JP2019049514A - Object detection device and object detection method - Google Patents

Abstract

To more suppress a snowflake, a rain drop, a leaf, or the like from being wrongly detected as a detection target object in an object detection device and object detection method, and to improve detection accuracy of the detection target object therein.SOLUTION: An object detection device comprises: a laser sensor unit that emits laser light for measurement and can change an emission direction, and receives reflection light of the laser light to output a signal as an electric signal; and a signal processing unit that conducts detection of objects on the basis of an output signal of the laser sensor unit. The signal processing unit is configured to extract an inside-area measurement point located in a preset detection target area, and determine whether the inside-area measurement point is noise or not on the basis of an outside-area measurement point adjacent to the inside-area measurement point and located outside the detection target area.SELECTED DRAWING: Figure 1

Description

  The present invention relates to an object detection apparatus and an object detection method.

  BACKGROUND ART Conventionally, an object detection device using a laser sensor unit has been used in various places such as a roadway and a level crossing. For example, on a roadway, the presence or absence of a vehicle is detected by an object detection device, and a traffic congestion state is grasped from time-series changes in measurement results. Moreover, at the level crossing, the object detection device detects the presence or absence of the obstacle in the level crossing, and the safety of the train traveling is confirmed. In such an object detection apparatus, the reflected light of the measurement laser beam emitted from the laser sensor unit is received to detect the reflection position (measurement point) of the measurement laser beam, and this measurement point is a detection target set in advance. An object is detected based on whether or not it is located in an area (see, for example, Patent Document 1).

Japanese Patent Laid-Open No. 2002-140790

  By the way, in the object detection apparatus, it is required to prevent erroneous detection of snow flakes, raindrops, leaves of trees and the like as vehicles or obstacles. For example, in the area near the laser sensor unit with high resolution, the measurement point obtained by reflecting the laser light on a snowflake or the like can be used as a vehicle or a noise by removing one having a small number of measurement points constituting the object as noise. It is conceivable to prevent false detection as an obstacle. However, even in such a case, it is difficult to prevent erroneous detection due to all the snow flakes and the like, and it is desirable to further reduce erroneous detection and to improve the detection accuracy of the object detection device.

  The present invention has been made in view of the above-described problems, and in the object detection apparatus and the object detection method, it is possible to further suppress false detection of snow flakes, raindrops, leaves of trees and the like as objects to be detected. The object is to improve the detection accuracy of an object.

  The present invention adopts the following configuration as means for solving the above-mentioned problems.

  According to a first aspect of the present invention, there is provided a laser sensor unit capable of emitting a laser beam for measurement and changing the emitting direction and receiving the reflected light of the laser beam and outputting it as an electric signal, and an output signal of the laser sensor unit And a signal processing unit configured to detect an object based on the signal processing unit, wherein the signal processing unit extracts an in-area measuring point located in a predetermined detection target area, and sets the in-area measuring point as the in-area measuring point. A configuration is employed in which it is determined whether the in-area measurement point is noise or not based on the out-of-area measurement point adjacent to and located outside the detection target area.

  In a second invention according to the first invention, the signal processing unit is configured to measure the horizontal width of the object indicated by the in-area measuring point based on the out-of-area measuring point adjacent to the side of the in-area measuring point. When the dimension is determined, and the number of in-area measurement points included in the object is equal to or less than a preset threshold and the width dimension is equal to or less than a preset threshold, the in-area measurement points are noise It adopts a configuration of determining that

  In a third invention according to the first or second invention, the signal processing unit is configured such that the number of in-area measurement points included in the object indicated by the in-area measurement point is equal to or less than a preset threshold value and When there is an out-of-area measurement point adjacent below the in-area measurement point, a configuration is adopted in which the in-area measurement point is determined to be noise.

  In a fourth invention according to any one of the first to third inventions, the signal processing unit uses the in-area measurement point as a reference as the distance from the laser sensor unit to the in-area measurement point increases. A configuration is adopted in which the estimated shape of the object is largely determined.

  In a fifth invention according to any one of the first to fourth inventions, the signal processing unit is adjacent when the distance from the laser sensor unit to the in-area measurement point is shorter than a preset threshold. A configuration is employed in which it is determined whether the in-area measurement point is noise based on the out-of-area measurement point.

