JP2006021641A - Vehicle periphery display device and its display method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a vehicle periphery display device for displaying the position of an object around own vehicle without feeling odd. <P>SOLUTION: A relative position of another vehicle around the own vehicle is calculated from a photographed image of a camera 1 by an existence position calculation part 23 and the calculated relative position data are accumulated on an existence position history memory part 25 for a predetermined time. A grouping processing part 26 performs grouping processing of first stage, eliminates the relative position data having no continuity caused by erroneous detection and pitching of the vehicle and performs grouping of the relative position data of the predetermined time by grouping processing of second stage. Average coordinate and covariance matrix are calculated by a statistic calculation part 27 in every group of the relative position data and reliability distribution is calculated at a reliability distribution calculation part 28. Monitor display of the existence position of another vehicle sets an object marker having a size corresponding to a size of theory error onto a plane coordinate making the own vehicle viewed from above as a center and displays it with brightness corresponding to a value of reliability added to the object marker. <P>COPYRIGHT: (C)2006,JPO&NCIPI

Description

  The present invention relates to a vehicle periphery display device that detects other vehicles around the host vehicle and measures a relative position with the host vehicle, and a display method of the measured relative position of the other vehicle.

  Conventionally, a vehicle that detects other vehicles in the vicinity of the host vehicle while traveling and presents the positional relationship between the detected other vehicle and the host vehicle to a driver or the like or issues an alarm according to the distance from the host vehicle. There is a peripheral display device. As such a vehicle periphery display device, a laser radar is mounted on the vehicle, and from the time from the emitted laser pulse to the reflected laser pulse reception, the presence of other vehicles around the own vehicle and the distance from the own vehicle to the other vehicle There are known ones that perform calculation and those that include a camera mounted on a vehicle and calculate the presence or absence of other vehicles around the host vehicle and the distance from the host vehicle to another vehicle from an image captured by the camera.

For example, in a vehicle periphery display device that sequentially measures the inter-vehicle distance with a preceding vehicle with a laser radar for each predetermined period and displays it on a monitor, the light emitting part of the laser radar head attached to the front part of the vehicle has the direction of its optical axis, The initial setting is such that the pulse laser beam is emitted toward the rear reflector or the like of the preceding vehicle. However, if the attitude of the vehicle changes upward or downward due to pitching or the like on an uneven road, the optical axis of the laser is also swung upward or downward following this change in the attitude of the vehicle.
The inter-vehicle distance measured at this time becomes infinite when the optical axis of the laser is swung upward, for example, when the optical axis deviates from the preceding vehicle and heads toward the sky, while the laser optical axis is swung downward. In some cases, the measured inter-vehicle distance may vary greatly and include an error, for example, the optical axis may deviate from the preceding vehicle toward the road surface, resulting in a shorter distance than the actual inter-vehicle distance.

When a measurement error due to pitching occurs, sequentially displaying an unstable inter-vehicle distance that fluctuates until the pitching converges on the monitor may impair a comfortable information display environment in the vehicle interior.
In order to prevent this, a technique is disclosed in which, when the inter-vehicle distance exceeds a predetermined measurement upper limit value, it is determined that a measurement error due to pitching has occurred, and the monitor display of the inter-vehicle distance is not updated (Patent Document 1). reference).
JP-A-10-31800

In such a conventional method, the display is not updated because an error due to pitching occurs when the measurement upper limit is exceeded, but measurement errors due to factors other than pitching are considered. There was a problem that not.
As a measurement error due to factors other than pitching, for example, an error caused by a measurement technique can be considered.

As an example of an error caused by the sensor measurement method, a distance calculation method using a camera as a distance sensor will be described below.
For example, the number of pixels in the vertical direction from the lowest end of the image to the ground contact position of the vehicle is shot with a camera mounted in a predetermined orientation so that the camera optical axis is horizontal to the ground b is detected. The distance L to the vehicle in the predetermined area that has been photographed is determined using the height h of the camera optical axis from the ground and the total number of pixels a in the vertical direction on the photographed image,
L = h / tan ((ab) / a)
Is detected.
In this case, the distance per pixel at a farther position on the photographed image indicates a larger distance, and the distance resolution depends on the resolution.
Therefore, a distant vehicle vibrates unstablely in the calculated distance just like the case of laser radar based on vehicle pitching, as the position on the captured image changes only one pixel in vibration.

  In order to solve the above problems, the present invention provides a vehicle periphery display device and a display that displays the distance to an object around the host vehicle without a sense of incongruity even when there is a distance measurement error caused by a sensor measurement method or the like. It aims to provide a method.

  For this reason, the present invention acquires relative position information from the host vehicle to surrounding objects, further acquires the theoretical error of the relative position information, and uses the relative position information and the theoretical error to calculate the object position information. The target position information is calculated and displayed on the monitor.

