JP5918689B2 - Parking assistance device - Google Patents

Description

  The present invention relates to a parking support device, and more particularly to a parking support device that detects a detection target such as a parked vehicle or a parking frame line.

  2. Description of the Related Art Conventionally, parking assistance devices that detect a parked vehicle or a parking frame line using a distance measuring sensor or a camera and set a parking space or a target parking position based on the detection result are known (for example, Patent Documents 1 and 2). reference). In the invention of Patent Document 1, the distance to the parked vehicle is sequentially detected by a distance measuring sensor such as an ultrasonic sensor when the parking space (parked vehicle) passes by the side, and the parked vehicle (detection target) is based on the distance detection result. The contour point is detected. In the invention of Patent Document 2, a camera that captures the rear of the vehicle is mounted on the vehicle, an edge is extracted from an image captured by the camera, and a parking frame line (detection target) is detected based on the extracted edge. ing.

JP 2008-21039 A JP 2012-80497 A

  By the way, as shown in FIG. 15, there is a scene where there is a step 200 in front of the parked vehicle 6 parked in parallel in Europe or the like. In this scene, when the distance is detected by the distance measuring sensor 2 mounted on the vehicle 5, the step is detected. 200 may be detected. In FIG. 15, the detection distance point 73 of the step 200 is indicated by a black circle. In particular, when the specification of the distance measuring sensor 2 is a specification in which only the first reflected wave received after transmission of an exploration wave such as an ultrasonic wave is detected, the parked vehicle 6 behind the step 200 is detected. As a result, the parking space 100 adjacent to the parked vehicle 6 cannot be detected.

  Moreover, in a parking assistance device that detects a three-dimensional object such as a parked vehicle using a distance measuring sensor, the distance of the road surface may be erroneously detected due to the transmission range of the exploration wave reaching the road surface. In addition, in a parking assistance device that detects a parking frame line using a camera, it is difficult to determine whether an edge extracted from an image is an edge on a road surface or an edge of a three-dimensional object. The parking frame line may be erroneously detected in the range of the three-dimensional object.

  Thus, in the conventional parking assistance device, there is a problem that detection of the detection target or erroneous detection may occur.

  This invention is made | formed in view of the said problem, and makes it a subject to provide the parking assistance apparatus which can prevent generation | occurrence | production of the detection failure | detection of a detection target, or a misdetection.

In order to solve the above-described problem, the parking assist device of the present invention sequentially transmits exploration waves to the periphery of the vehicle, receives the reflected wave reflected by the exploration wave and hits the object, and based on the reflected wave, the distance to the object Distance detecting means for sequentially detecting,
Imaging means for imaging the periphery of the vehicle;
Extraction means for extracting feature points in the image captured by the imaging means;
Object detection means for detecting a detection target in or around the parking space using both the distance detection result by the distance detection means and the extraction result of the feature points;
Assuming that

  According to the present invention, both a distance detection unit (ranging sensor) that detects a distance to an object and an imaging unit (camera) that images the periphery of the vehicle are mounted on the vehicle. When the distance detection means is used, it is easy to detect a three-dimensional object having a certain height or higher, but it is difficult to determine what the detected three-dimensional object is (a step or a parked vehicle). On the other hand, when the imaging means is used, it is easy to determine the shape of the object and the road surface range shown in the image, while the object having the feature points extracted from the image is an object on the road surface (parking frame line). It is difficult to determine whether it is an object (such as a parked vehicle). In the present invention, since the detection target is detected using both the distance detection result by the distance detection means and the extraction result of the feature point, the disadvantage of the detection using the distance detection means is compensated by the advantage of the imaging means, or the imaging means The shortcomings of detection using can be compensated by the advantages of the distance detection means. Therefore, it is possible to prevent the detection target from being undetectable and erroneous detection.

As a first specific aspect of the present invention, the distance detection means transmits the exploration wave to the side of the vehicle when passing sideways in a parking space,
The imaging means captures a range including at least the side of the vehicle when passing laterally in a parking space;
The extraction means extracts an edge in the image as the feature point,
The object detection means includes
When the point sequence of the detection distance detected by the distance detection unit is linear, and the linear edge is extracted in the vicinity of the linear point sequence, the detection range of the distance detection unit is defined as a straight line. Detection range changing means for changing the detection distance to be farther than the linear detection distance, which is the detection distance to be a point sequence;
And a three-dimensional object detection unit that detects a three-dimensional object adjacent to a parking space as the detection target based on the detection distance detected after the detection range is changed by the detection range changing unit.

  When there is a step between a vehicle and a three-dimensional object adjacent to the parking space in a scene when passing through the side of the parking space, the point sequence of the detection distance detected by the distance detection means is linear. In this case, a linear edge is included in the vicinity of the point sequence of the linear detection distance in the image captured by the imaging means. According to the present invention, when a point sequence of a linear detection distance is detected and a linear edge is extracted in the vicinity of the point sequence, there is a step between the vehicle and the three-dimensional object. The detection range of the distance detection means is changed farther than the step detection distance (linear detection distance). For this reason, after the detection range is changed, it is possible to prevent detection of an object (step) in the range from the vehicle to the linear detection distance, and it is adjacent to the detection target that is far from the linear detection distance, that is, the parking space. It is easy to detect a three-dimensional object such as a parked vehicle. In particular, among the reflected waves that the distance detection means receives in one transmission of the exploration wave, only the first received wave having an amplitude exceeding a predetermined threshold is valid, and the object from the reflected wave to the object is valid. Even if it is the specification which detects distance, the detection of the solid object in the back of a level | step difference is attained.

As a second specific aspect of the present invention, the extraction means in the present invention extracts an edge in the image as the feature point,
The target detection unit is configured to detect a low level difference as the detection target when the linear edge is extracted and a point of the detection distance detected by the distance detection unit is not obtained in the vicinity of the linear edge. A low step determining means for determining that there is a line marked on the road surface is provided.

  When there is a low step or line on the road surface, an edge of the low step or line (straight edge) is included in the image. On the other hand, in the detection by the distance detection means, it is difficult to receive a reflected wave from an object that does not reach a certain height, that is, a low step or a line, and therefore a detection distance point cannot be obtained in the vicinity of a straight edge. According to the present invention, when a straight edge is extracted and a point of the detection distance cannot be obtained in the vicinity of the edge, it is determined that there is a low step or line. That is, it is possible to accurately detect a low step or line that is difficult to detect with only the distance detection unit and that may be erroneously detected as an edge of a three-dimensional object with only the imaging unit.

As a third specific aspect of the present invention, the extraction means in the present invention extracts an edge in the image as the feature point,
The target detection means has a high step as the detection target when the straight edge is extracted and a point of the detection distance detected by the distance detection means is obtained in the vicinity of the linear edge. It is characterized by comprising a high step discrimination means for discriminating that it is present.