  A sixth invention is an object detection method using a laser sensor unit that is capable of emitting a laser beam for measurement and changing the emitting direction, and receiving the reflected light of the laser beam and outputting it as an electric signal, The in-area measurement point is extracted based on the out-of-area measurement point adjacent to the in-area measurement point while extracting the in-area measurement point located in the predetermined detection target area. A configuration is adopted in which it is determined whether or not it is noise.

  According to the present invention, not only the measurement points (measurement points in the area) in the detection target area used in the conventional object detection apparatus and object detection method but also measurement points outside the detection target area (measurement points outside the area) ) Is also used to determine whether the measurement point in the detection target area is noise. When the measurement point adjacent to the in-area measurement point is an out-of-area measurement point, the in-area measurement point is an extremely small object, which may be an object not to be detected, such as snow flakes, raindrops and leaves of trees. Therefore, as in the present invention, it is determined whether or not the in-area measurement point is noise based on the out-of-area measurement point adjacent to the in-area measurement point and located outside the detection target area. Thus, it is possible to more reliably prevent an object not to be detected, such as snow flakes, raindrops and leaves of trees, from being erroneously detected as an object to be detected. Therefore, according to the present invention, in the object detection device and the object detection method, false detection of snow flakes, raindrops, leaves of trees and the like as false detection objects is further suppressed, and detection accuracy of the detection target objects is further improved. Is possible.

It is a block diagram showing a schematic structure of an object detection device in an embodiment of the present invention. It is a flowchart for demonstrating the object detection method by the object detection apparatus in one Embodiment of this invention. It is a flowchart which shows the process which a search part performs with respect to one in-area measurement point in the scan search process by the object detection apparatus in one Embodiment of this invention. It is a flowchart which shows the process which a search part performs with respect to the measurement point in one area in the swing search process by the object detection apparatus in one Embodiment of this invention. It is a flowchart which shows the process which a search part performs with respect to one object in the deletion process by the object detection apparatus in one Embodiment of this invention. It is a schematic diagram for demonstrating the effect | action of the object detection apparatus in one Embodiment of this invention, and an object detection method. It is a schematic diagram for demonstrating the effect | action of the object detection apparatus in one Embodiment of this invention, and an object detection method.

  Hereinafter, an embodiment of an object detection apparatus and an object detection method according to the present invention will be described with reference to the drawings. In the following drawings, the scale of each member is appropriately changed in order to make each member a recognizable size.

  FIG. 1 is a block diagram showing a schematic configuration of an object detection device 1 of the present embodiment. As shown in this figure, the object detection device 1 of the present embodiment includes a laser sensor unit 2, a signal processing unit 3, and a unit control unit 4.

  The laser sensor unit 2 can emit laser light for measurement and change the emitting direction, and receives the reflected light of the laser light and outputs it as an electric signal. The structure of the laser sensor unit 2 is not particularly limited. For example, a light source for emitting laser light, a scanning mechanism such as a polygon mirror for scanning laser light emitted from the light source in the horizontal direction, and each polygon mirror A swing mechanism that shakes the laser light in the vertical direction by tilting it, a light receiving unit that receives the reflected light of the laser light, converts it into an electric signal, and outputs it. In order to obtain measurement data of one frame, such a laser sensor unit 2 is under the control of the unit control unit 4 and from one end to the other end of the laser beam in the left-right direction (hereinafter, scan direction) The scanning operation of continuously scanning is performed each time the swing angle (angle in the vertical direction) is changed stepwise.

  The signal processing unit 3 extracts the position (measurement point) where the laser beam is reflected based on the output signal of the laser sensor unit 2, and the detection target object is positioned in the detection target area set in advance based on the measurement point. Output the result of whether or not As shown in FIG. 1, such a signal processing unit 3 includes an extraction unit 3a, a search unit 3b, a grouping unit 3c, a conversion unit 3d, a noise deletion unit 3e, an object synthesis unit 3f, a determination unit 3g, and object information addition. It has functional units such as a unit 3h and an output unit 3i.

  In the object detection device 1 of the present embodiment, among the spaces that can be measured with laser light for measurement, a partial space set in advance is set as a detection target area, and a detection target object is present in the detection target area Whether or not is output as a detection result. The extraction unit 3a extracts measurement points (hereinafter referred to as in-area measurement points) located in the detection target area among a plurality of measurement points obtained by changing the emission direction of the laser light.