  According to the present invention, the object existence position information indicating the existence position of the object is calculated using the theoretical error of the acquired relative position information, so that the object existence position information according to the magnitude of the theoretical error is generated, By sequentially displaying the numerical values of the conventionally detected inter-vehicle distance, it is possible to reduce the uncomfortable feeling that the inter-vehicle distance vibrates unstablely due to a theoretical error.

Embodiments of the present invention will be described below.
FIG. 1 is a block diagram of a vehicle periphery display device according to the present embodiment.
The vehicle periphery display device according to the present embodiment detects a vehicle 1 that acquires a captured image for detecting a traveling vehicle around the host vehicle 5 (see FIG. 2), and detects other vehicles by performing image processing on the captured image. The vehicle relative display control unit (hereinafter referred to as vehicle peripheral display C / U) 2 for calculating the relative position of the other vehicle from the own vehicle 5 and displaying the relative position, and the detected relative position of the other vehicle. It is comprised from the monitor 3 displayed on a passenger | crew.

  The vehicle periphery display C / U2 includes an image processing processor including a vehicle detection unit 21 that performs image processing on a captured image and detects a vehicle, and an image storage unit 22 that temporarily stores the captured image from the camera 1 and image data being processed. Based on the vehicle detection results from the unit 11 and the vehicle detection unit 21, the relative position of the other vehicle from the host vehicle 5 is calculated, the relative position data is processed, and the object marker is used as the object presence position information of the present invention. And an arithmetic control unit 12 that graphically displays.

The calculation control unit 12 includes an existence position calculation unit 23, a theoretical error storage unit 24, an existence position history storage unit 25, a grouping processing unit 26, a statistic calculation unit 27, a certainty factor distribution calculation unit 28, and a display control unit 29. Has been.
Although one camera 1 that captures the surroundings of the host vehicle 5 is shown as a representative in FIG. 1, the number of cameras 1 may be arbitrary. For example, as shown in FIG. The surroundings of the host vehicle may be monitored using four cameras: a camera 1B that captures the rear, a camera 1L that captures the left rear side of the host vehicle, and a camera 1R that captures the right rear side of the host vehicle.

The image processing processor unit 11 may be individually prepared for each of these cameras, or the vehicle detection may be performed by shifting the timing of camera image acquisition and image processing by one image processing processor unit 11.
In any case, the timing of image acquisition and image processing from each camera is controlled by the arithmetic control unit 12 in the image processing processor unit 11.

The functions of the image processing processor unit 11 and the calculation control unit 12 will be described below.
The vehicle detection unit 21 detects a vehicle from an image taken by the camera 1. As a vehicle detection method, for example, the method disclosed in JP-A-8-320999 is used. In this method, the vehicle is detected by paying attention to the point where the vertical edge detected as the side surface of the vehicle is interrupted at the lowermost end of the vehicle and the left-right symmetry that is a characteristic of the vehicle. Image processing is performed on the captured image to detect a vertical edge in the image, and the point at which the vertical edge is interrupted is taken as the vehicle bottom end candidate on the left or right of the rear of the vehicle or the front of the vehicle. When the vehicle bottom end candidates are substantially the same height, it is determined that the detected object is a vehicle.

  When the vehicle detection unit 21 detects the vehicle, the vehicle detection unit 21 assigns an identification number m to the detected vehicle, and the vehicle lower end coordinates (horizontal and vertical pixel positions of the vehicle lowermost end) and size (horizontal of the vehicle) on the image. The number of pixels in the direction), camera shooting time data t, and camera number n are added to the calculation control unit 12 and output. The arithmetic control unit 12 has a RAM (not shown) and temporarily stores the data in the RAM.

The existence position calculation unit 23 of the arithmetic control unit 12 controls image acquisition and image processing timing in the image processing processor unit 11. Further, based on the vehicle detection result from the vehicle detection unit 21, the relative position from the own vehicle 5 as the presence position of the other vehicle is calculated.
A method for calculating the relative position will be described below. The existence position calculation unit 23 reads the vehicle bottom end coordinates and size from the RAM, and in the camera image coordinate system with the center coordinates of the detected camera image as the origin, determines the horizontal intermediate point coordinates of the vehicle bottom end coordinates. As a representative point of the vehicle, a horizontal distance and a relative direction from the mounting position of the camera 1 to the other vehicle are calculated from the focal length of the camera 1, the mounting height of the camera 1, and the mounting direction of the camera 1.
It should be noted that data such as the focal length of the camera 1, the mounting height of the camera 1, and the mounting direction of the camera 1 are stored in advance in a non-illustrated nonvolatile memory of the arithmetic control unit 12.