  When there is a high step on the road surface, the edge of the high step (straight edge) is extracted from the image. Further, in the case of a high level difference, there is a high possibility of being detected by the distance detection means. Therefore, in the present invention, when a straight edge is extracted and a detection distance point is obtained in the vicinity of the edge, it is determined that there is a high step. As a result, when the distance detecting means and the imaging means are each used alone, it is possible to discriminate between a low step, a high step, a line, and a high step that are difficult to distinguish.

As a 4th specific aspect of this invention, the said object detection means in this invention is:
One-side detection means for detecting the detection object using one of the detection distance and the feature point detected by the distance detection means;
Of the detection distance and the feature point, the one used for detection of the detection target by the one-side detection means is the first information, and the one not used for the detection is second information, and the detection target based on the second information And invalidating means for invalidating the first information unrelated to the.

  According to the present invention, the detection target is detected using the detection distance of the distance detection means and one of the feature points obtained from the image (first information), and the other (second information) is supplemented during the detection. Used for. That is, since the second information has advantages and disadvantages different from the advantages and disadvantages of the first information, the second information is used to determine the first information (false detection information) unrelated to the detection target. it can. In the present invention, since the erroneous detection information is invalidated, erroneous detection of the detection target can be prevented.

As a more specific aspect of the fourth specific aspect, the one-side detection means in the present invention detects a parking frame line as the detection target based on the feature point,
The invalidating unit includes a three-dimensional object range setting unit that sets a range of a three-dimensional object in the image based on the point sequence of the detection distance, and invalidates the feature points belonging to the range of the three-dimensional object. And

  The detection of the distance detection unit is suitable for detection of a three-dimensional object, and the point sequence of the detection distance detected by the distance detection unit is considered to reflect the position of the three-dimensional object. To set the range of the three-dimensional object in the image. Since the feature points belonging to the range of the three-dimensional object are irrelevant to the parking frame line, the feature points are invalidated. Therefore, since feature points irrelevant to the parking frame line are removed, the detection accuracy of the parking frame line can be improved.

As yet another specific aspect of the fourth specific aspect, the one-side detection means in the present invention detects a three-dimensional object adjacent to a parking space as the detection target based on the detection distance,
The invalidating means includes road surface range setting means for setting a road surface range based on the feature points, and invalidates the detection distance points belonging to the road surface range.

  Since it is easy to determine the range of the road surface using the image (feature point) of the imaging means, in the present invention, the range of the road surface is set based on the feature point. Since the detection distance point belonging to the road surface range is irrelevant to the three-dimensional object, the detection distance point is invalidated. Therefore, since the detection distance point unrelated to the three-dimensional object is removed, the detection accuracy of the three-dimensional object can be improved.

1 is a block diagram illustrating a configuration of a parking assistance device 1. FIG. It is the figure which showed the detection range 27 of the ranging sensor 2, and the imaging range 31 of the camera 3. FIG. It is the figure which showed the exploration wave and the reflected wave on the time axis. It is the figure which showed the scene which detects the parking space 100 adjacent to the parked vehicle 6 when the parked vehicle 6 passes by the side. It is the figure which showed the scene where the level | step difference 200 exists between the parked vehicle 6 and the vehicle 5. FIG. It is a flowchart of the parking assistance process of 1st Embodiment. It is the figure which showed the exploration wave and the reflected wave on the time axis | shaft, and has shown the state which changed the detection range in the distance from the 1st reflected wave 261. FIG. FIG. 5 is a diagram showing a scene where a step 205 exists on the side of the vehicle 5. It is a flowchart of the process relevant to a ranging sensor among the parking assistance processes of 2nd-4th embodiment. It is a flowchart of the process relevant to a camera among the parking assistance processes of 2nd Embodiment. It is the figure which illustrated the scene of 3rd Embodiment. It is a flowchart of the process relevant to a camera among the parking assistance processes of 3rd Embodiment. It is the figure which illustrated the scene of 4th Embodiment. It is a flowchart of the process relevant to a camera among the parking assistance processes of 4th Embodiment. It is the figure which showed the scene where the parking vehicle 6 and the parking space 100 cannot be detected since the level | step difference 200 exists between the parked vehicle 6 and the vehicle 5. FIG.

(First embodiment)
Hereinafter, a first embodiment of a parking assistance apparatus according to the present invention will be described with reference to the drawings. FIG. 1 is a block diagram showing the configuration of the parking assistance device 1 of the present embodiment. The parking assistance device 1 is mounted on a vehicle 5 (see FIG. 2). As shown in FIG. 1, the parking assistance apparatus 1 includes a distance measuring sensor 2, a camera 3, a vehicle speed sensor 41, a steering angle sensor 42, and an ECU 10 connected thereto.

  The distance measuring sensor 2 is a sensor that detects a distance to an object such as a parked vehicle existing around the distance measuring sensor 2. Specifically, the distance measuring sensor 2 includes a control unit 21, a transmission / reception unit 22, a threshold setting unit 23, and a comparison unit 24. Based on an instruction from the ECU 10, the control unit 21 causes the transmission / reception unit 22 configured by a piezoelectric element or the like that transmits a search wave such as an ultrasonic wave to transmit the search wave. At this time, the transmission / reception unit 22 transmits an exploration wave at predetermined intervals (for example, every 100 milliseconds) in the front direction of the distance measuring sensor 2 (transmission / reception unit 22). The transmission / reception unit 22 receives a reflected wave reflected by the transmitted exploration wave hitting an object. The threshold setting unit 23 sets a threshold for the amplitude of the reflected wave.

  The reflected wave received by the transmission / reception unit 22 and the threshold set by the threshold setting unit 23 are input to the comparison unit 24. The comparison unit 24 compares the amplitude of the input reflected wave with a threshold value, and outputs a signal indicating that the amplitude exceeds the threshold value when the amplitude of the reflected wave exceeds the threshold value. The signal output from the comparison unit 24 is input to the control unit 21, and the control unit 21 receives the reflected wave based on the input of the signal (for example, the timing when the amplitude of the reflected wave exceeds the threshold value). ) (Hereinafter referred to as reception time) is notified to the ECU 10. Note that the control unit 21 can change the threshold set by the threshold setting unit 23.

  The distance measuring sensor 2 may be any sensor that transmits an exploration wave and receives a reflected wave of the exploration wave, and uses a radio wave regardless of whether it uses a sound wave or a light wave. May be. As the distance measuring sensor 2, for example, a sensor such as an ultrasonic sensor, a laser radar, or a millimeter wave radar can be used.

  FIG. 2 is a diagram for explaining the mounting positions of the distance measuring sensor 2 and the camera 3 in the vehicle 5 and shows the vehicle 5 as viewed from above. As shown in FIG. 2, the distance measuring sensor 2 is mounted toward the side of the vehicle 5 on the left and right side surfaces 51 of the vehicle 5 (in FIG. 2, the front portion of the left and right side surfaces 51). That is, the distance measuring sensor 2 mounted on the left side surface 51 transmits an exploration wave toward the left side of the vehicle 5, and the distance measuring sensor 2 mounted on the right side surface 51 searches toward the right side of the vehicle 5. Send a wave. FIG. 2 illustrates an object detection range 27 (search wave transmission range) of the distance measuring sensor 2. The directivity θ (the directivity of the exploration wave) of the detection range 27 is, for example, about 70 ° to 120 °. The maximum detection distance MaxL that can be detected by the distance measuring sensor 2 is, for example, about 2.5 m to 6 m.

