JP2019185535A - Travel path recognition device of vehicle and travel controller - Google Patents

Abstract

To extend a distance of compartment lines which can be calculated in a travel path recognition device which calculates right and left compartment lines of a travel path from a picked-up image around a vehicle to recognize a road shape.SOLUTION: A travel path recognition device 10 of a vehicle is provided with: an edge calculation part 12 constituted to calculate coordinates of an edge point at which a predetermined pixel parameter changes by a predetermined reference value or more on an image on the basis of the image picked up by an imaging part; and a compartment line recognition part 14 constituted to recognize right and left compartment lines of a travel path on the basis of a calculation result of the edge calculation part. The edge calculation part changes the reference value to be used for calculating the coordinates of the edge point so that a recognized distance becomes longer when the recognized distance of the compartment lines recognized by the compartment line recognition part is short.SELECTED DRAWING: Figure 2

Description

  The present disclosure relates to a vehicle travel path recognition apparatus and a travel control apparatus that are mounted on a vehicle, calculate left and right lane markings of the travel path of the vehicle from captured images around the vehicle, and recognize the road shape of the travel path.

  As this type of travel path recognition device, as described in Patent Document 1, an edge point whose luminance change is equal to or greater than a threshold value is extracted from a captured image around the vehicle, and linearly continuous edge points are connected. Thus, there is known an apparatus configured to calculate left and right lane markings of a vehicle traveling path.

  In Patent Document 1, when the number of lane markings calculated based on the extracted edge points exceeds a predetermined number, the threshold value used for edge point extraction is changed to make it difficult to perform edge point extraction. It has been proposed. That is, by changing the threshold value for edge point extraction, the left and right lane markings of the travel path can be calculated more accurately.

JP 2007-257449 A

  However, in the captured image, line paint such as a white line drawn on the road surface of the traveling road is difficult to identify farther away from the vehicle, and the luminance change occurring on both sides of the line paint becomes smaller as far away.

  In addition, pixel parameters such as luminance used for edge point extraction vary depending on the environment of the traveling road, such as weather, the presence / absence of a building around the traveling road, the entrance / exit of a tunnel, and the blur of line paint.

  For this reason, if the threshold value used for edge point extraction is uniformly changed in the entire captured image as in the above proposal, an unnecessary lane line obtained from the edge point can be removed, but finally obtained. The distance between the marked lines may be shortened.

  In addition, when the calculated lane marking distance is shortened, the road shape distance recognized from the lane marking is also shortened, so that the vehicle is automatically driven according to the road shape recognized by the travel path recognition device. In the travel control device, vehicle control cannot be performed normally.

  In other words, in the vehicle travel control device, when the road shape distance recognized by the travel path recognition device is shortened, a response delay occurs in steering during automatic travel, and the vehicle cannot travel stably. The problem arises. In this case, the vehicle control must be switched from automatic steering for lane keeping to manual steering by the driver, and the function of the travel control device is impaired.

  According to one aspect of the present disclosure, it is desirable to extend the distance of a computable lane line in a lane recognition device that recognizes a road shape by calculating the lane markings on the left and right sides of the lane from captured images around the vehicle.

  The traveling path recognition device of the present disclosure is mounted on a vehicle including an imaging unit (4) that captures an image around the vehicle and a sensor unit (6) that detects a traveling state of the vehicle, and the imaging unit and the sensor unit. It is an apparatus which recognizes the shape of the travel path of a vehicle using the information obtained from (1).

The traveling path recognition device according to one aspect of the present disclosure includes an edge calculation unit (12) and a lane marking recognition unit (14).
Then, the edge calculation unit calculates the coordinates of the edge point where the predetermined pixel parameter changes by a predetermined reference value or more on the captured image based on the image captured by the imaging unit, and the lane marking recognition unit Based on the calculation result of the calculation unit, the left and right lane markings of the traveling road are recognized.

Further, the edge calculation unit changes a reference value used for calculating the coordinates of the edge point so that the recognition distance becomes long when the recognition distance of the lane marking recognized by the lane marking recognition unit is short.
Therefore, according to the traveling path recognition device of one aspect of the present disclosure, the edge calculation unit changes the reference value of the predetermined pixel parameter used to detect the edge point on the captured image, thereby recognizing the lane marking. It can suppress that the recognition distance of the lane marking recognized in a part becomes short.