  The search unit 3b is a measurement point located outside the detection target area (hereinafter referred to as an out-of-area measurement point), and searches for a measurement point adjacent to the in-area measurement point in the scan direction or the swing direction. Further, when there is an out-of-area measurement point adjacent to the in-area measurement point as a result of the search, the search unit 3b includes the in-area measurement point based on the out-of-area measurement point. Limit the size of the scan direction of the assumed object area. In addition, the "object assumed area" said here means the area | region estimated that the object (namely, object which reflected the laser beam) which caused the measurement point in an area to exist exists. An object assumed area based on one in-area measurement point is an area having a size in the scan direction, the swing direction, and the depth direction orthogonal to these, and the size is determined in advance according to the resolution of the object detection device 1 or the like. It is set. The resolution of the object detection device 1 decreases with distance from the laser sensor unit 2. Therefore, for example, when the distance from the laser sensor unit 2 to the in-area measurement point is small, the size of the object assumed area in the scanning direction (that is, the estimated shape of the object) is reduced. As the distance to the measurement point is larger, the size of the object assumed area in the scan direction is increased. In the object detection device 1 of the present embodiment, when there is no out-of-area measurement point adjacent to the in-area measurement point in the scanning direction, the area according to the distance from the laser sensor unit 2 to the in-area measurement point The size of the assumed object area based on the inner measurement point is uniquely set. On the other hand, when there is an out-of-area measurement point adjacent to the in-area measurement point in the scan direction, the search unit 3b limits the size of the object assumed area to the position in the scan direction of the out-of-area measurement point. Further, when there is an out-of-area measurement point adjacent below the in-area measurement point as a result of the search, the search unit 3b determines that the object indicated by the in-area measurement point is floating.

  The grouping unit 3c groups in-area measurement points in which object assumed areas overlap each other based on the in-area measurement points. In addition, the grouping unit 3c performs the single in-area measurement even for a single in-area measurement point in which the assumed-object region based on another in-area measurement point does not overlap the assumed object region based on itself. Make points a group. The conversion unit 3d converts the in-area measurement point into an object. At this time, the conversion unit 3 d sets one group set by the grouping unit 3 c as one object, and sets a space occupied by the object assumed area as a space where the object exists.

  The noise removing unit 3e is configured such that the number of in-area measurement points included in the object converted by the conversion unit 3d is equal to or less than a preset threshold (hereinafter, configuration measurement score threshold), and the width in the scanning direction of the object When it is determined that the value is equal to or less than a preset threshold (width threshold) or when it is determined that the object is floating, the object is deleted as a noise component (that is, the object is not a detection target object). In the object detecting device 1 of the present embodiment, the noise removing unit 3e satisfies the above conditions for an object based on the in-area measuring point present at a position separated from the laser sensor unit 2 by a predetermined distance or more. Do not delete even. The distance from the laser sensor unit 2 not to delete the object can be arbitrarily set according to the resolution of the object detection device 1.

  The object combining unit 3 f combines the previously generated object and the later generated object. In the object detection device 1 of the present embodiment, an object is generated for each one scanning operation (operation in which laser light is continuously scanned from one end of one frame to the other end of one frame). When the objects obtained by different scanning operations overlap each other, the object combining unit 3 f combines these objects into one object.

  The determination unit 3g determines whether the composition of the object is completed for one frame. The object information adding unit 3 h adds object information such as an ID number and a size as related information to the object combined by the object combining unit 3 f. The output unit 3i outputs the detection result, that is, the detection target object located in the detection target area and the related information to the outside.

  The unit control unit 4 controls the laser sensor unit 2. For example, the unit control unit 4 exchanges data with the signal processing unit 3 as needed, and controls the laser sensor unit 2 based on the data. The signal processing unit 3 and the unit control unit 4 are embodied by, for example, a computer having a central processing unit (CPU) for performing arithmetic processing, a memory for storing data and software, and the like.

  Next, the operation (object detection method) of the object detection device 1 according to the present embodiment configured as described above will be described with reference to the flowcharts of FIGS. In the following description, it is assumed that the scanning operation is repeatedly performed while the laser sensor unit 2 performs the swing operation in stages under the control of the unit control unit 4.

  First, the unit controller 4 moves the laser beam emitted from the laser sensor unit 2 by one scan. During this time, the signal processing unit 3 performs a measurement process of acquiring a measurement point for one scan based on the signal input from the laser sensor unit 2 (step S1). In addition, since the signal represented by polar coordinates is output from the laser sensor unit 2, the signal processing unit 3 converts the polar coordinates into orthogonal coordinates.