  Further, when the distance measuring sensor 2 receives a plurality of reflected waves by one transmission of the exploration wave (when there are reflected waves from a plurality of objects), only the reflected wave exceeding the first threshold received is received. This sensor is valid and other reflected waves are invalid. FIG. 3 illustrates the exploration wave and the reflected wave on the time axis, and specifically illustrates the exploration wave 25 and the two reflected waves 261 and 262. FIG. 3 also shows a threshold line 231 set by the threshold setting unit 23. As shown in FIG. 3, when two reflected waves 261 and 262 are generated by one transmission of the exploration wave 25, only the first reflected wave 261 is validated (received). The next reflected wave 262 is invalid (not received). In the case of FIG. 3, the distance measuring sensor 2 notifies the ECU 10 of the reception time t0 of the first reflected wave 261 (the time from when the exploration wave 25 is transmitted until the reflected wave 261 is received).

  The camera 3 is mounted on the rear part of the vehicle 5, for example, and images the rear of the vehicle 5 at predetermined intervals based on an instruction from the ECU 10. As the camera 3, for example, a rear camera used to display a rear image of the vehicle 5 on a screen (not shown) in the vehicle interior during back parking, or to detect a parking frame line based on the rear image. Can do. As shown in FIG. 2, the imaging range 31 of the camera 3 is set to include at least a part 311 of the lateral range of the vehicle 5 in addition to the rear range of the vehicle 5. Therefore, the horizontal angle of view φ of the camera 3 is set to a wide angle (for example, about 150 ° to 180 °). An image captured by the camera 3 is input to the ECU 10.

  The vehicle speed sensor 41 is a sensor that detects the vehicle speed of the vehicle 5, and is, for example, a rotary encoder that detects the number of rotations of a wheel. The steering angle sensor 42 is a sensor that detects the steering angle (steering angle) of the vehicle 5. Detection information (vehicle speed, steering angle) detected by the vehicle speed sensor 41 and the steering angle sensor 42 is input to the ECU 10.

  The ECU 10 is mainly configured by a microcomputer (microcomputer) including a CPU, a ROM, a RAM, and the like, and is based on each detection information input from the distance measuring sensor 2, the camera 3, the vehicle speed sensor 41, and the steering angle sensor 42. The various processes which support the parking of 5 are performed. In addition, the ECU 10 includes a memory 11 such as a ROM or a RAM that stores various types of information necessary for processing executed by itself. A specific example of processing executed by the ECU 10 will be described. The ECU 10 instructs the distance measurement sensor 2 to transmit an exploration wave, and the distance measurement sensor 2 notifies the reflected wave reception time t0 (see FIG. 3). If there is, the distance to the object is calculated based on the reception time t0 and the speed of the exploration wave (the speed of sound when using an ultrasonic wave as the exploration wave). The distance measuring sensor 2 may calculate the distance to the object. In this case, the distance measuring sensor 2 notifies the ECU 10 of the calculated distance to the object.

  FIG. 4 is a diagram exemplifying a scene where the parking assistance process is performed by the ECU 10, and in detail, a basic scene for detecting the parking space 100 adjacent to the parked vehicle 6 when passing through the side passage 101 of the parked vehicle 6. Show. In FIG. 4, two parked vehicles 61 and 62 parked in parallel as the parked vehicle 6 are illustrated. The ECU 10 causes the distance measuring sensor 2 to sequentially detect the distance to the parked vehicle 6 when passing through the side passage 101. Then, the ECU 10 calculates the contour point of the parked vehicle 6 based on the detection distance detected by the distance measuring sensor 2. Specifically, the ECU 10, for example, the position of the distance measuring sensor 2 when distance detection is performed based on the vehicle speed detected by the vehicle speed sensor 41 and the steering angle detected by the steering angle sensor 42 (hereinafter referred to as sensor position). And a point 7 (hereinafter referred to as a distance detection point) that is separated from the sensor position by the detection distance in the front direction of the distance measuring sensor 2 is calculated as a contour point of the parked vehicle 6.

  In FIG. 4, the distance detection point for the first parked vehicle 61 through which the vehicle 5 passes is denoted by reference numeral “71”, and the distance detection point for the second parked vehicle 62 is denoted by reference numeral “72”. Since the detection range 27 of the distance measuring sensor 2 has a certain extent in the horizontal direction, the distance detection by the distance measuring sensor 2 is started before the vehicle 5 is positioned in front of the parked vehicle 6. The distance detection continues for a while after passing through the parked vehicle 6. As a result, as shown in FIG. 4, the point sequence of the distance detection points 7 includes a portion 7 b protruding from the parked vehicle 6 in addition to the portion 7 a detected on the surface of the parked vehicle 6. Since the protruding portion 7b is detected before and after the parked vehicle 6 passes, it is detected in the depth direction (direction perpendicular to the traveling direction of the vehicle 5) than the portion 7a. Therefore, the combined shape of the portions 7a and 7b is a curved shape.

  The ECU 10 sets a space between the distance detection point 71 for the first parked vehicle 61 and the distance detection point 72 for the second parked vehicle 62 as the parking space 100. Specifically, the ECU 10 obtains the position of the corner 611 of the first parked vehicle 61 based on the point sequence of the distance detection points 71. For example, the ECU 10 obtains a point offset from the end point 711 of the point sequence of the distance detection point 71 by a predetermined distance in the direction opposite to the traveling direction of the vehicle 5 as the position of the corner 611. Or based on the principle of triangulation based on the two adjacent sensor positions and the two detection distances at the two sensor positions under the assumption that the exploration wave is reflected at the same point between the two adjacent sensor positions. A point (hereinafter referred to as triangulation point) is obtained. Two adjacent sensor positions are defined as a first sensor position and a second sensor position, a detection distance at the first sensor position is defined as a first detection distance, and a detection distance at the second sensor position is defined as a second detection distance. The survey point is the intersection of the first arc with the first sensor position as the center and the radius of the first detection distance and the second arc with the second sensor position as the center and the radius of the second detection distance. The triangulation point is detected at a position substantially coincident with the vehicle body surface of the parked vehicle 6 (does not protrude from the parked vehicle 6). Then, for example, the end point of the point sequence of the triangulation points for the first parked vehicle 61 is obtained as the position of the corner 611 of the parked vehicle 61. Similarly, the ECU 10 determines the position of the corner 621 of the second parked vehicle 62 based on the point sequence of the distance detection points 72 or the triangulation point sequence of the second parked vehicle 62. Then, the ECU 10 sets the obtained space 100 between the corners 611 and 621 as a parking space.

