into a predetermined angle among the angular ranges corresponding to the omnidirection in space of the lidar sensor, The L_k-th layer, which is any one of the L_1 th layer to the L_n th layer, includes L_k_1 th cells to L_k_mk cells,
The processor calculates a cell-by-cell cumulative count of L_k-th busy data for each L_k_1 cell to the L_k_mk cell of the L_k-th monitoring data acquired through the L_k-th layer, and from among the L_k_1 cell to the L_k_mk cell An apparatus characterized in that it is determined whether the accumulated count of at least one L_k_p-th cell exceeds a threshold count for each cell, and the L_k_p-th cell determined to exceed the threshold count for each cell is determined to be contaminated. 18. The method of claim 17,
In the process (II) above,
In determining whether the accumulated count of the L_k_p cell of at least one of the L_k_1 cell to the L_k_mk cell by performing the (I) process a plurality of times according to time exceeds the threshold count for each cell, the processor , initializing the accumulated count of at least one L_k_p-th cell determined to include the ground data among the L_k_1 th cell to the L_k_mk cell. delete 16. The method of claim 15,
Prior to the (I) process,
(I-0) the processor, with reference to at least a part of an installation position, an installation angle, and a three-dimensional shape of a mounting body in which the lidar sensor is installed, an observable viewing angle range of the lidar sensor and a process of setting a LiDAR ground estimation area corresponding to the L_1 th layer to the L_n th layer;
The (I) process is
The processor obtains the L_1 th monitoring data to the L_n th monitoring data through each of the L_1 th layer to the L_n th layer with reference to the LiDAR ground estimation area, and the L_1 th monitoring data to the L_n th monitoring data and performing a process of determining whether at least a portion of each of the L_n monitoring data is the ground data or the non-ground data. 16. The method of claim 15,
In the process (I) above,
The processor further acquires U_1 th monitoring data to U_s th monitoring data through each of U_1 th layer to U_s th layer determined not to interact with the ground among the plurality of layers of the lidar sensor, and monitoring the U_1 th Data to at least a portion of each of the U_s-th monitoring data is U-th object data including information about an object detected at a predetermined threshold distance or more by at least one of the U_1-th layer to the U_s-th layer judge more,
In the process (II) above,
The processor acquires at least some of the U_1th monitoring data to the U_sth monitoring data multiple times over time, and the U-th type object data included in each of at least some of the U_1th monitoring data to the U_sth monitoring data Further determining that the at least one U_th specific layer corresponding to the at least one specific point determined to exceed the second threshold count among the accumulated counts corresponding to the number of missing points is contaminated device characterized. An apparatus for detecting contamination of a lidar sensor, comprising:
at least one memory storing instructions; and
at least one processor configured to execute the instructions;
(I) a process of obtaining, by the processor, U_1 th monitoring data to U_s th monitoring data through each of U_1 th layer to U_s th layer determined not to interact with the ground among a plurality of layers of the lidar sensor; and (II) obtaining at least a portion of the U_1th monitoring data to the U_sth monitoring data multiple times over time by performing the (I) process multiple times over time to monitor the U_1th monitoring data to the U_sth monitoring data It is determined whether or not the U_1th threshold count is exceeded among accumulated counts, which is the number of missed counts for each specific point, which is at least some of the missing detection points among all points corresponding to the target object data included in at least a portion of the data, and the U_1th A process of determining that at least one specific point determined to have exceeded the threshold count as at least one problem point and at least one U_th specific layer corresponding to the problem point is contaminated
do, but
In the process (II) above,
When the processor determines whether at least a portion of the accumulated count exceeds the U_1th threshold count, the detection of all points corresponding to the target object data among the U_1th monitoring data to the U_sth monitoring data is performed For each of the specific points that are at least some of the missing points, the accumulated count is initialized for at least some specific points that are determined to have been detected again by at least some of the U_1th monitoring data to the U_sth monitoring data with
Prior to the (I) process,
(I-1) The processor can observe the U_k-th layer, which is any one of the U_1 th layer to the U_s th layer, in an angular range corresponding to omnidirection in space of the lidar sensor A plurality of cells are generated by subdividing the viewing angle range into a predetermined angle to constitute a first axis, and subdividing the separation distance from the lidar sensor into a predetermined unit distance to constitute a second axis, and among the plurality of cells, the target Further performing a process of determining at least one specific cell corresponding to the object data,