  Subsequently, the signal processing unit 3 performs an extraction process of extracting the in-area measurement point from the measurement point acquired in step S1 by the extraction unit 3a (step S2). Here, the extraction unit 3a determines whether or not the coordinates of the measurement point acquired in step S1 are located in a preset detection target area. Then, the extraction unit 3a extracts measurement points whose coordinates are located in the detection target area as in-area measurement points, and for example, numbers them sequentially in the scan direction and stores them.

  Subsequently, the signal processing unit 3 causes the search unit 3b to perform a scan search step (step S3) and a swing search step (step S4). In the scan search step, when there are out-of-area measurement points adjacent in the scan direction for all in-area measurement points extracted in step S2, the search unit 3b assumes an object based on the in-area measurement points. Limit the width of the area scan direction.

  FIG. 3 is a flowchart showing a process performed by the search unit 3b on one in-area measurement point in the scan search step (step S3). As shown in this figure, the search unit 3b first sets the N value to "0" (step S3a), and then sets the N value to "N + 1" (step S3b). Subsequently, the search unit 3b determines whether the N value is smaller than the scan search upper limit (step S3c). Here, the “scan search upper limit” is set to a value at which the number of consecutive in-area measurement points adjacent in the scan direction should be reliably determined to be the detection target object. In the present embodiment, the scan search step (step S3) is performed to delete an object such as a snowflake having a small width in the scan direction as a noise component. Therefore, by setting the scan search upper limit, it is possible to prevent the scan search process from being unnecessarily continued with respect to the in-area measurement point obtained by the object of the size that should not be deleted as the noise component. Can.

  If the N value is smaller than the scan search upper limit in step S3c, the search unit 3b determines whether N adjacent measurement points in the plus scan direction are out-of-area measurement points (step S3d). Then, when the N adjacent measurement points are out-of-area measurement points, or when it is determined in step S3 c that the N value is not smaller than the scan search upper limit, the search unit 3b takes the in-area measurement points as a reference. The X distance (the size in the plus scan direction) in the plus scan direction of the assumed object area is calculated (step S3d). For example, when the N adjacent measurement points are the out-of-area measurement points, the search unit 3b sets the distance from the in-area measurement points in the scan direction to the out-of-area measurement points as the X distance. When the search unit 3b determines in step S3c that the N value is not smaller than the scan search upper limit, the distance in the scan direction from the in-area measurement point to the measurement point with the largest N value is the X distance. Do. When the out-of-area measurement point can not be found in the plus scan direction, the search unit 3b determines the X distance in the plus scan direction of the object assumed area according to the distance from the laser sensor unit 2 to the target in-area measurement point. Set to a fixed value.

  If the search unit 3b determines that the N adjacent measurement points in the plus scan direction are not the out-of-area measurement points in step S3d, the search unit 3b returns to step S3b and increases the number of N values by one.

  When the X distance in the positive scan direction of the object assumed area is calculated in step S3e, the search unit 3b sets the N value to "0" (step S3f), and subsequently sets the N value to "N + 1". (Step S3g). Subsequently, the search unit 3b determines whether the N value is smaller than the scan search upper limit (step S3h).

  If the N value is smaller than the scan search upper limit in step S3h, the search unit 3b determines whether N adjacent measurement points in the minus scan direction are out-of-area measurement points (step S3i). Then, when the N adjacent measurement points are out-of-area measurement points, or when it is determined in step S3 h that the N value is not smaller than the scan search upper limit, the search unit 3b takes the in-area measurement points as a reference. The X distance (the size in the minus scan direction) in the minus scan direction of the assumed object area is calculated (step S3j). For example, when the N adjacent measurement points are the out-of-area measurement points, the search unit 3b sets the distance from the in-area measurement points in the scan direction to the out-of-area measurement points as the X distance. When the search unit 3b determines in step S3h that the N value is not smaller than the scan search upper limit, the distance in the scan direction from the in-area measurement point to the measurement point with the largest N value is the X distance. Do. When the out-of-area measurement point can not be found in the negative scan direction, the search unit 3b determines the X distance in the negative scan direction of the object assumed area according to the distance from the laser sensor unit 2 to the target in-area measurement point. Set to a fixed value.

  If the search unit 3b determines that the N adjacent measurement points in the minus scan direction are not the out-of-area measurement points in step S3i, the search unit 3b returns to step S3g again and increases the number of N values by one.