  However, as described above in the “Problem to be Solved by the Invention”, there is a possibility that the parked vehicle cannot be detected well in an exceptional scene where there is a step between the parked vehicle and the vehicle. FIG. 5 shows the exceptional scene, in which a step 200 exists between the parked vehicle 6 and the vehicle 5. The step 200 extends in a straight line along the side passage 101 in front of the parked vehicle 6 and the parking space 100 (in front of the vehicle 5). The height of the step 200 is, for example, about several centimeters. In the scene of FIG. 5, the distance measuring sensor 2 may detect the distance of the step 200. As described with reference to FIG. 3, since the distance measuring sensor 2 can receive only the first reflected wave (it is not valid), if the distance of the step 200 is detected, the distance of the parked vehicle 6 existing behind the step 200 is detected. It becomes impossible to detect the parking space 100. In FIG. 5, a distance detection point 73 for the step 200 is illustrated.

  ECU10 performs the parking assistance process which enables the detection of the parked vehicle 6 (parking space 100) also in the scene of FIG. The details of the parking assistance process will be described below. FIG. 6 shows a flowchart of a parking support process executed by the ECU 10. The process of FIG. 6 is started when a support start switch (not shown) is operated by a driver, for example.

  When the process of FIG. 6 is started, first, the distance measurement sensor 2 is instructed to transmit a search wave and receive a reflected wave (S11). In addition, the camera 3 is instructed to capture an image (camera image) captured by the camera 3 (S11). Next, it is determined whether or not the amplitude of the reflected wave received by the distance measuring sensor 2 has exceeded the threshold set by the threshold setting unit 23 (see FIG. 1) (S12). Specifically, as described above, the control unit 21 of the distance measuring sensor 2 notifies the ECU 10 of the reception time of the reflected wave when the amplitude of the reflected wave exceeds the threshold value. Whether or not the reception time is notified of whether or not the amplitude of the reflected wave exceeds the threshold value is determined.

  When the amplitude of the reflected wave exceeds the threshold value, that is, when the distance measurement sensor 2 notifies the reception time (S12: Yes), the distance to the object based on the notified reception time and the speed of the exploration wave (Detection distance) is calculated (S13). Next, the calculated detection distance is stored in the memory 11 (see FIG. 1) (S14). After S14, the process proceeds to S15. The ECU 10 calculates the sensor position based on the vehicle speed and the steering angle when the distance measuring sensor 2 detects the distance, and the calculated sensor position is also stored in the memory 11 in association with the detected distance. In the scene of FIG. 5, the point sequence of the distance detection points 73 (detection distance and sensor position) is accumulated in the memory 11 as the vehicle 5 moves. In S12, when the amplitude of the reflected wave does not exceed the threshold value or when the reflected wave is not received in the first place, that is, when there is no notification of the reception time from the distance measuring sensor 2 (S12: No), the process proceeds to S15. .

  In S15, it is determined whether or not the number of accumulated sensing distances in the memory 11 exceeds a predetermined number of times, and the values of the sensing distances are equal. In S15, it is determined whether or not a point sequence (point sequence of linear distance detection points) of the distance detection points 73 (see FIG. 5) with respect to the step 200 is detected. Whether or not the values of the detection distances are the same may be determined, for example, by determining whether or not the detection distances are within a predetermined range. More specifically, for example, when the accumulated detection distances are L1, L2, and L3, the difference (L1-L2, L1-L3, L2-L3) between the detection distances L1, L2, and L3 is within a predetermined range. It is judged whether it is settled in.

  When the accumulated number of detection distances is less than the predetermined number (S15: No), there is no object such as the step 200, the parked vehicle 6 or the like around the vehicle 5, or even if it exists, immediately after the start of distance detection. Assuming that there are not enough data for the detection distance that can be used for various determinations, the process returns to S11, and the distance detection by the distance measuring sensor 2 and the imaging by the camera 3 are continued.

  Even when the number of accumulated detection distances exceeds the predetermined number, if the values of the respective detection distances are not equivalent (S15: No), an object exists around the vehicle 5, but the object has a step. It returns to S11, without performing the process after S16 noting that it is not a linear object, such as. In this case, assuming that it is the basic scene in FIG. 4, the distance detection points 71 and 72 of the parked vehicles 61 and 62 are not changed as described in FIG. Based on this, the parking space 100 is detected. Note that the determination in S15 is synonymous with determining whether the point sequence of the distance detection points is linear. And even if a part of the point sequence of the distance detection points is a straight line like the distance detection points 71 and 72 in FIG. 4, a curved part (part 7b or a part where the part 7a and the part 7b are combined) ) Is included, it is determined that the point sequence of the distance detection points is not linear (S15: No).

  On the other hand, when the number of accumulated detection distances in the memory 11 exceeds a predetermined number of times and the values of the respective detection distances are equal (S15: Yes), the scene of FIG. ), The processing after S16 is executed. That is, a straight edge (straight line) is extracted from the camera image acquired in S11 (S16). Various known methods can be adopted as a method of extracting a linear edge from a camera image. Here, for example, first, an edge point (luminance change point) is extracted from a camera image. Next, the extracted edge point is subjected to bird's-eye conversion. Finally, a well-known straight line fitting such as a Hough vote is applied to the edge point after bird's eye conversion to detect a straight edge.

  In S16, the distance detection points stored in the memory 11 may be superimposed on the camera image, and the linear edge may be extracted by limiting to the area near the superimposed distance detection points. As a result, the calculation load can be reduced. Considering FIG. 5 as a camera image, FIG. 5 shows a straight edge 201 extracted in S16 as an edge of the step 200. FIG. The edge 201 is extracted in the vicinity of the point sequence of the distance detection points 73.

  Next, it is determined whether or not the straight edge extracted in S16 exists in the vicinity of the detection distance (distance detection point) accumulated in the memory 11 (S17). Specifically, the point sequence of the distance detection points accumulated in the memory 11 is superimposed on the image including the linear edge extracted in S16. Then, it is determined whether or not a linear edge exists within a predetermined range near the point sequence of the superimposed distance detection points. When there is no straight edge (S17: No), the detection distance (distance detection point) accumulated in the memory 11 is a distance detection result of a straight three-dimensional object (such as a wall) or the like. It returns to S11 not being the distance detection result of.

  If a straight edge exists in the vicinity of the distance detection point (S17: Yes), it is determined that there is a step on the side of the vehicle 5, and the process proceeds to S18. Is a point sequence detected in the back of the parking space 100. As shown in FIG. 4, when a step 203 such as a curb exists in the back of the parking space 100, the step 203 may be detected as a distance. The point sequence of the distance detection points 75 of the step 203 is also linear, and a linear edge exists in the vicinity of the distance detection point 75. In S18, the detected point sequence of the distance detection points is the point sequence of the distance detection points 75 in FIG. 4 (the point sequence behind the parking space 100) or the point sequence of the distance detection points 73 in FIG. 5 (the parking space 100). This is a process for determining whether or not the point sequence is in front of. In the scene of FIG. 4, the distance detection point 71 of the first parked vehicle 61 is first detected, and then the distance detection point 75 of the step 203 is detected at a position away from the distance detection point 71 in the depth direction. The Therefore, in S18, for example, based on the positional relationship between the point sequence of the distance detection point currently detected and the point sequence of the distance detection point detected immediately before, the point sequence of the distance detection point currently detected is determined. It is determined whether the point sequence is detected in the back of the parking space 100. More specifically, in S18, the point sequence of the distance detection point currently detected is a predetermined distance in the depth direction from the point sequence of the distance detection point detected immediately before (for example, about the vehicle width (about 2 m)). It is determined whether or not the point sequence is detected in the back of the parking space 100 by determining whether or not it is detected at a position away from the above. When the point sequence of the linear distance detection points is a point sequence detected in the back of the parking space 100 (S18: Yes), the detection range of the distance measuring sensor 2 is changed as the scene of FIG. If not, return to S11. Thereby, it can prevent that the distance detection point 72 (refer FIG. 4) of the 2nd parked vehicle 62 is no longer detected because the detection range of the ranging sensor 2 is changed to a distant place.

