JP2002140696A - Image pickup and image conversion device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To automatically detect the variation of a camera pitch angle from picked-up lane marks and automatically adjust a converted image. SOLUTION: The image pickup and image conversion device loaded in a vehicle and fixed on a point C of height (h) and capable of picking up the image of a road surface on the back of the vehicle from a camera and converting the picked up image into a bird's-eye image detects lane marks to be parallel lines and judges whether the lane marks are parallel lines or not. When the lane marks are not parallel lines, the variation dP of the camera pitch angle Pc is obtained from width Wi1 on a position most close to the vehicle in the bird's eye image, width Wi2 on a position most far from the vehicle and an angle P in the vertical direction of the camera in an area to be converted into the bird's-eye image and Pc+dP is substituted for the camera pitch angle to prepare a bird's-eye image again.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image pickup and image conversion apparatus for converting an image picked up by an image pickup means into a converted image by an image conversion means and outputting the converted image. Specifically, there is a method in which a captured image obtained by obliquely imaging a road surface or the like is converted into a plan view obtained by imaging from above before being output to, for example, an image recognition device.
INDUSTRIAL APPLICABILITY The present invention is particularly effective for an imaging and image conversion apparatus that provides a top view (bird's eye view) to a driver or the like in driving a vehicle or the like, maneuvering a ship or an aircraft, or external monitoring.

[0002]

2. Description of the Related Art For example, as an image pickup and image conversion apparatus for picking up an image of a road surface and a road condition ahead or behind to provide a top view (bird's eye view) to a driver of a vehicle, see Japanese Patent Application Laid-Open Publication No. HEI 9-29139.
No. 46727 and JP-A-7-218280 are known. These are to fix the imaging device including the imaging direction in advance as designed, and to input parameters.

[0003]

By the way, when the vehicle body sinks or leans depending on the number of passengers and the amount of luggage, for example, like a passenger car, the position and the imaging direction of the imaging device are designed or input. It often happens that the parameters differ from the parameters set. Then, for example, while traveling on a road surface with completely parallel lane marks,
Even if a plan view (bird's eye view) from above is obtained from the imaging and image conversion device, distortion occurs.

This is shown in FIG. As shown in FIG. 8A, when the pitch angle Pc of the camera C provided in the vehicle M (the angle between the vertical line passing through the lens center of the camera C and the imaging direction of the camera) is α, the imaging is performed. It is assumed that the processing device is configured so that the image (FIG. 8B) is a correct converted image (FIG. 8C). That is, it is assumed that, when parallel lines (thick line segments and broken lines) on the road surface are captured, the converted image is adjusted to be parallel lines. At this time,
When the pitch angle of the camera becomes large as α + dα (dα> 0) or becomes small as α−dα (dα> 0), the captured image is low in FIG. Position) and (e) (the horizontal line H is at a higher position in the captured image). Then, the image conversion method changes the area surrounded by the dotted line in (b), (d), and (e) of FIG. 8 into the screen frame of (c), (f), and (g) in FIG. Since conversion is to be performed, (d) and (e) of FIG.
Will be as shown in FIGS. 8F and 8G. That is, there is a problem that the converted image does not become a correct plan view (bird's eye view) from above due to the inclination of the vehicle, and correct information is not instructed to the driver even when the converted image is output to the image recognition device.

SUMMARY OF THE INVENTION The present invention has been made to solve the above-mentioned problems, and an object of the present invention is to provide an image pickup and image conversion apparatus configured to automatically correct the image pickup direction of an image pickup apparatus and output a corrected image. It is to provide a device.

[0006]

According to a first aspect of the present invention, there is provided an image capturing apparatus comprising: an image capturing unit configured to convert an image captured by the image capturing unit and output a converted image; In an imaging and image conversion apparatus having an image analysis means for analyzing a captured image, an imaging direction parameter of the imaging means of the image conversion means is corrected based on an analysis result of the image analysis means, and the corrected converted image is output. It is configured to output. This is shown as a block diagram in FIG. Incidentally, as the imaging direction parameter,
The position of the imaging means (camera height, etc.), the angle (pitch angle) between the imaging center direction and a straight line vertically lowered from the imaging means, and the angle between the plane formed by them and the set imaging center direction set (Horizontal blur angle) and the torsion angle of the image pickup means with the image pickup center direction as an axis.