The existence position calculation unit 23 further corrects the vehicle left-right and front-back distance differences between the reference point M of the host vehicle 5 (for example, the plane center of the host vehicle 5) and the mounting position of the camera 1, and The relative position from the reference point M is converted.
At this time, the relative position from the host vehicle 5 is that the forward direction with the reference point M as the origin is the y coordinate (plus), the backward direction is the y coordinate (minus), the right direction is the x coordinate (plus), and the left direction is x Expressed in relative coordinates (minus). Further, data related to the mounting position of the camera 1 necessary for conversion into a relative position from the reference point M of the host vehicle 5 is also stored in advance in a non-illustrated nonvolatile memory of the arithmetic control unit 12.
The relative position data with the reference point M of the host vehicle 5 as the origin corresponds to the relative position information of the present invention.

The existence position calculation unit 23 adds the relative position data {x _mnt , y _mnt } from the calculated reference point M to the camera photographing time data t, the camera number n, and the identification vehicle number m, and the existence position history storage unit The data is output to 25 and stored.
Here, m = 0, 1, 2,..., Mt, and Mt means that the number detected as a vehicle from the image acquired by the camera n at time t was Mt.

The existence position history storage unit 25 stores the relative position data obtained from the camera images of all the cameras 1 at a predetermined cycle, deletes the old relative position data, and determines the predetermined time (hereinafter, the upper limit of the width of the predetermined time). The relative position data {x _mnt , y _mnt } of the other vehicle detected from the camera image photographed at the time t1 to t2 is displayed for convenience).

The theoretical error storage unit 24 calculates the theoretical error k _x {x _mnt , y _mnt } of the x coordinate, which is the theoretical error in calculating the relative position data {x _mnt , y _mnt } based on the image captured by the camera n, and the y coordinate. The theoretical error k _y {x _mnt , y _mnt } is calculated in advance from the resolution data, focal length accuracy, mounting position accuracy, etc. of the camera n, and the function of the relative position data {x _mnt , y _mnt } or the form of a lookup table Is stored in.

The grouping processing unit 26 first creates a histogram for each camera n and identification vehicle number m that captures the distribution of the relative position data {x _mnt , y _mnt } stored in the presence position history storage unit 25, for example, k-mean A first-stage grouping process is performed by a clustering technique. There is no continuity in the grouping process, and the x-coordinate theoretical error k _x {x _mnt , y _mnt } and the y-coordinate theoretical error k _y {x _mnt , y _mnt } read from the theoretical error storage unit 24 are Thus, the relative position data {x _mnt , y _mnt } greatly _deviating is regarded as relative position data due to erroneous detection and is deleted from the existing position history storage unit 25.
A histogram of the distribution of the relative position data {x _mnt , y _mnt } of all the cameras 1 that remain without being deleted is created again, and a second-stage grouping process is performed by the k-mean clustering method, and the relative position data {x _mnt , y _mnt } are classified into grouped relative position data { _xigt , _yigt }.
Here, g represents a group number (g = 1 to Nt).

ここで、Ｉ_ｇｍａｘはグループｇの所定時間（時刻ｔ１〜ｔ２の間）に含まれる相対位置データ｛ｘ_ｉｇｔ，ｙ_ｉｇｔ｝の総数を示す。 The statistic calculation unit 27 calculates the following statistic for each grouped relative position data { _xigt , _yigt }.
First, the average coordinates {x _cg , y _cg } of the relative position data { _xigt , _yigt } of each group are calculated by the following formula.

Here, _Igmax indicates the total number of relative position data { _xigt , _yigt } included in a predetermined time (between times t1 and t2) of group g.

である。 Further, a covariance matrix R with respect to the average coordinates {x _cg , y _cg } is obtained as follows.

It is.

である。
（ν_ｘｇ，ν_ｙｇ）は、グルーピングされた相対位置データ｛ｘ_ｉｇｔ，ｙ_ｉｇｔ｝の分布のｘ座標およびｙ座標をそれぞれカメラ撮影時刻ｔとｘ座標、カメラ撮影時刻ｔとｙ座標との２次元分布に投影し、各２次元分布において、ｘ座標、ｙ座標の時間に対する変化率の主軸の傾きを求めることによって得られる。 The certainty factor distribution calculation unit 28 calculates a certainty factor distribution P (x, y), which is an index indicating the probability that another vehicle exists at the relative position (x, y).
At this time, from the relative position data included in the location history storing unit 25 with the stored predetermined time (between time t1 to t2), consider the confidence distribution in the current time t _f.
First, the center coordinates (x _ug , y _ug ) of the distribution of the relative position data of the group g at time t _f are obtained using the equation (4).