  When the point sequence of the linear distance detection points is not the point sequence detected in the back of the parking space 100, that is, when the point sequence is detected in front of the parking space 100 (S18: No), the front side 5, the detection range of the distance measuring sensor 2 is changed farther than the detection distance (linear detection distance) of the step accumulated in the memory 11 (S19). Specifically, when a representative value representing the detection distance accumulated in the memory 11 (for example, the maximum detection distance) is Ld, a range from the distance measuring sensor 2 to the representative value Ld as shown in FIG. 271 is changed to a non-detection range in which no object is detected. As a result, the detection range of the distance measuring sensor 2 becomes a range 272 farther than the non-detection range 271.

  A specific method of changing to the far range 272 will be described with reference to FIG. FIG. 7 illustrates the exploration wave 25 and two reflected waves 261 and 262 reflected by the exploration wave 25 when hitting an object on the time axis. FIG. 7 corresponds to the scene of FIG. 5. The first reflected wave 261 is a reflected wave from the step 200, and the second reflected wave 262 is a reflected wave from the parked vehicle 6. As described above, only the first reflected wave 261 can be received before the detection range of the distance measuring sensor 2 is changed.

  In S19, as a first method of changing to the far range 272, as shown in FIG. 7A, the threshold value of the period T1 from when the exploration wave 25 is transmitted until the reception of the first reflected wave 261 is completed is set. The amplitude of the reflected wave in the period T1 is increased so as not to exceed the threshold value. In FIG. 7A, the changed threshold value 232 is illustrated. The threshold value 231 after the period T1 is not changed. The ECU 10 instructs the control unit 21 (see FIG. 1) of the distance measuring sensor 2 to change the threshold value, and the control unit 21 sets the threshold value set by the threshold value setting unit 23 based on the instruction as the threshold value 232 in the period T1. Change to Accordingly, the reflected wave 261 from the step 200 received in the period T1 becomes invalid when the threshold value is less than 232, so that the reflected wave 262 from the parked vehicle 6 can be received as an effective reflected wave. The period T1 corresponds to the non-detection range 271 in FIG. 2, and the period T1 and after corresponds to the far range 272. Since the first reflected wave 261 was received before the detection range was changed, when the detection range was changed, the period T1 was set based on the reception time of the reflected wave 261 received so far. Can be set.

  Further, as shown in FIG. 7B, only the period T2 in the vicinity of the first reflected wave 261 (the reception period of the reflected wave 261) may be changed to a large threshold 232. Since there is no reflected wave exceeding the threshold 231 before the period T2, the reflected wave 262 after the period T2 can be received as an effective reflected wave.

  As a second method of changing to the far range 272, the reflected wave in the period T1 may not be received in the period T1 in FIG. By starting reception of the reflected wave after the period T1 has elapsed, the reflected wave 262 from the parked vehicle 6 can be received as an effective reflected wave.

  After the detection range is changed to the far range 272 (see FIG. 2) in S19, the process returns to S11 to transmit the exploration wave and receive the reflected wave in the changed detection range. At this time, the reception time of the reflected wave 262 from the parked vehicle 6 is notified from the distance measurement sensor 2 to the ECU 10 (S12: Yes), and the ECU 10 calculates the distance to the parked vehicle 6 based on the reception time (S13). . The ECU 10 stores the calculated distance and sensor position in the memory 11 (S14).

  Since the detection distance of the parked vehicle 6 is accumulated in the memory 11 after the change to the far range 272, the point sequence of the distance detection points of the parked vehicle 6 includes a curved portion, and the values of the detection distances are not equal (S15). : No), after changing to the far range 272, the processing of S16 to S19 is not performed. In this case, as described in the basic scene of FIG. 4, the parking space 100 is detected based on the distance detection points 71 and 72 of the parked vehicles 61 and 62. When the parking space 100 can be detected, the ECU 10 sets a target parking position based on the detected parking space 100 and calculates a route to the target parking position, for example. Thereafter, the steering motor of the vehicle 5 is controlled to guide the vehicle 5 to the target parking position according to the calculated route.

  As described above, in the present embodiment, even when a distance measuring sensor that can receive only a single reflected wave is used, a three-dimensional vehicle such as a parked vehicle that exists in the back of the step in a scene in which there is a step in front. The distance of the object can be detected, and consequently the parking space adjacent to the three-dimensional object can be detected.

(Second Embodiment)
Next, a second embodiment of the parking assistance apparatus according to the present invention will be described focusing on the differences from the above embodiment. In this embodiment, as shown in FIG. 8, when there is a step 205 around or in the parking space on the side of the vehicle 5, it is possible to determine whether the step 205 is a high step or a low step (or line). It is an embodiment. Note that it is useful for parking support to determine the level of a step, and specifically, for example, it can be determined whether or not a vehicle can get over the step due to the difference in level of the step. For example, if the existence of a high step (eg, a step of 10 cm or more) such as a curb or a ring stop in a parking space can be determined, the vehicle may warn that it cannot get over that step or park at a position near the step. And can help. On the other hand, if it is possible to determine the presence of a low level difference (for example, about 3 cm) or a line, the vehicle can get over that level difference or line, and an unnecessary warning can be prevented.

  The structure of the parking assistance apparatus of this embodiment is the same as the structure (FIG. 1) of 1st Embodiment. The parking assist process executed by the ECU 10 is different from the process of the first embodiment. Hereinafter, the details of the parking assistance process of this embodiment will be described. In addition, this embodiment assumes the scene (scene of FIG. 4, FIG. 5) which detects the parking space at the time of the side passage of parking space similarly to 1st Embodiment.

  ECU10 is performing the process of the flowchart of FIG. 9, and the process of the flowchart of FIG. 10 in parallel as a parking assistance process. The process of FIG. 9 is a process related to the distance measuring sensor 2, and corresponds to S11 to S14 of FIG. The processing in FIG. 10 is processing related to the camera 3 (including determination of the level of the step). 9 and 10 is started when a support start switch (not shown) is operated by a driver, for example. First, the processing of FIG. 9 will be described. When the process of FIG. 9 is started, first, the distance measurement sensor 2 is instructed to transmit a search wave and receive a reflected wave (S21). Next, similarly to S12 of FIG. 6, it is determined whether or not the amplitude of the reflected wave received by the distance measuring sensor 2 has exceeded a threshold value (S22). When the amplitude of the reflected wave is less than the threshold value (S22: No), the process returns to S21 to transmit / receive the exploration wave and the reflected wave.