According to a second aspect of the present invention, there is provided an image pickup means, an image conversion means for converting an image picked up by the image pickup means and outputting a converted image, and a conversion image output from the image conversion means. An image capturing and image converting apparatus having an image analyzing means for analyzing, wherein an image capturing direction parameter of the image capturing means included in the image converting means is corrected based on an analysis result of the image analyzing means, and a corrected converted image is output. It is characterized by having done. This is shown as a block diagram in FIG. The imaging direction parameter is set to the position of the imaging means (camera height or the like), the angle (pitch angle) between the imaging center direction and a straight line vertically lowered from the imaging means, and the plane formed by them. The angle (horizontal shake angle) with the set imaging center direction, and the torsion angle of the imaging unit with the imaging center direction as an axis.

According to a third aspect of the present invention, in the image capturing and image converting apparatus according to the first or second aspect, the image capturing means is attached to a vehicle and mainly captures an image of a road surface. I do. It should be noted that the road surface is not limited to a so-called road, but means all areas where the vehicle moves.
That is, it includes the parking area.

According to a fourth aspect of the present invention, the image analysis means estimates and extracts a parallel line or a known figure on a road surface, and analyzes a captured image or a converted image of the parallel line or the known figure. Thus, the correction amount of the imaging direction parameter is determined. Here, the road surface parallel line is
For example, road shoulder, lane mark, parking area dividing line etc.
The known figure means, for example, a circular manhole, a stop line, a traveling direction, and other road markings.

According to a fifth aspect of the present invention, the image analyzing means analyzes the captured image or the converted image of the optical flow on the road surface by interlocking with the speed measuring means of the vehicle to obtain the image capturing direction. It is characterized in that a parameter correction amount is determined. Here, the optical flow refers to a flow recognized by the image analysis unit as moving on the captured image or the converted image, such as a dashed lane mark formed on the road surface.

Further, according to the present invention, two or more image pickup means are provided at a predetermined distance, and the image analysis means is provided with the image picked up by the plurality of image pickup means or the image conversion. The correction amount of the imaging direction parameter is determined by analyzing the converted image by the means.

Further, according to the present invention, the image pickup means is provided with a posture detecting means for detecting a direction of the image pickup means, and the image pickup means is provided with respect to the vehicle detected by the posture detecting means. The correction amount of the imaging direction parameter is determined using the direction parameter together. Here, the attitude detecting means means a gyro, a combination of an acceleration detector and an arithmetic unit, and the like.

[0013]

According to the first and second aspects of the present invention, even when the imaging means is not oriented in the imaging direction as set, the imaging means or the converted image is analyzed to analyze the captured image or the converted image. The deviation from the set value of the direction can be obtained. Therefore, a corrected converted image can be obtained by correcting the imaging direction parameter in the image conversion unit without adjusting the imaging unit itself. Therefore, fine adjustment is not required even when the imaging unit is fixed. Such an imaging and image conversion apparatus is particularly effective for an imaging and image conversion apparatus that forms a plan view (bird's eye view) from above after, for example, obliquely imaging a road surface or the like. Thereby, for example, in conjunction with the image recognition device, more accurate recognition of the surrounding situation can be performed, and the surrounding situation information can be given to the driver, the monitor, and the like.

If the imaging means is mounted on the vehicle and analyzes the captured image of the road surface or the converted image, the correction amount of the imaging direction parameter can be easily determined. In particular, according to the analysis of parallel lines, known figures, and optical flows on the road surface, it is easy and reliable (claims 4 and 5). If a plurality of image pickup means are provided, the image pickup direction can be reliably determined from the stereo image, and the correction thereof is easy (claim 6). In addition to these, if it has the attitude detection means of the imaging means itself,
Correction becomes easier (claim 7).

The present invention also includes a combination of the first and second aspects of the present invention. That is, the image processing apparatus includes first image analysis means for analyzing a captured image, and first image analysis means for analyzing a converted image, and one or more imaging direction parameters based on the analysis result of the first image analysis means. And the second
And correcting one or more imaging direction parameters based on the analysis result of the image analysis means. One or a plurality of imaging direction parameters for which the correction amount is easy to determine by analyzing the captured image and one or a plurality of imaging direction parameters for which the correction amount is easy to determine by analyzing the converted image, respectively, by the first image analysis means and the second image analysis means. And image analysis means. Thereby, the correction amount can be easily determined.