  In the swing search step, when there is an out-of-area measurement point adjacent to the lower side with respect to all in-area measurement points extracted in step S2, the search unit 3b assumes an object assumed area based on the in-area measurement point The floating determination flag is set with respect to.

  FIG. 4 is a flowchart showing a process performed by the search unit 3b on one in-area measurement point in the swing search process (step S4). As shown in this figure, the search unit 3b first sets the N value to "0" (step S4a), and then sets the N value to "N + 1" (step S4b). Subsequently, the search unit 3b determines whether the N value is smaller than the swing search upper limit (step S4c). Here, the “swing search upper limit” is set to a value at which the number of consecutive in-area measurement points adjacent to the lower side should be reliably determined to be a detection target object. In the present embodiment, the swing search step (step S4) is performed to delete a floating small object such as a snowflake as a noise component. For this reason, by setting the swing search upper limit, the area obtained by a floating object of a size not to be deleted as a noise component (or a part of the object placed on the ground that is projected in the horizontal direction) Unnecessary swing search process can be prevented from continuing to the measurement point.

  If the N value is smaller than the swing search upper limit in step S4c, the search unit 3b determines whether or not the N adjacent measurement points are close to each other in the negative swing direction (downward). (Step S4d). In step S4d, it is determined whether N adjacent measurement points in the minus swing direction (downward) are distant out-of-area measurement points (step S4e). Then, when the N adjacent measurement points are the out-of-area measurement points, or when the search unit 3b determines that the N value is not smaller than the swing search upper limit in step S4c, the swing search process (step S4) The floating determination flag is set to the in-area measurement point which is the target of (step S4f).

  Further, the search unit 3b determines in step S4c that it is not an out-of-area measurement point where N adjacent measurement points are close in the negative swing direction, and in step S4d, the N adjacent measurement points outside the distant area If it is determined that the point is not a measurement point, the process returns to step S4b again to increase the number of N values by one.

  Returning to FIG. 2, when the swing search step (step S4) is completed, the signal processing unit 3 performs a grouping step (step S5) by the grouping unit 3c. The grouping unit 3c groups in-area measurement points in which object assumed areas overlap each other based on the in-area measurement points. In addition, the grouping unit 3c performs the single in-area measurement even for a single in-area measurement point in which the assumed object area based on another in-area measurement point does not overlap the assumed object area based on itself. Make points a group.

  Subsequently, the signal processing unit 3 performs the conversion process by the conversion unit 3d (step S6). The conversion unit 3d converts the in-area measurement point into an object. The conversion unit 3 d sets one group set by the grouping unit 3 c as one object, and sets the space occupied by the object assumed area as the space where the object exists.

  Subsequently, the signal processing unit 3 causes the noise removing unit 3e to perform the removing process (step S7). FIG. 5 is a flowchart showing a process performed by the noise removing unit 3e on one of the objects converted in step S6 in the removing step (step S7). As shown in this figure, the noise removing unit 3e determines whether the in-area measurement points forming the object are less than or equal to a preset configuration score threshold (step S7a).

  If it is determined in step S7a that the in-area measurement points that constitute the object are not less than or equal to the configuration score threshold, the noise deletion unit 3e deletes the target object without deleting it (step S7). Finish). On the other hand, when it is determined in step S7a that the in-area measurement points constituting the object are equal to or less than the configuration score threshold, the noise removing unit 3e sets in advance the X width (width dimension in the scanning direction) of the object. It is determined whether it is less than or equal to the width threshold (step S7b). If the noise removing unit 3e determines in step S7b that the X width of the object is not equal to or less than the width threshold, the noise removing unit 3e determines whether the floating determination flag is set for the in-area measurement point configuring the object. (Step S7c). Note that, for example, when there are a plurality of in-area measurement points that configure an object, the noise removal unit 3e sets the floating determination flag to all of the in-area measurement points that configure the object. In step S7c, it is determined that the ascent determination flag is set for the in-area measurement point constituting the object.

  When the noise removing unit 3e determines in step S7b that the X width of the object is equal to or less than the width threshold, or in step S7c, the floating determination flag is set for the in-area measurement point that configures the object. If it is determined that the object is a noise component (that is, an object that is not a detection target object), the object is deleted. In step S7b, the noise removing unit 3e determines that the X width of the object is not equal to or less than the width threshold, and further determines that the floating determination flag is not set for the in-area measurement point configuring the object in step S7c. In this case, the deletion process (step S7) for the object is ended without deleting the object of interest.