It is.
(Ν _xg , ν _yg ) is an x-coordinate and a y-coordinate of the distribution of the grouped relative position data { _xigt , _yigt }, which are 2 of the camera photographing time t and the x coordinate, respectively. It is obtained by projecting onto a dimensional distribution and determining the inclination of the principal axis of the rate of change with respect to time of the x-coordinate and y-coordinate in each two-dimensional distribution.

ｔ_ｉｇ：グループｇの相対位置データ｛ｘ_ｉｇｔ，ｙ_ｉｇｔ｝のカメラ撮影時刻
である。 The inclination of the main axis is obtained by the following general principal component analysis. Here, the x coordinate is illustrated.

t _ig is the camera shooting time of the relative position data {x _igt , y _igt } of group g.

この結果得られる固有値の内、最大となる固有値λ_１ｇに対して、式（７）から固有ベクトルＡｇ＝［ａ_１ｇ，ａ_２ｇ］を求める。

である。
ν_ｘｇは式（８）で与えられる。

ｙ座標についても同様に求められる。 Next, the eigenvalue λ _{g of} V _g is obtained. The eigenvalue λ _g satisfies Expression (6).

Among the eigenvalues obtained as a result, the eigenvector Ag = [a _1g , a _2g ] is obtained from the equation (7) for the maximum eigenvalue λ _1g .

It is.
ν _xg is given by equation (8).

The y coordinate is obtained in the same manner.

ここで、νとｕ_ｇは式（１０）、（１１）で定義される二次元ベクトルである。

確信度の分布Ｐ（ｘ，ｙ）は、上式の二次元正規確率分布を（１２）式で示す中心座標（ｘ_ｕｇ，ｙ_ｕｇ）における理論誤差Ｑ（ｘ_ｕｇ，ｙ_ｕｇ）と、中心座標（ｘ_ｕｇ，ｙ_ｕｇ）における確率Ｇ（ｘ_ｕｇ，ｙ_ｕｇ）で割った値であり、式（１３）で表される。

なお、αは任意の定数である。

The two-dimensional normal probability distribution G (x, y) of the relative position distribution of other vehicles in the group g extracted from the camera image at the current time t _f obtained in this way is the center coordinates (x _ug , y _ug ) and the covariance matrix R defined by Equation (3), and expressed by Equation (9).

Here, [nu and _{u g} is the formula (10) is a two-dimensional vector defined by (11).

The certainty distribution P (x, y) has a theoretical error Q (x _ug , y _ug ) and a center at the center coordinates (x _ug , y _ug ) indicating the two-dimensional normal probability distribution of the above equation by the equation (12). This is a value divided by the probability G (x _ug , y _ug ) at the coordinates (x _ug , y _ug ), and is expressed by equation (13).

Α is an arbitrary constant.

ここで、βは任意の係数である。 Based on the calculated certainty distribution P (x, y) and the theoretical error, the display control unit 29 indicates an object marker 31 (FIG. 3, FIG. 3) indicating the relative position of the other vehicle in the group g obtained from the camera image. 4) and the luminance distribution within the display area of the object marker 31 are set. The display control unit 29 causes the monitor 3 to display the size and luminance distribution of the set object marker 31.
The shape of the object marker 31 is, for example, an ellipse, and the size of the ellipse is defined by using the center coordinates (x _ug , y _ug ) of the group g represented by the equation (4) as the center of the object marker 31 (x _{ug -k x (x ug, y ug} ), y ug -k y (x ug, y ug)) and _{(x ug + k x (x} ug, y ug), y ug + k y (x ug, y ug)) These two points are inscribed in a rectangle that is parallel to the front-rear direction of the host vehicle 5 and that has two points at both ends of the diagonal line.
The luminance I (x, y) of the object marker 31 is expressed by Expression (14).

Here, β is an arbitrary coefficient.

Therefore, the luminance distribution within the display area of the elliptical object marker 31 has the highest luminance at the center coordinates (x _ug , y _ug ) where the center of the target marker 31 is located, and the certainty toward the outside of the ellipse. The luminance decreases according to the value of the distribution P (x, y) of. The certainty factor, which is a numerical value between 0 and 1, may be divided into a predetermined number of steps Ni, and the luminance level may be associated with each step, and the luminance may be decreased stepwise from the center of the distribution. If the luminance value of the monitor 3 is, for example, 256 gradations, (luminance value) = 256 × (confidence value) may be set.
3 and 4 show display examples of the object marker 31 indicating other vehicles around the own vehicle in the plane coordinates with the center of the own vehicle 5 seen from above as a reference point M. FIG.
A plurality of predetermined distance zones are provided for the distance from the outermost contour of the host vehicle 5 to the center of the target marker 31, and the target marker 31 having the center in the distance zone close to the host vehicle 5 is more alert. The color is displayed in a high degree of color and indicates the degree of approach to the driver. For example, colors are set such that red is less than 2 m, orange is 2 m or more and less than 5 m, pink is 5 m or more and less than 10 m, and green is 10 m or more.