The present invention also includes a combination of claims 4 to 7 of the present invention. For example, for an image that only substantially fluctuates on the scale of the converted image, such as a fluctuation only in the camera height, a correction amount is obtained from the relationship with the vehicle speed by means of claim 5, and other imaging direction parameters are obtained. Is corrected according to claim 4.
Further, for example, a gyro is incorporated in the image pickup means as the posture detecting means of claim 7 so that an image pickup direction parameter such as a torsion angle, which is difficult from image analysis, can be corrected, and other image pickup direction parameters can be corrected. 5 for correction. In addition, any combination of claims 4 to 7 is included in the present invention.

[0017]

DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, specific embodiments of the present invention will be described. The following embodiment is a specific example including correction of a pitch angle, but other configurations of the present invention can be easily configured by learning the following embodiment.

[First Embodiment] FIG. 2 is a block diagram showing the configuration of an image pickup and image conversion apparatus 10 according to a first embodiment of the present invention. It is assumed that a camera C is mounted on a vehicle M as shown in FIG. An image captured by the camera (imaging unit) 1 in FIG. 2 is input to a viewpoint conversion unit (image conversion unit) 2. Here, the image is converted into a bird's-eye view image based on the imaging direction parameter which is an initial setting value. Here, the bird's-eye view image is output to the image analysis unit 3. The image analysis unit 3 includes a lane mark detection unit 31 and a parameter correction amount calculation unit 32. The lane mark detector 31 detects a lane mark to be marked as a parallel white line. Here, the lane mark may not be a vertical parallel line group in the image, but may be inclined. Next, in the parameter correction amount calculation unit 32,
When the lane mark is inclined, the correction amount dP of the pitch angle Pc of the camera 1 is calculated and output to the viewpoint conversion unit 2. The viewpoint conversion unit 2 calculates the correction amount d of the pitch angle Pc of the camera 1
Image conversion is performed again based on P. In this way, the routine is repeated until the inclination of the lane mark, which should be marked as a parallel white line in the bird's-eye image, disappears.

FIG. 3 is a flowchart showing a procedure performed by the imaging and image conversion apparatus 10 according to the present embodiment as a whole.
"S", such as S1, S2, etc., means a step in the flowchart. First, in S1, an image captured by the camera 1 is captured. Next, in S2, viewpoint conversion is performed using the imaging direction parameter. This is the operation of the viewpoint conversion unit 2. Next, in S3, two lane marks are detected from the bird's-eye view image. Next, in S4, the width W of the two lane marks is determined.
_i1 and _Wi2 are calculated. Next, the difference between the widths _Wi1 and _Wi2 is determined in S5. If _Wi1 = _Wi2 , the process proceeds to S6, where a bird's-eye image is output. If W _i1 ≠ W _i2 , there is an inclination of the lane mark, so the process proceeds to S7, where the correction amount dP of the pitch angle Pc is calculated, and the process proceeds to S8, where the pitch angle Pc of the imaging direction parameter is set to Pc.
Replace with + dP and return to S2.

The method of calculating the correction amount dP of the pitch angle Pc of the camera 1 from the inclination of the lane mark of the bird's-eye image is as follows. As shown in FIG. 4 (a), the camera C provided at the height h, the foot of the perpendicular lowered to the road surface G, G
_1, and the ranges there bird's-eye image from G ₁ to the rear G _2. Let P be the size of the corner G ₁ CG ₂ . Further, as shown in FIG. 4B, it is assumed that points R ₁ and R ₂ are located at a position W which is perpendicular to G ₁ G ₂ from points G ₁ and G ₂ . View point converter 2 in FIG.
Are coincident with the actual imaging direction parameters of the camera C in FIG. 4, the points G ₁ , G ₂ , R
_1, the corresponding point on the bird's-eye image of the _{R 2 I (G 1),} I (G 2), I
For (R ₁ ) and I (R ₂ ), the lengths I (G ₁ ) I (R ₁ ) and I (R ₁ )
(G ₂ ) I (R ₂ ) is equal.

Next, the pitch angle Pc of the camera C is shifted by dP.
Is assumed. That is, as shown in FIG.
The corner G of P₁'CG_Two'Is the corner G₁CG_TwoAlso shifted by dP
Assume that At this time, the magnitude θ₁Corner G₁CR₁And large
Size θ_TwoCorner G_TwoCR_TwoThe area of the square pyramid surrounded by
When the pitch angle Pc of the camera C shifts by dP, the size θ
₁Corner G₁'CR₁'And the magnitude θ_TwoCorner G_Two'CR_TwoSurrounded by '
Change into a quadrangular pyramid region. Where G₁R₁= G_TwoR_Two=
W, G₁'R₁'= W + dW₁, G_Two'R_Two'= W + dW_Twofar
And the following relationship holds.