The (I) process is
characterized in that the processor determines at least some specific monitoring data corresponding to the specific cell from among the U_1th monitoring data to the U_sth monitoring data,
The (II) process is
The processor determines whether the accumulated count of the number of missed counts of the specific point for each cell, which is a point for which detection is missed, among at least one point for each cell corresponding to each of the specific cells for each specific cell exceeds a threshold count for each cell, It is characterized in that it is determined that some specific cells that are determined to exceed the threshold count for each cell are contaminated,
Before the process (I-1) above,
(I-0) the processor further performs a process of further acquiring other sensor data from another sensor, wherein the other sensor includes at least some of an image sensor and a radar sensor;
The (I-1) process is
A process of determining the specific cell corresponding to the target object data is performed with respect to the U_k-th layer, which is any one of the U_1 th layer to the U_s th layer, with reference to the other sensor data . delete delete delete 23. The method of claim 22,
In the process (II) above,
In determining whether the accumulated count for the number of missed counts of the specific point for each cell, which is a point for which detection is missing, among at least one point for each cell corresponding to each of the specific cells for each specific cell exceeds the threshold count for each cell, the The processor initializes the accumulated count for each cell corresponding to at least some of the specific points for each cell when it is determined that at least some of the specific points for each cell have been detected again by at least a part of the U_k th monitoring data device to do. 23. The method of claim 22,
In the process (II) above,
Targeting the specific cell, in a state where the separation distance on the second axis is constant, a target point corresponding to the target cell among consecutive cells arranged on the first axis is not detected. When a neighboring point corresponding to an adjacent neighboring cell is detected, the processor counts the number of missed times of the target point corresponding to the target cell, and targeting the specific cell, the second axis In a state in which the separation distance of an image is constant, when all target points corresponding to all consecutive cells arranged on the first axis are not detected, the number of missed counts of all target points corresponding to all target cells is calculated and performing at least a portion of the counting process. An apparatus for detecting contamination of a lidar sensor, comprising:
at least one memory storing instructions; and
at least one processor configured to execute the instructions;
(I) a process of obtaining, by the processor, U_1 th monitoring data to U_s th monitoring data through each of U_1 th layer to U_s th layer determined not to interact with the ground among a plurality of layers of the lidar sensor; and (II) obtaining at least a portion of the U_1th monitoring data to the U_sth monitoring data multiple times over time by performing the (I) process multiple times over time to monitor the U_1th monitoring data to the U_sth monitoring data It is determined whether or not the U_1th threshold count is exceeded among accumulated counts, which is the number of missed counts for each specific point, which is at least some of the missing detection points among all points corresponding to the target object data included in at least a portion of the data, and the U_1th A process of determining that at least one specific point determined to have exceeded the threshold count as at least one problem point and at least one U_th specific layer corresponding to the problem point is contaminated
do, but
In the process (II) above,
When the processor determines whether at least a portion of the accumulated count exceeds the U_1th threshold count, the detection of all points corresponding to the target object data among the U_1th monitoring data to the U_sth monitoring data is performed For each of the specific points that are at least some of the missing points, the accumulated count is initialized for at least some specific points that are determined to have been detected again by at least some of the U_1th monitoring data to the U_sth monitoring data with
The U_1 th monitoring data to the U_s th monitoring data include: (i) a U_1 th type including information on an object detected at less than a predetermined threshold distance by at least one of the U_1 th layer to the U_s th layer Object data, (ii) U_2 th type object data including information on an object detected at a predetermined threshold distance or more by at least one of the U_1 th layer to the U_s th layer - Among the U_2 th type object data At least a portion is the target object data; Device characterized in that.

Images