If the relative speed between the host vehicle 5 and the other vehicle is large or the distance detection accuracy of the other vehicle is poor, that is, if the theoretical error is large, the relative position data obtained in a predetermined time (between times t1 and t2) varies. As a result, the spread of the distribution increases and the probability at the center coordinates of the two-dimensional normal probability distribution decreases. Whether the reason why the spread of the distribution is large is due to the relative velocity or the theoretical error cannot be judged only by the two-dimensional normal probability distribution.
Therefore, as the certainty factor, (certainty factor) = (two-dimensional normal probability distribution value) / {(central coordinates of the two-dimensional normal probability distribution) as shown in the equation (13) without using the two-dimensional normal probability distribution. Probability) × (theoretical error in center coordinates)}.

  In the case of a two-dimensional normal probability distribution with the same spread, the probability at the center coordinate of the probability distribution is the same, so the certainty varies depending on the magnitude of the theoretical error at the center coordinate of the distribution. When the spread of the distribution depends on the theoretical error, the certainty factor becomes small because the theoretical error of the denominator is large. On the contrary, when the spread of the distribution is large due to the large relative speed, the certainty factor increases because the theoretical error value of the denominator is small.

Next, the reliability of the other vehicle having a higher relative speed and the other vehicle having a lower relative speed at one point relative to the reference point M of the host vehicle 5 will be compared.
Since the camera coordinate position is the same, the theoretical error is the same in both cases where the relative speed is large and small. When the relative speed is large, the confidence distribution is wide and the attenuation is small as the distance from the center coordinate is increased. When the relative speed is small, the certainty distribution is small and the attenuation increases as the distance from the center coordinate increases.
From Equation (13), when the theoretical error is the same, the certainty factor at the center coordinate is the same regardless of the spread of the distribution.
That is, the certainty factors of other vehicles having different relative speeds at relative positions from the own vehicle 5 having the same theoretical error have the same value at the distribution center, and the higher the relative velocity, the higher the certainty factor region and the smaller the relative velocity. The region with higher certainty is narrowed.

FIG. 5 shows a flowchart of a control flow for displaying the relative positions of other vehicles around the host vehicle on the monitor.
These controls are processed as programs in the image processing processor unit 11 and the calculation control unit 12.
In step 101, the image storage unit 22 acquires a photographed image from the camera 1 together with the camera number n and the camera photographing time t.
In step 102, the vehicle detection unit 21 detects the vehicle by edge processing, and adds the camera bottom coordinate and size of the detected vehicle to the camera number n, the camera shooting time t, and the detected vehicle identification number m, and controls the calculation. To the unit 12. The arithmetic control unit 12 temporarily stores them in the RAM.

In step 103, the presence position calculation unit 23 reads the detected vehicle lower end coordinates and size of the vehicle from the RAM, and calculates relative position data {x _mnt , y _mnt } from the reference point M of the host vehicle 5.
In step 104, the existence position calculation unit 23 sets the relative position data {x _mnt , y _mnt } as a relative position history for a predetermined time (between times t1 and t2) and stores the camera number n, The camera photographing time t and the detected vehicle identification number m are added and stored.
In step 105, the existence position calculation unit 23 calculates the theoretical error k _x {x _mnt , y _mnt } of the x coordinate corresponding to the relative position data {x _mnt , y _mnt }, the theoretical error k _y {x _mnt , y _mnt } is acquired from the data stored in the theoretical error storage unit 24 and stored in the existing position history storage unit 25 in correspondence with the relative position data {x _mnt , y _mnt }.

In step 106, the grouping processing unit 26 performs a grouping process.
First, the relative position data {x _mnt , y _mnt }, the x-coordinate theoretical error k _x {x _mnt , y _mnt }, and the y-coordinate theoretical error k _y {x _mnt , y _mnt } are stored from the existing position history storage unit 25. The first stage grouping process is performed on the relative position data {x _mnt , y _mnt } corresponding to the camera n and the vehicle identification number m for a predetermined time (between times t1 and t2). Relative position data {x _mnt , y _mnt } is deleted. Next, a second-stage grouping process for grouping relative position data for a predetermined time (between times t1 and t2) by all the cameras n is performed.

In step 107, for each grouped relative position data group, the statistic calculator 28 calculates the average coordinates according to the expressions (1) and (2), the statistic of the covariance matrix R according to the expression (3), and the expression ( The center coordinates are calculated from 4) based on equation (8).
In step 108, the certainty factor distribution calculation unit 28 acquires a theoretical error for the center coordinates (x _ug , y _ug ) from the theoretical error storage unit 24.
In step 109, confidence factor distribution calculation unit 28, the confidence distribution of presence positions of other vehicles detected at the current time t _f using theoretical error obtained by the statistics and step 108 calculated in step 107 It calculates based on Formula (13).