  In the present embodiment, the noise removing unit 3e sets the X width of the object at step S7b for the object based on the in-area measurement point present at a position separated from the laser sensor unit 2 by a predetermined distance or more. If it is determined that the value is equal to or less than the threshold value, or if it is determined in step S7c that the rising determination flag is set for the in-area measurement point that configures the object, deletion is not performed.

  Returning to FIG. 2, when the deletion process (step S5) is completed, the signal processing unit 3 performs an object combining process (step S8) by the object combining unit 3f. When the object generated in the main scan operation overlaps the object generated in another scan operation, the object synthesis unit 3 f synthesizes these objects into one object.

  Subsequently, the signal processing unit 3 determines whether the combination of all objects for one frame is completed by the determination unit 3g (step S9). If it is determined by the determination unit 3g that synthesis of all objects for one frame is not completed, the emission direction of the laser light for measurement is changed to the negative swing direction, and the signal processing unit 3 performs step S1 again. Return to

  If it is determined in step S9 that the synthesis of all objects for one frame has been completed, the signal processing unit 3 determines the ID number and the size of the object synthesized by the object information adding unit 3h. Object information as related information (step S10), and the output unit 3i outputs the detection result, that is, the detection target object located in the detection target area and its related information to the outside (step S11).

  According to the object detection device 1 and the object detection method of the present embodiment as described above, not only measurement points (measurement points within the area) in the detection target area used in the conventional object detection device and object detection method The measurement point outside the detection target area (measurement point outside the area) is also used to determine whether the measurement point in the detection target area is noise. When the measurement point adjacent to the in-area measurement point is an out-of-area measurement point, the in-area measurement point is an extremely small object, which may be an object not to be detected, such as snow flakes, raindrops and leaves of trees. Therefore, as in the present invention, it is determined whether or not the in-area measurement point is noise based on the out-of-area measurement point adjacent to the in-area measurement point and located outside the detection target area. Thus, it is possible to more reliably prevent an object not to be detected, such as snow flakes, raindrops and leaves of trees, from being erroneously detected as an object to be detected. Therefore, according to the object detection device 1 and the object detection method of the present embodiment, false detection of snow flakes, raindrops, leaves of trees and the like as false detection objects is further suppressed, and detection accuracy of the detection target object is further improved. It becomes possible.

  Further, in the object detection device 1 of the present embodiment, the signal processing unit 3 assumes an object based on the out-of-area measurement point adjacent to the side (scan direction) of the in-area measurement point in the scan search step (step S3). The horizontal width dimension of the area (that is, the object indicated by the in-area measuring point) is determined, and the number of in-area measuring points included in the object in the deleting step (step S7) is equal to or less than the configuration score threshold and the width dimension is the width threshold In the following cases, it is determined that the object (that is, the in-area measurement point that constitutes the object) is noise. According to the object detection device 1 of this embodiment, as shown in the schematic view of FIG. 6A, when an object such as a vehicle is located in the detection target area, the in-area measurement that constitutes the object is performed. The object is not deleted as noise because the number of points (black circles in FIG. 6A) is large and the width dimension of the object is large. On the other hand, as shown in the schematic view of FIG. 6 (b), when an object such as a snowflake (white circle in FIG. 6 (b)) is located in the detection target area, an in-area measurement point (figure Because the number of black circles in 6 (b) is small and the width dimension of the object is small, the object is eliminated as noise.

  Further, in the object detection device 1 of the present embodiment, the signal processing unit 3 determines that the number of in-area measurement points included in the object in the deletion step (step S7) is equal to or less than the configuration score threshold and below the in-area measurement points. If there is an out-of-area measurement point adjacent to the point, it is determined that the object (that is, the in-area measurement point that constitutes the object) is noise. According to the object detection device 1 of this embodiment, as shown in the schematic view of FIG. 7A, in the case of a non-floating object such as a vehicle, an in-area measurement point (an object) The object is not deleted as noise because the number of black circles in FIG. 7A is large and the out-of-area measurement point does not exist below the in-area measurement point configuring the object. On the other hand, as shown in the schematic view of FIG. 7 (b), in the case of an object such as a snowflake (white circle in FIG. 7 (b)), an in-area measuring point (black circle in FIG. 7 (a)) constituting the object. The object is eliminated as noise because the number of) is small and an out-of-area measurement point occurs below.