  When the amplitude of the reflected wave exceeds the threshold (S22: Yes), the detection distance and sensor position are calculated in the same manner as S13 in FIG. 6 (S23), and the calculated detection distance and sensor position are stored in the memory 11. (S24). Then, it returns to S21 and repeats above-mentioned S21-S24. As a result, distance detection points (detection distances) are accumulated in the memory 11 as the vehicle 5 moves.

  Next, the process of FIG. 10 will be described. When the process of FIG. 10 starts, first, the camera 3 is instructed to capture an image (camera image) captured by the camera 3 (S31). Next, an edge point (luminance change point) is extracted from the camera image (S32). Next, a straight edge (straight line) is detected based on the extracted edge point in the same manner as S16 in FIG. 6 (S33). Next, it is determined whether or not a straight edge has been detected (S34). In S34, when it is assumed that the curb stone 203 (see FIG. 4) in the back of the parking space 100 or the step 200 (see FIG. 5) in front of the parking space 100 is detected, the side of the vehicle 5 Within a predetermined distance, it may be determined whether or not a linear edge that is substantially parallel to the vehicle 5 (the angle formed with the direction of the vehicle 5 is within a predetermined angle (for example, 10 °)) can be detected. Thereby, it is possible to prevent an object existing in a range other than the side of the vehicle 5 (for example, another vehicle behind the vehicle 5) from being erroneously determined as a step.

  If the straight edge cannot be detected (S34: No), the process returns to S31 because there is no step around the vehicle 5. On the other hand, when a linear edge can be detected (S34: Yes), the process after S35 is performed on the assumption that there is a step around the vehicle 5. In S35, it is determined whether or not the distance detection point (detection distance) accumulated in the memory 11 in S24 of FIG. 9 exists within a predetermined range in the vicinity of the linear edge. Specifically, each distance detection point stored in the memory 11 is superimposed on an image including a linear edge, and the distance between each superimposed distance detection point and the linear edge in the image is within a predetermined range. It is determined whether it is within.

  The reason why the process of S35 is performed will be described with reference to FIG. 8. As shown in FIG. 8, when a step 205 exists in the detection range 27 of the distance measuring sensor 2, reflection from the step 205 depends on the level of the step 205. Waves may or may not return. Specifically, when the level difference 205 is high, the area of the reflection surface 205a of the level difference 205 is large, so that the distance measuring sensor 2 can receive a reflected wave exceeding the threshold and detect the distance of the level difference 205. That is, a distance detection point exists near the step 205. On the other hand, when the step 205 is low, the exploration wave does not hit the step 205, or even if it hits, the area of the reflecting surface 205a is small, so the distance measuring sensor 2 can only receive a reflected wave less than the threshold value, The distance of the step 205 cannot be detected. That is, there is no distance detection point near the step 205. Therefore, it is possible to determine the level of the step by determining whether a linear edge is detected and whether a distance detection point exists in the vicinity of the linear edge. In addition, in order to determine the presence of a level | step difference accurately, in S35, you may add the conditions that the point sequence of a distance detection point is linear. That is, in S35, it may be determined whether or not the point sequence of the distance detection points is linear and whether or not the linear point sequence exists in the vicinity of the linear edge. Whether or not the point detection point sequence is linear may be determined in the same manner as in S15 of FIG.

  In S35, when a distance detection point exists in the vicinity of the linear edge (S35: Yes), the process proceeds to S36, and it is determined that there is a high step around the vehicle 5. Specifically, it is determined that there is a high step (for example, a step of 10 cm or more that cannot be overtaken by the vehicle 5 such as a curb or a ring stop) at the position of the detected linear edge or distance detection point (S36). Then, it returns to S31 and repeats the above-mentioned process.

  In S35, when there is no distance detection point in the vicinity of the straight edge (S35: No), the process proceeds to S37 and a low step (for example, the vehicle 5 can easily get over the detected position of the straight edge). It is determined that there is a line (such as a parking frame line) marked on the road surface. Then, it returns to S31 and repeats the above-mentioned process.

  As described above, in this embodiment, when there is a step around the vehicle, the level of the step can be determined with high accuracy. In addition, the invention of the present embodiment can be applied not only when the parking space is passed laterally but also when a parking operation such as back parking is performed. When the present invention is applied during back parking, it is preferable to mount the distance measuring sensor toward the rear of the vehicle. This makes it possible to determine the type (level difference of steps) such as a wheel ring and a parking frame line that exists behind the vehicle during back parking, so that back parking can be appropriately supported.

(Third embodiment)
Next, a third embodiment of the parking assistance apparatus according to the present invention will be described with a focus on differences from the above embodiment. In the present embodiment, the parking frame line 202 (see FIG. 11) is detected. The parking assistance device of this embodiment has the same configuration as that of the first and second embodiments (FIG. 1). However, the distance measuring sensor 2 may have a specification capable of receiving a plurality of reflected waves. The parking assist process executed by the ECU 10 is different from the above embodiment. Hereinafter, the details of the parking assistance process of this embodiment will be described. In addition, FIG. 11 has shown the scene which this embodiment assumes. FIG. 11 shows a scene in which the vehicle 5 is back parked in the parking space 100 on which the parking frame line 202 is marked after passing the side of the parking space 100 sandwiched between the two parked vehicles 6. . In addition, in FIG. 11, although it is the scene of parallel parking (the scene where the parked vehicle 6 is parking in parallel), the scene of parallel parking may be sufficient.

  ECU10 is performing the process of the flowchart of FIG. 9, and the process of the flowchart of FIG. 12 in parallel as a parking assistance process. The process of FIG. 9 is a process related to the distance measuring sensor 2 as described in the second embodiment. The process of FIG. 12 is a process related to the camera 3. 9 and 12 starts when a support start switch (not shown) is operated by a driver, for example.

  If the parking assist process is executed in the scene of FIG. 11, the distance to the parked vehicle 6 is sequentially detected by the process of FIG. 9 when passing through the side passage 101, and the detected distance is accumulated in the memory 11. Go (S21-S24). FIG. 11 shows a distance detection point 7 (a point away from the distance measurement sensor 2 in the front direction of the distance measurement sensor 2 by a detection distance), and the distance detection point 7 is stored in the memory 11. become. In FIG. 11, the distance detection point 7 is detected on the front surface 63 of the parked vehicle 6.