  Therefore, according to the road recognition device of one aspect of the present disclosure, the distance of the road shape obtained from the left and right lane lines of the road recognized by the lane line recognition unit is shortened, and the vehicle is moved according to the road shape. It is possible to suppress the occurrence of abnormality in the control in the travel control device that automatically travels.

Next, the traveling path recognition apparatus of the other aspect of this indication is provided with the edge calculation part (12) and the lane marking recognition part (14) similarly to the traveling path recognition apparatus of the said one situation.
Then, as described above, the edge calculation unit calculates the coordinates of the edge point at which a predetermined pixel parameter changes by a threshold value or more on the captured image based on the image captured by the imaging unit.

On the other hand, the lane marking recognition unit includes a paint extraction unit (20), a line candidate calculation unit (22), and a trace edge unit (26).
Here, the paint extraction unit is configured to extract line paint drawn on the road surface as a lane marking of the traveling road based on the calculation result of the edge calculation unit.

  The line candidate calculation unit is configured to calculate a line candidate that is a candidate for a lane line based on the line paint extracted by the paint extraction unit and the running state of the vehicle detected by the sensor unit. .

  Further, the trace edge portion is configured to search for an edge point existing in the extending direction of the line candidate calculated by the line candidate calculation unit based on the coordinates of the edge point calculated by the edge calculation unit.

Then, the line candidate calculation unit registers an edge point existing in the extending direction of the line candidate as an edge point of the line candidate based on the search result of the edge point by the trace edge part.
That is, in the travel path recognition device of the present disclosure, there is an edge point searched for in the trace edge portion in the extending direction of the line candidate calculated based on the line paint and the running state of the vehicle by the line candidate calculation unit. In this case, the edge point is registered as the edge point of the line candidate.

As a result, according to the traveling path recognition device of the present disclosure, it is possible to increase the distance between the line candidates calculated by the line candidate calculation unit and to suppress the distance between the lane lines recognized by the lane line recognition unit from being shortened. .
Therefore, also in the travel path recognition device of the present disclosure, the distance of the road shape obtained from the lane line recognized by the lane line recognition unit is shortened according to the road shape, as in the travel path recognition device of the above aspect. It is possible to suppress the occurrence of abnormality in the control in the travel control device that automatically travels the vehicle.

  Next, the vehicle travel control device according to one aspect of the present disclosure includes an imaging unit (4) configured to capture an image around the vehicle, and a sensor unit (6) configured to detect a traveling state of the vehicle. And a travel path recognition unit (10) and a control unit (8).

  The control unit is configured to control the traveling state of the vehicle based on information such as the shape of the traveling path obtained from the traveling path recognition unit, and the traveling path recognition unit is one aspect or the other of the present disclosure described above. It is comprised with the travel path recognition apparatus of the situation of.

  Therefore, according to the vehicle travel control device of one aspect of the present disclosure, it is possible to suppress the distance of the travel path recognized by the travel path recognition unit from being shortened, and to detect the travel path recognized by the travel path recognition unit. According to the shape, the running state of the vehicle can be stably controlled without a response delay.

  In addition, the code | symbol in the parenthesis described in this column and a claim shows the correspondence with the specific means as described in embodiment mentioned later as one aspect, Comprising: The technical scope of this invention is shown. It is not limited.

It is a block diagram showing the structure of the whole traveling control apparatus of the vehicle of embodiment. It is a block diagram showing the structure of the traveling path recognition part shown in FIG. It is explanatory drawing explaining the edge detection operation | movement in an edge calculation part. It is a flowchart showing the line candidate calculation process performed as a line candidate calculation part. It is a flowchart showing the exclusion determination process performed in S150 of FIG. It is a flowchart showing the reconnection determination process performed in S160 of FIG. It is explanatory drawing explaining the connection operation | movement of the line candidate in S110 and S140 of FIG. It is explanatory drawing explaining the exclusion operation | movement of the line candidate by the exclusion determination process shown in FIG. It is explanatory drawing which represents the recognition result of the lane marking by the traveling path recognition part of embodiment compared with the conventional apparatus. It is explanatory drawing explaining the reference value update operation | movement in an edge calculation part. It is a flowchart showing the reference value update process performed in order to implement | achieve the reference value update operation | movement shown in FIG.

Embodiments of the present invention will be described below with reference to the drawings.
As shown in FIG. 1, the vehicle travel control apparatus of the present embodiment includes an in-vehicle camera 4, a sensor unit 6, a travel path recognition unit 10, and a control unit 8.