  Further, in the object detection device 1 of the present embodiment, the signal processing unit 3 increases the distance from the laser sensor unit 2 to the in-area measurement point as the object assumed area (that is, the object based on the in-area measurement point). Determine the estimated shape of) largely. For this reason, in the area | region where the resolution of the laser sensor unit 2 falls, it becomes possible to prevent that a detection target object is deleted as noise and to improve detection accuracy of a detection target object more.

  Further, in the object detection device 1 of the present embodiment, the signal processing unit 3 measures the in-area measurement point when the distance from the laser sensor unit 2 to the in-area measurement point is at a predetermined distance or more. The object based on is the target of determination of deletion. That is, in the object detection device 1 of the present embodiment, when the distance from the laser sensor unit 2 to the in-area measurement point is shorter than a preset threshold, it is determined whether the in-area measurement point is noise. Do. For this reason, in the area | region where the resolution of the laser sensor unit 2 falls, it becomes possible to prevent that a detection target object is deleted as noise and to improve detection accuracy of a detection target object more.

  Although the preferred embodiments of the present invention have been described above with reference to the accompanying drawings, it goes without saying that the present invention is not limited to the above embodiments. The shapes, combinations, and the like of the constituent members shown in the above-described embodiment are merely examples, and various changes can be made based on design requirements and the like without departing from the spirit of the present invention.

  For example, in the above embodiment, as the distance from the laser sensor unit 2 to the in-area measurement point is larger, the shape of the object assumed area is enlarged or the distance from the laser sensor unit 2 to the in-area measurement point is preset. When the length is shorter than the threshold, it is determined whether or not the in-area measurement point is noise. However, the present invention is not limited to this. Regardless of the distance from the laser sensor unit 2 to the in-area measurement point, the shape of the object assumed area is the same, and the in-area measurement point is noise regardless of the distance from the laser sensor unit 2 to the in-area measurement point It is also possible to adopt a configuration that makes a determination as to whether or not.

1 ... object detection device 2 ... laser sensor unit 3 ... signal processing unit 3 a ... extraction unit 3 b ... search unit 3 c ... grouping unit 3 d ... conversion unit 3 e ... noise deletion unit 3 f ... object combination Unit 3g ... Determination unit 3h ... Object information addition unit 3i ... Output unit 4 ... Unit control unit

Claims (6)

A laser sensor unit capable of emitting a laser beam for measurement and changing the emitting direction and receiving the reflected light of the laser beam and outputting it as an electric signal, and detecting an object based on an output signal of the laser sensor unit An object detection device including a signal processing unit to
The signal processing unit extracts an in-area measurement point located in a preset detection target area, and based on an out-of-area measurement point adjacent to the in-area measurement point and located outside the detection target area. An object detection apparatus characterized in that it is determined whether the in-area measurement point is noise.   The signal processing unit determines the horizontal width dimension of the object indicated by the in-area measurement point based on the out-of-area measurement point adjacent to the side of the in-area measurement point, and the in-area measurement included in the object The in-area measurement point is determined to be noise when the number of points is less than or equal to a preset threshold and the width dimension is less than or equal to a preset threshold. Object detection device.   In the signal processing unit, the number of in-area measurement points included in the object indicated by the in-area measurement point is equal to or less than a preset threshold value, and an out-of-area measurement point exists below the in-area measurement point The object detection device according to claim 1 or 2, wherein the in-area measurement point is determined to be noise in the case where it is determined.   The said signal processing part calculates | requires largely the presumed shape of the object on the basis of the said in-area measurement point, so that the distance from the said laser sensor unit to the said in-area measurement point is large. An object detection device according to any one of the preceding claims.   When the distance from the laser sensor unit to the in-area measurement point is shorter than a preset threshold value, the signal processing unit may cause the in-area measurement point to be noise based on the adjacent out-of-area measurement point. The object detection device according to any one of claims 1 to 4, wherein it is determined whether or not it is. An object detection method using a laser sensor unit capable of emitting a measurement laser beam and changing an emission direction and receiving a reflected beam of the laser beam and outputting it as an electric signal.
The in-area measurement point is extracted based on the out-of-area measurement point adjacent to the in-area measurement point while extracting the in-area measurement point located in the predetermined detection target area. An object detection method characterized in that it is judged whether it is noise or not.

Images