(Equation 1)

Here, assuming that dP is sufficiently small, the following applies.

(Equation 2)

Using equations (2-1) and (2-2),
Equations (1-1) and (1-2), Equations (1-3) and (1-
From 4), dW ₁ and dP can be obtained as follows.

(Equation 3)

Now, since these are relations in a real image system, what is happening in a bird's-eye view image will be considered. As shown in FIG. 4C, corresponding points I (G ₁ ′), I (G ₂ ′), I (R ₂ ) of the points G ₁ ′, G ₂ ′, R ₁ ′, and R ₂ ′ in the bird's-eye view image. ₁ for _{'), I (R 2'} ),
Length _{I (G 1 ') I (} R 1') and _{I (G 2 ') I (} R 2') are equal. By taking a point r on the real image system line segment G ₂ 'R ₂ ' so that G ₂ 'r = W, the corresponding point I (r) of the point r in the bird's-eye image
Is located on the line segment I (G ₂ ′) I (R ₂ ′) in the bird's-eye image.
Exactly the same applies to the straight line G of the points R ₁ , R ₂ , R ₁ ′, R ₂ ′, r.
_{This is performed} for the line symmetry points L ₁ , L ₂ , L ₁ ′, L ₂ ′, 1 at ₁ G ₂ . Then, a straight line passing through points R ₁ , R ₂ , and r and points L ₁ ,
Assuming two parallel lane marks having a width of 2 W, which are straight lines passing through L ₂ and L ₂ , the point R ₁ ′ and the point L ₁ ′ are obtained from the equation (3-1).
Is clearly on each lane mark. Therefore, a lane mark such as a bold line in FIG. 4C can be assumed.

As described in Japanese Patent No. 2946727, the bird's-eye image is for converting a captured image of a Gaussian imaging system. The ratio of the line segment in the bird's-eye image is determined in the horizontal axis direction (the imaging center line of the camera). (In the direction on the road surface perpendicular to the vertical axis) in the real image system. Therefore, the ratio of dW ₂ = rR ₂ ′ and G ₂ ′ R ₂ ′ = W + dW _{2 in} the real image system is I (r) I (R ₂ ′) and I (G ₂ ′) I (R ₂ ′) in the bird's-eye view image. equal to the ratio _{of), I (R 2 ')} I (L 2') -I (r) I (l) (= 2I (r) I
(R ₂ ')) and I (L ₂ ') I (R ₂ ') (= 2I (G ₂ ') I (R ₂ '))
Equal to the ratio of I (L ₂ ') I (R ₂ ') = I (L ₁ ') I (R ₁ ') =
If W _i1 , I (r) I (l) = W _i2 , the equation (3-
2) can be replaced by equation (4) in the following bird's-eye image.

(Equation 4)

The meaning of equation (4) is clear. That is, the vehicle in the bird's-eye image of the lane mark that continues from the vehicle nearest to the vehicle to the rear far away from the bird's-eye image detected by the lane mark detection unit 31 of the image analysis unit 3 in FIG. The latest width W
The blur dP of the camera pitch angle Pc is obtained from _i1 , the width W _i2 at the _rear end of the bird's-eye image, and the vertical imaging angle P captured as the bird's-eye image in the real image system. Therefore, by replacing the camera pitch angle Pc of the imaging direction parameters with Pc + dP and performing viewpoint conversion again, a correct bird's-eye view image can be obtained without correcting the imaging direction of the camera itself.

[Second Embodiment] The height h of the camera C is h +
Let us consider the case where it changes to dh. As in FIG. 4, a set camera height h and a vertical imaging angle P captured as a bird's-eye view image are set as shown in FIG. The same applies to points G ₁ , G ₂ , R ₁ , R ₂ , length W, angles θ ₁ , θ ₂ . Next, as shown in FIG. 5 (b), when the height becomes a point C ′ of h + dh, a point R becomes an angle G ₁ C′R ₁ ′ of a size θ _1.
The distance between ₁ ′ and the point G ₁ ′ is W + dW ₁ , and the angle G ₂ C ′ of size θ ₂
The distance between the point R ₂ ′ that becomes R ₂ ′ and the point G ₂ ′ is set as W + dW ₂ . Then, the following relationship clearly holds.

(Equation 5)

From the equations (5-1) and (5-3), the following relationship is established.

(Equation 6)

By substituting equation (6) into equation (5-4), the following is obtained.

(Equation 7)

When the equations (5-2) and (7) are compared, the following equation is established.