The in-vehicle camera 4 corresponds to the imaging unit of the present disclosure, and is installed in the vehicle 2 so as to capture an image ahead in the traveling direction.
The sensor unit 6 is for detecting the traveling state of the vehicle 2, and in order to be able to predict the behavior of the vehicle 2, at least a vehicle speed sensor for detecting the vehicle speed, a yaw rate sensor for detecting the yaw rate of the vehicle 2, and .

  The traveling path recognition unit 10 corresponds to the traveling path recognition device of the present disclosure, and the vehicle 2 is traveling based on the image captured by the in-vehicle camera 4 and the traveling state of the vehicle detected by the sensor unit 6. Recognize the left and right lane markings on the road (specifically, the driving lane).

  That is, the travel path recognition unit 10 is configured by a microcomputer or an ASIC that can process an image captured by the in-vehicle camera 4, recognizes the left and right lane markings of the travel path (lane) from the captured image, and controls the control unit. 8 recognizes the road shape necessary for automatically driving the vehicle 2.

  The control unit 8 is mounted on the vehicle 2 by calculating, for example, a steering amount required to cause the vehicle to automatically travel along the travel path based on the road shape recognized by the travel path recognition unit 10. So-called lane keep control is performed to control the steering amount of the steering device.

  The control unit 8 is a so-called ECU, which is an electronic control device for vehicle control configured by a microcomputer. Further, the control unit 8 controls the vehicle steering amount, the driving force, and the braking force based on the road shape recognized by the traveling path recognition unit 10, thereby automatically traveling the vehicle 2 for automatic traveling control. ECU may be used. Since these lane keeping control and automatic traveling control are well known, detailed description on the control unit 8 is omitted here.

Next, the traveling path recognition unit 10 that is a main part of the present disclosure will be described.
As shown in FIG. 2, the travel path recognition unit 10 has functions as an edge calculation unit 12, a lane marking recognition unit 14, and a road shape estimation unit 16. Further, the lane marking recognition unit 14 has functions as a paint extraction unit 20, a line candidate calculation unit 22, a line candidate selection unit 24, and a trace edge unit 26.

  In the present embodiment, the traveling path recognition unit 10 is configured by a microcomputer including a CPU, a ROM, a RAM, and the like, and the function of each of the above units is executed by the CPU executing a control program stored in the ROM. , Shall be realized.

  Here, the edge calculation unit 12 detects, as an edge point, a pixel whose luminance as a pixel parameter changes more than a preset reference value on the captured image based on the captured image captured by the in-vehicle camera 4. It is a functional block for calculating the coordinates.

  Specifically, as shown in FIG. 3A, the edge calculation unit 12 scans the captured image in the X direction, which is the left-right direction toward the traveling direction, so that the luminance changes in the increasing direction and the decreasing direction over the reference value. The searched point is searched as an edge point, and the coordinates of the searched edge point are obtained.

  Further, scanning for edge point search is sequentially performed by changing the scanning position in the Y direction orthogonal to the X direction, and the edge point is searched for each scanning position. As a result, for example, when there are line paints such as white lines drawn as left and right dividing lines on the road surface at the scanning positions Y1 and Y2, as shown in FIGS. 3B and 3C, Edge points will be detected on both sides.

  On the other hand, the line paint on the captured image changes depending on the environment of the road, such as weather, the presence or absence of buildings around the road, the entrance of the tunnel, and the blur of the line paint. There is. In particular, the line paint tends to become unclear because the width decreases as the distance from the vehicle 2 increases.

  In this way, when the line paint becomes unclear on the captured image, the line paint is detected as an edge point of a discontinuous line, and the edge point is a white edge point having a high luminance or a black point having a low luminance. Are recognized as edge points.

  Next, the lane marking recognition unit 14 is a functional block that recognizes the lane markings on the left and right of the lane that will be the travel road based on the calculation result of the edge calculation unit 12, and the road shape estimation unit 16 is the lane marking recognition unit 14. This is a functional block that estimates the road shape of the traveling road based on the left and right lane markings recognized in step S5 and a preset road shape model.

  Further, in the lane marking recognition unit 14, the paint extraction unit 20 is a functional block that extracts line paint drawn on the road surface as the lane markings of the traveling road based on the coordinates of the edge points calculated by the edge calculation unit 12. is there.