In step 110, the display control unit 29, the size Satoshi object marker 31 the magnitude of the theoretical errors in the center coordinates of the confidence distribution group g at the current time t _f, around the center coordinates of the confidence distribution target The object marker 31 is set, and the luminance within the display area of the object marker 31 set by Expression (14) is set.
In step 111, the monitor 3 displays the position of the other vehicle with the position of the target marker 31 set by the display control unit 29 and the size and brightness of the target marker 31.

3A and 3B, the detected relative position of the other vehicle is far from the own vehicle 5, and the relative position theoretical errors k _x (x _ug , y _ug ), k _y (x _ug , y _ug). ) Is relatively large.
In the case of position detection using a camera image, the distance is calculated from the coordinate position of the pixel. Therefore, the distance from one vehicle increases as the distance from the host vehicle 5 increases, and the theoretical error increases. It becomes.
4A and 4B, the detected relative position of the other vehicle is close to the own vehicle 5, and the relative position theoretical errors k _x (x _ug , y _ug ), k _y (x _ug , y _ug). ) Is small.

Further, when the relative speed between the host vehicle 5 and the other vehicle is small, the change in the relative position is small, and as shown in (a) of FIG. 3 or (a) of FIG. It is displayed narrowly in the direction.
When the relative speed between the host vehicle 5 and the other vehicle is large, the change in the relative position is large, and as shown in FIG. 3B or FIG. 4B, the luminance is strong along the change direction of the relative position. A uniform central portion is displayed for a long time.

  Steps 101 to 104 in the flowchart of the present embodiment are the relative position information acquisition means of the present invention, steps 105 and 108 are the theoretical error acquisition means, and steps 106, 107, 109, and 110 are the object presence position information calculation. Configure the means.

As described above, according to the present embodiment, relative position data that is the position of another vehicle for a predetermined time (between times t1 and t2) is accumulated and detected by each camera n as a first-stage grouping process. The relative position data {x _mnt , y _mnt } is grouped for each vehicle identification number m, and the relative position data having a change larger than the theoretical error is not continuous as one group among the plurality of relative position data. Delete as false positive data.
As a result, in the conventional method, when an error within the upper limit of measurement occurs due to pitching, the displayed inter-vehicle distance oscillates unstable around the true value, and the inter-vehicle distance fluctuation displayed on the monitor is displayed. The problem of unnaturalness is improved, and the passengers are not discomforted.

  In addition, since the target vehicle is extracted by performing image processing on the captured image of the camera, optical disturbances such as direct incidence of external light on the camera and erroneous detection of shades and structures similar to the vehicle are likely to occur. In addition, misdetection data caused by extraneous light incident on the camera 1 is deleted from relative position data having no continuity as a group among a plurality of relative position data. Thus, the problem that the driver is prevented from recognizing the surrounding situation is improved, and the occupant is not discomforted.

Next, as a second stage grouping process, the entire relative position data {x _mnt , y _mnt } for a predetermined time (between times t1 and t2) when all the cameras are not deleted is grouped, The relative position data {x _mnt , y _mnt } of a plurality of other vehicles is _regrouped into the relative position data {x _igt , y _iigt } for each group g. Sonouede, and calculates the certainty factor distribution P (x, y) at the current time t _f to the relative position data group for each group g.

Thereafter, the display of the position of the other vehicle is displayed on the plane coordinate centered on the host vehicle 5 viewed from above, and the center coordinate of the certainty distribution P (x, y) for the relative position data group of the group g is set as the object marker. The size of the object marker 31 is determined based on the theoretical error in the center coordinates, and the luminance within the display area of the object marker 31 is set based on the certainty factor.
Therefore, the location of the other vehicle around the host vehicle can be displayed at a glance, and the driver can easily visually recognize the position of the other vehicle around.
Further, since the size of the object marker 31 is set according to the magnitude of the theoretical error, the reliability of the relative position data is easily understood at a glance.

Further, in other vehicles having a large relative speed with the own vehicle 5, a portion with a high luminance is displayed in an elongated manner within the display area of the object marker 31, and other vehicles with a low relative speed with the own vehicle 5 have a portion with a high luminance. Is displayed within the display area of the object marker 31 so that the relative speed with the other vehicle can be easily recognized intuitively.
Thus, since the certainty distribution is calculated and the luminance is displayed according to the value of the distribution, for example, the certainty distribution is not calculated, and the grouped relative position data is displayed simply as a dot display. Intuitively understand the relative speed.
Furthermore, since the object marker 31 of the vehicle that is closer to the outermost contour of the host vehicle 5 has a color with a high alerting degree, the driver can easily recognize the degree of approach.