(Equation 8)

From equation (8), it can be easily understood that even if the camera moves from the point C at the height h to the point C 'at the height h + dh, the parallelism of the lane marks is maintained in a bird's-eye view image. Therefore, the change dh in the camera height can be obtained by the following method.

In the bird's-eye image shown in FIG. 5 (c), it is assumed that the hatched lane mark disappears from the bird's-eye image after time t as shown in FIG. 5 (d). Here, if the image analysis means is linked with the speed measurement means of the vehicle, the distance D advanced during the time t is obtained. On the other hand, since the range of the image pickup range of the camera which is converted into a bird's-eye view image has a vertical angle of P, if the height is h + dh, the distance is (h + d
h) cosP. The distance D advanced during time t and (h +
dh) cosP and the variation dh of the camera position can be determined. In this manner, by repeatedly obtaining the bird's-eye image with the camera position set to h + dh, a correct bird's-eye image can be obtained.

[Third Embodiment] Determination of parameters in a captured image will be briefly described. For example, it is assumed that an image H of a horizontal line is detected in a captured image. At this time, FIG.
If the image H of the horizontal line is horizontal as shown in, there is no twist in the imaging direction of the camera. However, as shown in FIG. 6 (b), if the horizontal line imaging H forms an angle と with the horizontal, it can be detected that a twist of magnitude ψ has occurred in the imaging direction of the camera. Therefore, based on this, it is possible to correct the converted image (bird's eye image) without correcting the attitude of the camera itself.

[Fourth Embodiment] In many cases, it is difficult to detect a horizontal line image H as in the third embodiment. For example, this is due to terrain or a case where there are many obstacles such as buildings. At this time, if a gyro is mounted on the camera, it is possible to detect a twist in the imaging direction of the camera based on the information and correct the converted image (bird's-eye image) without correcting the attitude of the camera itself. it can. This makes it easy to correct a bird's-eye view image as in the first and second embodiments from an image in which twisting does not always occur.

[Fifth Embodiment] Correction in the case where the image pickup direction of the camera is shifted in the horizontal direction (lateral blur) will be described. When the camera shakes sideways, the captured image becomes as shown in FIG. 7A, and when converted into a bird's-eye image without correcting the imaging direction parameter, it becomes as shown in FIG. 7B. At this time, the pitch angle Pc and the torsion angle have been corrected by the first and third embodiments, and the dashed lane mark matches the traveling direction of the vehicle from the optical flow used in the second embodiment. When it is determined that the inclination of the lane mark is corrected, the lateral shake angle of the camera can be calculated. Therefore, it is possible to detect lateral blur in the imaging direction of the camera and correct the converted image (bird's-eye image) without correcting the attitude of the camera itself.

[Modification] In the first and second embodiments, the lane mark is analyzed. However, for example, a mark written on a road surface may be analyzed. When the sign includes a set of parallel lines parallel or perpendicular to the traveling direction of the vehicle or a rectangle, it is easy to use as an analysis target. The sign is "car"
(Especially the center (day), etc.), the part other than the arrow indicating the direction of travel, pedestrian crossings, stop lines, etc. are raised.

As described above, the present invention is effective as an imaging and image conversion device for obtaining an image such as a bird's-eye image with high accuracy. For example, a driver of a vehicle or a supervisor of a facility can be provided with a plan view that can easily grasp a sense of distance that is difficult to grasp with a normal captured image. As a result, the vehicle can be easily moved backward even if the driving ability is poor, such as a beginner or an elderly person.

The present invention not only provides a viewer with a converted image such as a bird's-eye view image, but also provides an image recognition device, so that a more sophisticated analysis of a captured surrounding situation can be performed. For example, by capturing an image of a vehicle or other moving object or the surroundings of a facility to which the imaging means of the present invention is fixed, and providing the converted image obtained by the present invention to the image recognition device, The moving object can be detected and discriminated, and a warning of collision avoidance and other instructions can be provided to the viewer. For example, by mounting on vehicles, pedestrians, obstacles,
The present invention can also be applied to a danger prevention device that warns a driver of approach to another vehicle, or an automatic vehicle traveling device that enables maintenance of the distance between vehicles and maintenance of lane center traveling.

[Brief description of the drawings]

FIG. 1 is a block diagram for explaining the contents of claims 1 and 2 of the present invention.

FIG. 2 is a block diagram showing a configuration of a first specific example of the present invention.

FIG. 3 is a flowchart showing the procedure of the first embodiment.