  That is, the paint extraction unit 20 extracts edge points that are linearly continuous from the coordinates of the edge points calculated by the edge calculation unit 12. Then, from the lines connecting the extracted edge points, two lines whose line intervals change constant or continuously within a tolerance set in advance for extracting a line paint such as a white line, etc. The left and right boundary lines of the line paint are specified, and the line paint is extracted from the specified left and right boundary lines.

Note that this line paint extraction method is also described in Patent Document 1 and is well known, and therefore detailed description thereof is omitted here.
Next, the line candidate calculation unit 22 is a functional block that calculates line candidates that are candidates for left and right division lines based on the line paint extracted by the paint extraction unit 20.

The function as the line candidate calculation unit 22 is realized by the CPU constituting the travel path recognition unit 10 executing the line candidate calculation process shown in FIG.
That is, as shown in FIG. 4, in the line candidate calculation process, in S110, a predicted line is obtained from the already calculated line candidates, and the predicted line is extracted from the line paint extracted by the paint extracting unit 20. Select a nearby line paint and connect it to the line candidate.

  Note that the predicted line is a predicted route where the vehicle is predicted to travel from the tip of the already calculated line candidate, and the shape of the already calculated line candidate and the current of the vehicle 2 detected by the sensor unit 6 are detected. Based on the running state, it is estimated as shown in FIG. 7A. Note that the vehicle speed and the yaw rate detected by the sensor unit 6 are used for estimating the prediction line.

  In S110, it is determined whether there is a line paint located within a predetermined range LA (for example, 0.4 m) from the predicted line, and there is a line paint located within the predetermined range LA from the predicted line. For example, the line paint is connected to the line candidate. In S110, if there is no line paint located within the predetermined range LA from the predicted line, connection of line candidates based on the predicted line is not performed.

  In S110, when connecting the line paint to the line candidate, one of the left and right boundary lines of the line paint, for example, the inner boundary line of the traveling road, is used as a line candidate. Connect to line candidates.

  Thus, when the connection process to the line candidate of the line paint based on the predicted line is performed in S110, the process proceeds to S120, and there is an edge point extending linearly from the last line paint connected as the line candidate. Judge whether or not to do. In the following description, an edge point extending linearly from the final line paint is referred to as a trace edge.

  The trace edge is obtained by the trace edge unit 26 based on the final line paint connected as a line candidate by the line candidate calculation unit 22 and the edge point on the captured image detected by the edge calculation unit 12. It is done.

  In other words, the trace edge unit 26 searches the edge points detected by the edge calculation unit 12 for edge points that are linearly continuous from the final line paint of the line candidates, whereby the trace edge shown in FIG. 7A is obtained. Is generated.

  Note that the trace edge unit 26 has high luminance (for example, white) edge points determined to have changed in the increasing direction by the edge calculating unit 12 or low luminance in which the luminance has been determined to have decreased in the decreasing direction. Each edge point (for example, black) is searched for. The trace edge unit 26 searches for edge points whose luminance or luminance change is within a certain range.

  For this reason, on the captured image, the trace edge looks like a white line, a black line, or the like, and the color appears to change gradually, such as from gray to black, or from white to gray. Will be generated by the line. The trace edge is generated if the edge points are arranged in a line even if the line is discontinuous on the captured image.

  In S120, when the trace edge is generated in the trace edge unit 26, it is determined that the trace edge exists, and the process proceeds to S130. If the trace edge is not generated, the process proceeds to S150.

  Next, in S130, it is determined whether or not there is a line paint that is located near the trace edge and can be connected via the trace edge in the line paint extracted by the paint extraction unit 20. This determination is performed by determining whether or not there is a line paint located within a predetermined range LB (for example, 0.2 m) from the trace edge.

  If it is determined in S130 that line paint exists in the vicinity of the trace edge, the process proceeds to S140, and as illustrated in FIG. 7B, line candidates are obtained from the trace edge to the line paint in the vicinity of the trace edge. Extend and connect line candidates to the line paint.

  In S140, when the line paint determined to exist near the trace edge does not exist within the predetermined range LC (for example, 0.96 m) larger than the LA from the predicted line, the line paint greatly deviates from the predicted line. Therefore, connection of line candidates is not performed.

If the process of S140 is performed or if it is determined in S130 that no line paint exists near the trace edge, the process proceeds to S150.
Next, in S150, an exclusion determination process is performed as a line candidate exclusion unit of the present disclosure.