In the present embodiment, the size of the object marker 31 for displaying the detected position of the other vehicle is determined based on the theoretical error. However, the region indicating the certainty level equal to or higher than a predetermined value in the belief distribution. May be defined as the size of the object marker 31. For example, the region of the certainty distribution of 0.5 or more is set as the size of the object marker.
In that case, when the relative speed with the other vehicle is large, the object marker increases in a shape extending in the direction of change of the relative position. Conversely, when the relative speed with the other vehicle is small, the object marker is relatively It becomes a small circle or ellipse.
Therefore, it can be understood intuitively whether the relative position has changed.

Furthermore, starting from the average coordinates {x _cg , y _cg } of the relative position data for a predetermined time (between times t1 and t2), the center coordinates (x of the certainty distribution P (x, y) at the current time t _{f ug} , y _ug ) as an end point, a relative displacement vector is defined, and an arrow indicating a vector direction for determining whether the relative displacement vector is in the direction of the host vehicle or on the other hand is a target object. It may be added near the marker 31.
In this manner, the direction of change of the relative position in time series can be determined based on the history of the relative position data within a predetermined time, and the occupant can always determine whether another vehicle is approaching the host vehicle 5 or not. Intuitive judgment.

In this embodiment, the distribution of the certainty value within the display area of the object marker 31 is displayed as a change in luminance. Instead, the larger the certainty value, the higher the certainty value is set according to the intensity of the same color. In this way, the confidence distribution may be displayed. The method of displaying the confidence distribution is not limited to this, and the color itself may be changed. For example, the order of red, orange, yellow, and blue may be set in order from the highest degree of certainty.
When the color of the object marker itself is changed according to the certainty value, the object marker may be trimmed with a color indicating the degree of approach.

Next, a first modification of the present embodiment will be described.
In this modification, after the first and second stage grouping processes of the embodiment, the average coordinates {x _cg , y _cg } of the relative position data for a predetermined time (between times t1 and t2) for each group g. And the center coordinates (x _ug , y _ug ) of the distribution of the relative position data at the current time t _f by principal component analysis.
Further, the display of the position of the other vehicle is set to a plane coordinate centered on the host vehicle 5 as viewed from above, and the center coordinate of the relative position data group of the group g is set to the center position of the object marker 31. Based on the theoretical error, the size of the elliptical object marker 31 is set. In addition, the relative position data of the group g included within the display area of the object marker 31 is plotted in, for example, dot display, and the average coordinates {x _cg , y _cg } are used as the starting point, and the center coordinates (x _ug , y _ug). The object marker in the shape of an arrow indicating a vector whose end point is the position of () is also superimposed and displayed.

With such a configuration, the other vehicle is relatively moved from the arrow-shaped object marker based on the accumulation of the relative position data of the other vehicle for a predetermined time (between times t1 and t2). Since the direction is indicated and the size of the elliptical object marker 31 indicates the magnitude of the theoretical error, the occupant can immediately determine the change in the relative position without having to observe the change in the inter-vehicle distance data for a while. Further, the reliability of the relative position can be easily determined by the size of the elliptical object marker 31 indicating the theoretical error.
Note that the relative position data due to erroneous detection is eliminated by the first-stage grouping process, and the relative position display of other vehicles around the host vehicle that is stable and easy to see is the same as in the embodiment.

Further, a second modification of the present embodiment will be described. In this modification, the calculation control unit 12 according to the embodiment includes an existing position calculation unit 23, a theoretical error storage unit 24, and a display control unit 29.
The display control unit 29 acquires the theoretical error corresponding to the calculated relative position coordinate from the theoretical error storage unit 24 without accumulating the relative position data calculated in the existence position calculation unit 23 for a predetermined time, and The center position is the calculated relative position coordinate, and the size of the object marker is set based on the value of the theoretical error as information indicating a range in which another vehicle may exist. The display control unit 29 sequentially displays the object marker thus obtained.
Since the object marker 31 having a large theoretical error is displayed large, even if the display position vibrates at a distance of the magnitude of the theoretical error, the display does not give a sense of incongruity. In addition, the occupant can intuitively grasp the size of the object marker as the range of possibility of existence of other vehicles.