FIG. 4 is a positional relationship diagram for explaining the operation content of the first embodiment.

FIG. 5 is a positional relationship diagram for explaining a second embodiment.

FIG. 6 is a conceptual diagram showing a captured image for explaining a third embodiment.

FIG. 7 is a conceptual diagram showing a captured image and a bird's-eye image for explaining a fifth embodiment.

FIG. 8 is an explanatory diagram showing a relationship between a shift in pitch angle and a bird's-eye view when correction is not performed.

[Explanation of symbols]

 M Vehicle C Camera and camera position G Road surface Pc Camera pitch angle H Horizontal line image P Vertical camera image angle captured in bird's-eye image

──────────────────────────────────────────────────続 き Continued on the front page (51) Int.Cl. ⁷ Identification FI FI Theme Court ゛ (Reference) H04N 7/18 H04N 7/18 JK (72) Inventor Akihiro Watanabe Nagakute-cho, Aichi-gun, Aichi Prefecture 41 No. 1 Inside Toyota Central Research Institute Co., Ltd. (72) Inventor Takahashi Shin 41 Aichi County, Nagakute-cho, Aichi-gun Oku-cho, Yokomichi 41 No. 1 Inside Toyota Central Research Institute Co., Ltd. (72) Inventor Yoshiki Ninomiya Aichi-gun, Aichi Prefecture (1) Inside the Toyota Central R & D Laboratories Co., Ltd. (72) Inventor Mitsuhiko Ota 41-41 Nagakute-cho Okucho Yokomichi, Aichi-gun, Aichi Prefecture Inside Toyota Central R & D Laboratories Co., Ltd. (72) Inventor Hisashi Satonaka 1 Toyota Town, Toyota City, Aichi Prefecture Inside Toyota Motor Corporation (72) Inventor Makoto Nishida 1 Toyota Town 1 Toyota City, Toyota City, Aichi Prefecture Inside the Automobile Co., Ltd. (72) Inventor Toshiaki Kakinami 2-1-1 Asahi-cho, Kariya-shi, Aichi F-term (reference) in Aisin Seiki Co., Ltd. 5B057 AA16 BA02 CA08 CA12 CA16 CB08 CB12 CB16 CD12 DA07 DA20 DB02 DC03 DC08 5C023 AA10 AA38 BA02 CA02 DA08 5C054 AA01 CA04 CC02 CE12 EA05 ED11 FC03 FC12 FD03 HA30 5H180 AA01 AA25 AA26 CC04 CC24 CC30 5L096 BA04 CA02 EA07 FA66 FA67 HA02

Claims (7)

[Claims] 1. An image capturing and image converting apparatus comprising: an image capturing unit; an image converting unit configured to convert a captured image of the image capturing unit and output a converted image; and an image analyzing unit configured to analyze the captured image. An imaging and image conversion apparatus, wherein an imaging direction parameter of an imaging unit of the image conversion unit is corrected based on an analysis result of the analysis unit, and a corrected converted image is output. 2. An image pickup apparatus comprising: an image pickup unit; an image conversion unit that converts a captured image of the image pickup unit to output a converted image; and an image analysis unit that analyzes the converted image output by the image conversion unit. In the conversion device, the imaging direction is corrected based on the analysis result of the image analysis unit and the imaging direction parameter of the imaging unit of the image conversion unit, and the corrected converted image is output. Conversion device. 3. The imaging and image conversion apparatus according to claim 1, wherein the imaging unit is mounted on a vehicle and mainly captures an image of a road surface. 4. The image analysis means estimates and extracts a parallel line or a known figure on a road surface, and analyzes a captured image or a converted image of the parallel line or the known figure to determine a correction amount of the imaging direction parameter. The apparatus according to claim 3, wherein the determination is made. 5. The image analysis unit, in conjunction with a vehicle speed measurement unit, determines an amount of correction of the imaging direction parameter by analyzing a captured image or a converted image of an optical flow on a road surface. The imaging and image conversion device according to claim 3, wherein 6. The apparatus according to claim 1, wherein two or more imaging units are arranged at a predetermined distance, and the image analysis unit analyzes images captured by the plurality of imaging units or images converted by the image conversion unit. The imaging and image conversion apparatus according to claim 3, wherein a correction amount of the imaging direction parameter is determined. 7. An image pickup means further comprising: an attitude detection means for detecting a direction of the image pickup means; and a direction parameter of the image pickup means for the vehicle detected by the attitude detection means. The imaging and image conversion apparatus according to claim 3, wherein the correction amount is determined.