  As shown in FIG. 5, in the exclusion determination process, first, in S210, one of the line paints constituting the line candidate is one before the final first line paint P1 farthest from the vehicle. The second line paint P2 is selected. Then, a fitting coefficient corresponding to the shape of the selected second line paint P2 is calculated.

  Next, in S220, as shown in FIG. 8A, based on the fitting coefficient calculated in S210, a stretched line stretched from the second line paint P2 along the shape of the second line paint P2 is obtained. And the first line paint P1 are calculated.

  In subsequent S230, it is determined whether or not the positional deviation amount calculated in S220 is equal to or larger than a preset threshold value. If the positional deviation amount is equal to or larger than the threshold value, the process proceeds to S240, and FIG. As shown in FIG. 5, the final first line paint P1 is excluded from the line candidates. As a result, the line candidates are limited to the second line paint P2.

  Further, if the final first line paint P1 is excluded from the line candidates in S240, or if it is determined in S230 that the positional deviation amount is less than the threshold value, the exclusion determination process ends. The process proceeds to S160.

In the present embodiment, the reason why the line candidate calculation unit 22 performs the exclusion determination process is to generate a trace edge in the trace edge unit 26 and extend the line candidate.
In other words, if the line candidate is extended based on the trace edge generated by the trace edge unit 26, it is considered that line paint that is not a traveling path is erroneously connected as a line candidate, so the exclusion determination process is performed. Therefore, it is possible to suppress the extension of the line candidate by mistake.

Therefore, by executing the exclusion determination process as the exclusion determination unit, it is possible to improve the recognition accuracy of the left and right lane markings of the lane that is the travel path of the vehicle 2.
Next, in S160, a reconnection determination process as a reconnection determination unit of the present disclosure is executed.

  As shown in FIG. 6, in the reconnection determination process, first, in S310, the road shape of the travel path along the line candidate is obtained based on the position and shape of the line paint connected as the line candidate, Estimate the road shape of the road that further extends from the candidates.

  In subsequent S320, the positional deviation amount between the road shape estimated in S310 and the unconnected line paint that is not connected as a line candidate among the line paints extracted by the paint extracting unit 20 is calculated.

  In S330, it is determined whether the displacement amount calculated in S320 is less than a preset threshold value, so that the unconnected line paint in which the estimated displacement amount from the road shape is less than the threshold value. It is determined whether or not exists.

  If it is determined in S330 that there is an unconnected line paint whose positional deviation from the estimated road shape is less than the threshold, the process proceeds to S340, and the unconnected line paint is connected to the line candidate. , Extend line candidates.

  If it is determined in S340 that the unconnected line paint is connected to the line candidate, or it is determined in S330 that there is no unconnected line paint whose positional deviation from the estimated road shape is less than the threshold value. The reconnection determination process is terminated.

  As a result, if there is unconnected line paint whose positional deviation amount is less than the threshold for the road shape estimated in S310, even if it is excluded from the line candidates in the exclusion determination process, The unconnected line paint is connected to the line candidate, and the line candidate is extended.

  As described above, the line candidate calculation unit 22 travels by connecting the line paint extracted by the paint extraction unit 20 in order along the predicted line set in consideration of the travel state of the vehicle 2. Line candidates for left and right lane markings on the road (lane) are generated.

  When line paint cannot be connected to a line candidate, a trace edge is generated by searching for an edge portion arranged linearly from the last line candidate, and line paint exists in the vicinity of the trace edge The line paint is connected as a line candidate. Therefore, according to the present embodiment, the distance between the line candidates can be increased as compared with the conventional apparatus that does not generate the trace edge.

  In addition, since the trace edge is generated by searching for an edge portion of the same system in which the luminance or luminance change that is the pixel parameter is within a predetermined range, even a white line on the captured image is a black line. However, it is generated as a trace edge.

  If there is a line paint near the trace edge, the line paint is connected as a line candidate. Therefore, even if the line paint is unclear at a position away from the vehicle 2, the line paint is a blurred line paint. Can be connected to extend the distance of the line candidate.

  Next, when there are a plurality of line candidates calculated by the line candidate calculation unit 22 on the left side or the right side of the travel path, the line candidate selection unit 24 illustrated in FIG. This is a functional block for selecting line candidates to be left and right partition lines.

  When there are a plurality of line candidates, for example, the line candidate selection unit 24 selects a clear line candidate having high luminance on the captured image obtained from the in-vehicle camera 4 or the progress of the vehicle 2 obtained from the sensor unit 6. A line candidate that becomes a partition line is selected by selecting a line candidate along the direction.