  On the other hand, when the process of FIG. 12 is started, first, the camera 3 is instructed to capture an image (camera image) captured by the camera 3 (S41). Next, image preprocessing such as distortion correction and gray scale conversion is performed on the acquired camera image (S42). Next, an edge point (luminance conversion point) is extracted from the camera image on which image preprocessing has been performed (S43). Next, the extracted edge point is converted to bird's-eye view (S44). Considering FIG. 11 as the edge point image after the bird's eye conversion in S44, FIG. 11 illustrates the edge point 8 (x point) extracted in S43 and bird's eye converted in S44. The edge point 8 includes an edge point 82 of a three-dimensional object (parked vehicle 6 in FIG. 11) in addition to the edge point 81 of the parking frame line 202.

  Next, a range in which a three-dimensional object exists in the edge point image after bird's-eye conversion (hereinafter referred to as a three-dimensional object range) is set (S45). The object detection by the distance measuring sensor 2 is suitable for detecting an object (three-dimensional object) having a certain height or more, and the distance detection point that is the distance detection result of the distance measuring sensor 2 is a point on the three-dimensional object. I can guess. Therefore, in S45, a three-dimensional object range is set based on the distance detection points stored in the memory 11 in S24 of FIG. Specifically, the distance detection point in the memory 11 is superimposed on the edge point image. Considering FIG. 11 as an edge point image on which the distance detection points 7 are superimposed, for example, as described in FIG. 4 of the first embodiment, the corner position 65 of the parked vehicle 6 is determined based on the point sequence of the distance detection points 7. Set. A predetermined amount (general vehicle length) range 600 is set as a three-dimensional object range from the set corner position 65 to the depth direction P2 (direction perpendicular to the traveling direction 1 of the vehicle 5 when the side passage 101 moves). .

  Next, edge points belonging to the three-dimensional object range are removed (S46). In the example of FIG. 11, the edge point 82 belonging to the three-dimensional object range 600 is removed. Next, a straight line (straight edge) is detected from the edge point image by performing known straight line fitting such as Hough voting on the edge point that has not been removed (edge point 81 in the example of FIG. 11) ( S47). In the example of FIG. 11, straight lines 301a and 301b corresponding to the edge of one parking frame line 202a and straight lines 302a and 302b corresponding to the edge of the other parking frame line 202b are detected.

  Next, a parking frame (parking space) is determined based on the detected straight line (S48). Various known methods (for example, the method described in Japanese Patent Application Laid-Open No. 2012-80497) can be adopted as a method for determining the parking frame. As a basic idea, two straight lines (pairs) that are parallel within a predetermined distance are found, and the straight line of the pair is used as one parking frame line that constitutes the parking frame. Then, a set of a plurality of parking frame lines having a predetermined relationship is determined as a parking frame. In the example of FIG. 11, the parking frame line 202a is detected by the straight lines 301a and 301b, and the parking frame line 202b is detected by the straight lines 302a and 302b.

  Next, a target parking position is calculated based on the confirmed parking frame (S49). Thereafter, the steering motor or the like is controlled to guide the vehicle 5 to the target parking position (S49). When the guidance to the target parking position is completed, the process of the flowchart in FIG.

  As described above, in the present embodiment, since feature points (edge points) belonging to the three-dimensional object range set from the distance detection result of the distance measuring sensor are removed, the edge of the three-dimensional object is erroneously set as a parking frame line. Can be prevented. That is, the detection accuracy of the parking frame line can be improved. Moreover, since edge points irrelevant to the parking frame line are removed, the calculation load can be reduced as compared with the case where all the edge points are calculated to detect the parking frame line.

(Fourth embodiment)
Next, a parking assistance device according to a fourth embodiment of the present invention will be described with a focus on differences from the above embodiment. The parking assistance apparatus of this embodiment has the same configuration as that of the above embodiment (FIG. 1). The parking assist process executed by the ECU 10 is different from the above embodiment. Hereinafter, the details of the parking assistance process of this embodiment will be described. FIG. 13 shows a scene assumed in the present embodiment. The scene of FIG. 13 is the same as the scene of FIG. 4 of the first embodiment. When the parked vehicle 6 passes through the side passage 101, the distance sensor 2 detects the distance of the parked vehicle 6, and the distance detection result This is a scene where the parking space 100 adjacent to the parked vehicle 6 is detected.

  ECU10 is performing the process of the flowchart of FIG. 9, and the process of the flowchart of FIG. 14 in parallel as a parking assistance process. The process of FIG. 9 is a process related to the distance measuring sensor 2 as described in the second embodiment. The process of FIG. 14 is a process related to the camera 3. In FIG. 14, the same processes as those in FIG. 12 are denoted by the same reference numerals. The processing in FIGS. 9 and 14 starts when a support start switch (not shown) is operated by a driver, for example.

  If the parking support process is executed in the scene of FIG. 13, the distance to the parked vehicle 6 is sequentially detected by the process of FIG. 9 as the vehicle 5 moves through the side passage 101, and the detected distance (distance) Detection points) are accumulated in the memory 11 (S21 to S24). FIG. 13 illustrates a point sequence of the distance detection points 7 accumulated in the memory 11. The distance measuring sensor 2 assumes distance detection of a three-dimensional object such as the parked vehicle 6, but the road surface may be erroneously detected due to the exploration wave transmission range 27 (see FIG. 13) reaching the road surface. . FIG. 13 shows distance detection points 74 detected on the road surface.

  On the other hand, when the processing of FIG. 14 is started, a camera image is acquired from the camera 3 (S41), image preprocessing is performed on the acquired camera image (S42), edge points are extracted (S43), and extracted. The edge point is converted into a bird's eye view (S44). Next, the road surface range in the edge point image after bird's-eye conversion is set (S51). Considering FIG. 13 as an edge point image, the processing of S51 will be described with reference to FIG. 13. In the edge point image, the portion 91 through which the vehicle 5 has passed immediately before and the luminance equivalent to that portion 91 are obtained. The connected portion 92 (the portion having no edge) is set as the road surface range 9. Since a method for extracting a road surface range from an image is known (see, for example, Japanese Patent Application Laid-Open No. 2007-310805), in S51, a known road surface range extraction method may be adopted to set the road surface range.

  Next, the distance detection points stored in the memory 11 are superimposed on the edge point image, and the distance detection points (detection distances) belonging to the road surface range are removed from the superimposed distance detection points (S52). In the example of FIG. 13, the distance detection point 74 belonging to the road surface range 9 is removed.

  Next, as described in the scene of FIG. 4 of the first embodiment, the position of the parked vehicle is determined based on the detection distance (distance detection point or triangulation point). That is, the corner 6 a of the parked vehicle 6 is obtained as the position of the parked vehicle 6. Next, the parking space 100 is set based on the obtained position of the corner 6a (S54). Then, the process of the flowchart of FIG. Thereafter, automatic parking is performed in the set parking space.

  As described above, in this embodiment, the detection distance within the road surface range unrelated to the three-dimensional object such as the parked vehicle is removed, so that the detection accuracy of the three-dimensional object such as the parked vehicle by the distance measuring sensor can be improved (incorrectly). Detection can be reduced).

  In addition, the parking assistance apparatus which concerns on this invention is not limited to the said embodiment, A various change is possible in the limit which does not deviate from description of a claim. For example, in the above-described embodiment, an example in which the rear camera is employed has been described, but a side camera that captures an image of the side of the vehicle may be employed. Further, the present invention can be applied to both parallel parking and parallel parking.