In the embodiment, the first modified example, and the second modified example, the method for acquiring the relative position information is based on the camera, but is not limited thereto. Relative position data may be obtained by detecting other vehicle positions around the host vehicle by scanning with a laser radar or the like.
For example, in a laser radar, it is determined that the measurement error due to factors other than pitching, for example, the pulse laser beam cannot irradiate the reflector at the rear of the preceding vehicle, the reflected laser beam with sufficient intensity cannot be received, and the measurement upper limit is exceeded. Error due to non-detection or false detection of an object such as a case where a pulse laser beam emitted from an oncoming vehicle is erroneously detected as a reflected laser beam, or an error caused by a sensor measurement technique.
In this case, in particular, in the embodiment and the first example, the first-stage grouping process is performed, and there is no continuity as one group among a plurality of relative position data, and the relative position of the change larger than the theoretical error. Delete the data as false positive data.
As a result, in the conventional method, when an error within the upper limit of measurement occurs due to pitching, the displayed inter-vehicle distance oscillates unstable around the true value, and the inter-vehicle distance fluctuation displayed on the monitor is displayed. The problem of unnaturalness is improved, and the passengers are not discomforted.

Note that although the theoretical error in the present embodiment and the first and second modified examples is attributed to the distance calculation method using the pixel position from the camera image, it is not limited thereto. Those that can be clearly quantified as errors in the generated relative position data can be treated as theoretical errors.
Although the theoretical error is calculated in advance and stored in the theoretical error storage unit 24, the present invention is not limited to this. The reliability of the relative position data acquired from time to time may be acquired by detecting an external factor such as pitching of the own vehicle at the acquisition timing, for example, by a vehicle state sensor, and acquiring an error reflecting it.

It is a figure which shows the structure of embodiment of this invention. It is a figure which shows the attachment state of a camera. It is a figure which shows the example of a display of the detected vehicle position. It is a figure which shows the example of a display of the detected vehicle position. It is a flowchart which shows the flow of display control about the detected vehicle position.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1, 1A, 1B, 1L, 1R Camera 2 Vehicle periphery display control unit 3 Monitor 5 Own vehicle 11 Image processing processor part 12 Arithmetic control part 21 Vehicle detection part 22 Image storage part 23 Existence position calculation part 24 Theoretical error storage part 25 Presence location history storage unit 26 Grouping processing unit 27 Statistics calculation unit 28 Certainty distribution calculation unit 29 Display control unit 31 Object marker

Claims (9)

Relative position information acquisition means for acquiring relative position information from the host vehicle to surrounding objects;
A theoretical error acquisition means for acquiring a theoretical error of the relative position information;
Using the relative position information and the theoretical error, target object position information calculating means for calculating target object position information;
A vehicle periphery display device comprising: a monitor for displaying the object presence position information. 2. The vehicle periphery display device according to claim 1, wherein the monitor displays a positional relationship between the object presence position information and the host vehicle in a plane coordinate. The relative position information acquisition means accumulates the relative position information for a predetermined time,
The object presence position information calculating means includes
Based on the accumulated relative position information, calculate the certainty distribution of the position where the object exists in the plane coordinates,
The vehicle periphery display device according to claim 2, wherein a display position and a size of the object presence position information in the plane coordinates are set based on the certainty factor distribution. The object presence position information calculating means includes
4. The vehicle periphery display device according to claim 3, wherein a color or luminance of the object presence position information is set based on the value of the certainty distribution. The relative position information acquisition means accumulates the relative position information for a predetermined time,
The object presence position information calculating means includes
Based on the accumulated relative position information, calculate the certainty distribution of the position where the object exists in the plane coordinates,
Based on the certainty distribution, set the display position of the object presence position information in the plane coordinates, and further color or brightness,
The vehicle periphery display device according to claim 2, wherein the size of the object presence position information is set based on the size of the theoretical error. The object presence position information calculating means calculates the certainty distribution of the position where the object exists in the plane coordinate at the current time from the distribution of the accumulated relative position information. ,
The center position of the certainty distribution is set as the display position of the object presence position information,
6. The vehicle periphery display device according to claim 4, wherein within the display area of the target object location information, the luminance level is set higher or darker as the value of the certainty distribution is higher. The relative position information acquisition means accumulates the relative position information for a predetermined time,
The object existence position information calculating means calculates the position of the average value of the distribution of the accumulated relative position information, and uses the position of the average value as the starting point of the vector,
Further, based on the distribution of the accumulated relative position information, a principal component analysis of the rate of change of the position of the object with respect to time is performed, and a center position where the object exists at the current time is calculated. Let the position be the end point of the vector,
The vehicle periphery display device according to claim 2, wherein the object presence position information is displayed as an arrow shape indicating the vector. The object presence position information calculation unit first performs a grouping process on the distribution of the relative position information in which the relative position information acquisition unit accumulates the relative position information for a predetermined time, and the relative position information having a variation greater than a predetermined value. The vehicle periphery display device according to any one of claims 1 to 7, wherein: Obtaining relative position information from the host vehicle to surrounding objects, obtaining a theoretical error of the relative position information, calculating object presence position information using the relative position information and the theoretical error, A display method in a vehicle periphery display device, wherein a positional relationship between presence position information and a host vehicle is displayed on a monitor in plane coordinates.

Images