  The line candidates calculated by the line candidate calculation unit 22 and selected as the lane markings by the line candidate selection unit 24 are estimated as road shapes from the lane marking recognition unit 14 as the lane markings on the left and right of the traveling road. Is output to the unit 16.

  Therefore, as shown in FIG. 9, the distance from the lane marking recognition unit 14 to the final ends Pel and Per of the lane markings output to the road shape estimation unit 16 is longer than that of the conventional device. Then, the road shape estimation unit 16 estimates the road shape of the traveling road based on the left and right lane markings input from the line candidate selection unit 24. For this reason, the length of the road shape estimated by the road shape estimation unit 16 is also longer than that of the conventional device.

In FIG. 9, FIG. 9A represents a lane line recognized by the conventional apparatus, and FIG. 9B represents a lane line recognized by the lane line recognition unit 14 of the present embodiment.
As is clear from this figure, the right lane marking drawn with a continuous white line among the left and right lane markings drawn on the road is recognized as one line paint, so the conventional apparatus and this implementation In the form, the recognition distance is substantially the same.

  On the other hand, the left lane line drawn with a short white line at a predetermined interval is difficult to connect as a line candidate because the line paint becomes unclear as it moves away from the vehicle on the captured image. Becomes shorter.

However, in this embodiment, the line paint is connected using the trace edge, so that the recognition distance is substantially the same as the right partition line.
Therefore, when the control unit 8 executes the lane keeping control based on the road shape output from the travel path recognition unit 10, the distance of the road shape output from the travel path recognition unit 10 is shortened. It is possible to suppress the occurrence of a steering delay during traveling and the vehicle 2 from becoming unstable.

  By the way, in this embodiment, when the line candidate calculation part 22 produces | generates the line candidate used as the division | segmentation line candidate of a driving path from line paints, such as a white line extracted from the captured image in the paint extraction part 20. FIG. In addition, the distance of the line candidate is extended by using the trace edge.

  However, since the line candidate is generated based on the edge part on the captured image extracted by the edge calculating unit 12, and the trace edge is also generated based on the edge part, the edge calculating unit 12 determines the distance between the line candidates. In order to extend the length, it is preferable that the edge portion can be accurately detected.

  That is, in the captured image, in a distant area located far from the vehicle 2, line paint such as a white line is likely to be unclear, and the weather, the presence or absence of a building around the traveling road, the entrance of the tunnel, the blur of the line paint, etc. The line paint also changes depending on the environment of the road.

  For example, as shown in FIG. 10A, when there is a building such as a bridge that generates a shadow on a distant road, a part of the captured image becomes dark due to the shadow of the building, and in the dark region, The brightness, which is a pixel parameter, decreases.

  Then, as in the dark area shown in FIG. 10A, when the luminance change becomes small in the far region and the edge of the white line portion cannot be detected from the luminance change, the distance between the line candidates that can be recognized by the lane marking recognition unit 14 is It will be shorter.

  Therefore, in the present embodiment, as one function of the edge calculation unit 12, the reference value update process shown in FIG. 11 is executed, so that the captured image is obtained when the distance between the lane markings output from the lane marking recognition unit 14 is short. The reference value used for detecting the edge portion from the luminance change is corrected.

  That is, in the reference value update processing, first, in S410, information on the left and right lane lines output from the lane line recognition unit 14 to the road shape estimation unit 16 is acquired, and the recognition distance between the left and right lane lines is calculated.

  Next, in S420, it is determined whether at least one of the recognition distances of the left and right lane markings is equal to or less than a preset threshold value. If the recognition distance is equal to or less than the threshold value, the process proceeds to S430 and edge extraction is performed. The reference value is changed to a value at which the edge point can be easily detected.

  Then, in S430, the reference value is changed, or when it is determined in S420 that the recognition distances of the left and right lane markings are both larger than the reference value, the reference value update process is ended.

  The updated reference value is held until the operation of the travel control device is stopped. For example, when the ignition switch of the vehicle 2 is turned on and the travel control device is started, the reference value is set to the initial value. .

  As a result, in the edge calculation unit 12, as shown in FIG. 10B, when the recognition distance of the lane marking is short, a threshold value used for detecting a pixel having a luminance change on the captured image as an edge point is The reference threshold value which is the initial value is corrected to a value smaller than the reference threshold value.