DESCRIPTION OF SYMBOLS 1 Parking assistance apparatus 2 Distance sensor 3 Camera 5 Vehicle 6, 61, 62 Parked vehicle 7, 71, 72, 73, 74, 75 Distance detection point 8, 81, 82 Edge point 200, 205 Level difference 202 Parking frame line

Claims (13)

Distance detection means (2) for sequentially transmitting exploration waves to the periphery of the vehicle (5), receiving reflected waves reflected by the exploration waves hitting the object, and sequentially detecting the distance to the object based on the reflected waves;
Imaging means (3) for imaging the periphery of the vehicle;
Extraction means (S16, S32, S43) for extracting feature points in the image captured by the imaging means;
Object detection means (10) for detecting a detection object (6, 61, 62, 205, 202) around or in the parking space using both the distance detection result by the distance detection means and the extraction result of the feature points When,
Equipped with a,
The distance detection means transmits the exploration wave to the side of the vehicle when passing sideways in a parking space,
The imaging means captures a range including at least the side of the vehicle when passing laterally in a parking space;
The extraction means extracts an edge in the image as the feature point,
The object detection means includes
The detection range of the distance detection unit when the point sequence (7) of the detection distance detected by the distance detection unit is linear and the linear edge is extracted in the vicinity of the linear point sequence. Detection range changing means (S15 to S17, S19) for changing the distance from the linear detection distance, which is the detection distance to be a linear point sequence,
A three-dimensional object detection unit (6, 61, 62) that detects a three-dimensional object (6, 61, 62) adjacent to the parking space (100) based on the detection distance detected after the detection range is changed by the detection range change unit ( And 10) a parking assistance device (1). The distance detection means is effective only for the first reflected wave having an amplitude exceeding a predetermined threshold among the reflected waves received by one transmission of the exploration wave, based on the reflected wave. The parking assistance device according to claim 1, wherein a distance to an object is detected. The detection range changing means (S19) is characterized in that the threshold value in the reception period of the reflected wave that is a detection source of the linear detection distance is increased so that the amplitude of the reflected wave does not exceed the threshold value. The parking assistance device according to claim 2. The parking support apparatus according to claim 1 or 2, wherein the detection range changing means (S19) does not accept reception of the reflected wave from a range from the vehicle to the linear detection distance. . When the detection distance point sequence includes a curved portion, even if the detection distance point sequence includes a linear portion, the detection range changing means (S15) The parking assistance device according to any one of claims 1 to 4, wherein the detection range is not changed. A back point sequence judging means (S18) for judging that the point sequence of the detection distance is a point sequence (75) detected in the back of the parking space;
  The detection range changing means (S19) is configured to detect the detection distance even if the point sequence of the detection distance is linear and the linear edge is extracted in the vicinity of the linear point sequence. 6. The change of the detection range is stopped when the back point sequence determination unit determines that the point sequence is a point sequence detected in the back of the parking space. The parking assistance device according to item. The extraction means extracts an edge in the image as the feature point,
  The target detection unit is configured to detect a low level difference as the detection target when the linear edge is extracted and a point of the detection distance detected by the distance detection unit is not obtained in the vicinity of the linear edge. The parking assist device according to any one of claims 1 to 6, further comprising low step determining means (S34, S35, S37) for determining that there is a line written on the road surface. The extraction means extracts an edge in the image as the feature point,
  The target detection means has a high step as the detection target when the straight edge is extracted and a point of the detection distance detected by the distance detection means is obtained in the vicinity of the linear edge. The parking assist device according to any one of claims 1 to 7, further comprising a high step determining means (S34, S35, S36) for determining that there is one. Distance detection means (2) for sequentially transmitting exploration waves to the periphery of the vehicle (5), receiving reflected waves reflected by the exploration waves hitting the object, and sequentially detecting the distance to the object based on the reflected waves;
  Imaging means (3) for imaging the periphery of the vehicle;
  Extraction means (S16, S32, S43) for extracting feature points in the image captured by the imaging means;
  Object detection means (10) for detecting a detection object (6, 61, 62, 205, 202) around or in the parking space using both the distance detection result by the distance detection means and the extraction result of the feature points When,
  With
The object detection means includes
  One-side detection means (10) for detecting the detection object using one of the detection distance detected by the distance detection means and the feature point;
  Of the detection distance and the feature point, the one used for detection of the detection target by the one-side detection means is the first information, and the one not used for the detection is second information, and the detection target based on the second information A parking assistance device (1), comprising invalidation means (S45, S46, S51, S52) for invalidating the first information unrelated to the vehicle. The one-side detecting means detects a parking frame line as the detection target based on the feature point,
  The invalid means (S46, S45) includes a three-dimensional object range setting means for setting a three-dimensional object range (600) in the image based on the point sequence of the detection distance, and the feature belonging to the three-dimensional object range. 10. The parking assistance device according to claim 9, wherein the point (82) is invalidated. The one-side detection means detects a three-dimensional object adjacent to a parking space as the detection target based on the detection distance,
  The invalidation means (S51, S52) includes road surface range setting means for setting a road surface range (9) based on the feature points, and invalidates the detection distance point (74) belonging to the road surface range. The parking support apparatus according to claim 9, wherein   Distance detection means (2) for sequentially transmitting exploration waves to the periphery of the vehicle (5), receiving reflected waves reflected by the exploration waves hitting the object, and sequentially detecting the distance to the object based on the reflected waves;
  Imaging means (3) for imaging the periphery of the vehicle;
  Extraction means (S16, S32, S43) for extracting feature points in the image captured by the imaging means;
  Object detection means (10) for detecting a detection object (6, 61, 62, 205, 202) around or in the parking space using both the distance detection result by the distance detection means and the extraction result of the feature points When,
  With
The extraction means extracts an edge in the image as the feature point,
  The target detection unit is configured to detect a low level difference as the detection target when the linear edge is extracted and a point of the detection distance detected by the distance detection unit is not obtained in the vicinity of the linear edge. A parking assist device (1), comprising low step determining means (S34, S35, S37) for determining that there is a line marked on the road surface. Distance detection means (2) for sequentially transmitting exploration waves to the periphery of the vehicle (5), receiving reflected waves reflected by the exploration waves hitting the object, and sequentially detecting the distance to the object based on the reflected waves;
  Imaging means (3) for imaging the periphery of the vehicle;
  Extraction means (S16, S32, S43) for extracting feature points in the image captured by the imaging means;
  Object detection means (10) for detecting a detection object (6, 61, 62, 205, 202) around or in the parking space using both the distance detection result by the distance detection means and the extraction result of the feature points When,
  With
  The extraction means extracts an edge in the image as the feature point,
  The target detection means has a high step as the detection target when the straight edge is extracted and a point of the detection distance detected by the distance detection means is obtained in the vicinity of the linear edge. A parking assist device (1), comprising high step determining means (S34, S35, S36) for determining that there is a parking space.

Images