  Therefore, the edge calculation unit 12 of the present embodiment adjusts the detection sensitivity when detecting the edge point from the luminance change of the captured image when the lane marking recognition distance is shortened due to the surrounding environment of the vehicle 2. Thus, the recognition distance of the lane marking can be increased.

  Therefore, according to the present embodiment, even when the reference value update operation in the edge calculation unit 12 is performed, the lane line recognition distance is increased, and the control unit 8 executes the lane keeping control. The distance of the road shape to be output is shortened, and the occurrence of a steering delay can be suppressed.

As mentioned above, although the form for implementing this invention was demonstrated, this invention is not limited to the above-mentioned embodiment, It can implement in various deformation | transformation.
For example, in the above embodiment, as shown in FIG. 2, the trace edge unit 26 searches the line candidate calculation unit 22 for an edge point extending linearly from the last line paint connected as a line candidate as a trace edge. Explained as what to do.

  However, since the line candidate calculation unit 22 also generates line candidates that are not selected as lane markings by the line candidate selection unit 24, when a plurality of line candidates are generated by the line candidate calculation unit 22, the trace edge The unit 26 generates a trace edge for each line candidate.

  Therefore, as indicated by a dotted arrow in FIG. 2, the left and right line candidates selected as the division lines by the line candidate selection unit 24 are input to the trace edge unit 26, and the edge points extending linearly from each division line Only a trace edge may be searched for.

If it does in this way, in runway recognition part 10, it will become possible to reduce processing load performed by CPU in order to realize a function as trace edge part 26.
In the above-described embodiment, when the edge calculation unit 12 extracts an edge from a captured image, the luminance is used as a pixel parameter, and the edge is detected from the luminance change. However, the pixel parameter used for extracting the edge portion may be a pixel value that changes due to line paint, such as a white line drawn on the road surface, and it is not always necessary to use luminance.

  Moreover, in the said embodiment, the traveling path recognition part 10 as a traveling path recognition apparatus is comprised with the microcomputer, and the function shown in FIG. 2 is that CPU performs the control program memorize | stored in ROM, It was described as being realized.

  However, for example, some functions that increase the processing load on the CPU, such as the trace edge section 26, may be configured by a dedicated circuit such as an ASIC, or all functions may be configured by a dedicated circuit such as an ASIC It may be configured.

  In addition, a plurality of functions that one constituent element in the above-described embodiment has may be realized by a plurality of constituent elements, or a single function that one constituent element has may be realized by a plurality of constituent elements. Further, a plurality of functions possessed by a plurality of constituent elements may be realized by one constituent element, or one function realized by a plurality of constituent elements may be realized by one constituent element. Moreover, you may abbreviate | omit a part of structure of the said embodiment. In addition, at least a part of the configuration of the above embodiment may be added to or replaced with the configuration of the other embodiment. In addition, all the aspects included in the technical idea specified only by the wording described in the claim are embodiment of this invention.

  DESCRIPTION OF SYMBOLS 2 ... Vehicle, 4 ... Car-mounted camera, 6 ... Sensor part, 8 ... Control part, 10 ... Running road recognition part, 12 ... Edge calculation part, 14 ... Track line recognition part, 16 ... Road shape estimation part, 20 ... Paint extraction Part, 22 ... line candidate calculation part, 24 ... line candidate selection part, 26 ... trace edge part.

Claims (12)

It is mounted on a vehicle (2) that includes an imaging unit (4) that captures an image around the vehicle and a sensor unit (6) that detects a traveling state of the vehicle. A vehicle travel path recognition device (10) for recognizing the shape of a travel path of a vehicle using information obtained from a sensor unit,
An edge calculation unit (12) configured to calculate coordinates of an edge point on the image where a predetermined pixel parameter changes by a predetermined reference value or more based on an image captured by the imaging unit;
A lane marking recognition unit (14) configured to recognize the lane markings on the left and right of the travel path based on the calculation result of the edge calculation unit;
The edge calculation unit is used to calculate the coordinates of the edge point so that the recognition distance becomes long when the recognition distance of the lane marking recognized by the lane marking recognition unit is short. A travel path recognition apparatus for a vehicle configured to change a reference value. It is mounted on a vehicle (2) that includes an imaging unit (4) that captures an image around the vehicle and a sensor unit (6) that detects a traveling state of the vehicle. A vehicle travel path recognition device (10) for recognizing the shape of a travel path of a vehicle using information obtained from a sensor unit,
An edge calculation unit (12) configured to calculate the coordinates of an edge point on the image where a predetermined pixel parameter changes by a threshold value or more based on the image captured by the imaging unit;
A lane marking recognition unit (14) configured to recognize the lane markings on the left and right of the travel path based on the calculation result of the edge calculation unit;
With
The lane marking recognition unit
A paint extraction unit (20) configured to extract line paint drawn on a road surface as a lane marking of the travel path based on a calculation result of the edge calculation unit;
Based on the line paint extracted by the paint extraction unit and the running state detected by the sensor unit, a line candidate calculation unit (22) configured to calculate line candidates that are candidates for the lane markings (22) )When,
Based on the coordinates of the edge point calculated by the edge calculation unit, a trace edge unit (26) configured to search for an edge point existing in the extending direction of the line candidate calculated by the line candidate calculation unit (26) )When,
And the line candidate calculation unit is configured to register the edge point existing in the extending direction of the line candidate as an edge point of the line candidate based on a search result of the edge point by the trace edge unit. A vehicle travel path recognition device.   The edge calculation unit uses the reference value used to calculate the coordinates of the edge point so that the recognition distance becomes long when the recognition distance of the lane marking recognized by the lane marking recognition unit is short. The vehicle travel path recognition apparatus according to claim 2, wherein the travel path recognition apparatus is configured to be changed. The line candidate calculation unit
Based on the search result of the edge point by the trace edge part and the line paint extracted by the paint extraction part, whether or not there is line paint connectable at the edge point searched by the trace edge part When the line paint which can be connected exists, it is comprised so that the said edge point searched in the said trace edge part may be registered as an edge point of the said line candidate. 4. A travel path recognition device for a vehicle according to 3. The trace edge portion is
The edge calculation unit is configured to search each of the edge point determined to have changed in the increasing direction and the edge point determined to have changed in the decreasing direction as one line. The travel path recognition device for a vehicle according to any one of claims 2 to 4, wherein The trace edge portion is
The edge point calculated by the edge calculation unit is configured to search for an edge point in which the pixel parameter or a change in the pixel parameter is within a certain range as one line. The vehicle travel path recognition apparatus according to claim 5. The line candidate calculation unit
Among the line paints used to calculate the line candidates, the line paint is extended along the shape of the second line paint from the second line paint immediately before the first line paint farthest from the vehicle. A line candidate exclusion unit configured to exclude the first line paint from the line candidates when the amount of displacement of the first line paint with respect to the stretched line is equal to or greater than a predetermined value. 22, S150),
The travel path recognition device for a vehicle according to any one of claims 2 to 6, further comprising: The line candidate calculation unit
The road shape in the vehicle traveling direction is estimated from a plurality of line paints constituting the line candidate, and the estimated road shape and the line paint extracted by the paint extraction unit are not connected as the line candidates. A reconnection determination unit (22, S160) configured to obtain a positional deviation amount with respect to the unconnected line paint and connect an unconnected line paint having the positional deviation amount less than a predetermined value as the line candidate;
The travel path recognition device for a vehicle according to any one of claims 2 to 7, further comprising: In the lane marking recognition unit, when a plurality of line candidates are calculated as the lane marking candidates in the line candidate calculation unit, the plurality of line candidates based on information from the imaging unit and the sensor unit A line candidate selection unit (24) configured to select a line candidate to be the division line from
A road shape estimation unit (16) configured to estimate the road shape of the traveling road based on the line candidates of the lane markings calculated by the line candidate calculation unit and selected by the line candidate selection unit When,
The travel path recognition device for a vehicle according to any one of claims 2 to 8, comprising:   The travel path of a vehicle according to claim 9, wherein the trace edge portion is configured to search for an edge point existing in an extending direction of the line candidate selected as the lane marking by the line candidate selection unit. Recognition device. An imaging unit (4) configured to capture an image around the vehicle;
A sensor unit (6) configured to detect a running state of the vehicle;
A travel path recognition unit (10) configured to recognize the shape of the travel path of the vehicle using information obtained from the imaging unit and the sensor unit;
A control unit (8) configured to control a traveling state of the vehicle based on information obtained from the traveling path recognition unit;
The travel path recognition unit is configured by the travel path recognition device according to any one of claims 1 to 10.   The vehicle travel control device according to claim 11, wherein the control unit is configured to control a steering amount of the vehicle based on a shape of the travel path recognized by the travel path recognition unit.

Images