WO2013094242A1 - 立体物検出装置 - Google Patents
立体物検出装置 Download PDFInfo
- Publication number
- WO2013094242A1 WO2013094242A1 PCT/JP2012/069095 JP2012069095W WO2013094242A1 WO 2013094242 A1 WO2013094242 A1 WO 2013094242A1 JP 2012069095 W JP2012069095 W JP 2012069095W WO 2013094242 A1 WO2013094242 A1 WO 2013094242A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- dimensional object
- image
- bird
- difference
- object detection
- Prior art date
Links
Images
Classifications
-
- G—PHYSICS
- G08—SIGNALLING
- G08G—TRAFFIC CONTROL SYSTEMS
- G08G1/00—Traffic control systems for road vehicles
- G08G1/16—Anti-collision systems
- G08G1/166—Anti-collision systems for active traffic, e.g. moving vehicles, pedestrians, bikes
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B60—VEHICLES IN GENERAL
- B60R—VEHICLES, VEHICLE FITTINGS, OR VEHICLE PARTS, NOT OTHERWISE PROVIDED FOR
- B60R1/00—Optical viewing arrangements; Real-time viewing arrangements for drivers or passengers using optical image capturing systems, e.g. cameras or video systems specially adapted for use in or on vehicles
- B60R1/002—Optical viewing arrangements; Real-time viewing arrangements for drivers or passengers using optical image capturing systems, e.g. cameras or video systems specially adapted for use in or on vehicles specially adapted for covering the peripheral part of the vehicle, e.g. for viewing tyres, bumpers or the like
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/13—Edge detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/50—Context or environment of the image
- G06V20/56—Context or environment of the image exterior to a vehicle by using sensors mounted on the vehicle
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/50—Context or environment of the image
- G06V20/56—Context or environment of the image exterior to a vehicle by using sensors mounted on the vehicle
- G06V20/58—Recognition of moving objects or obstacles, e.g. vehicles or pedestrians; Recognition of traffic objects, e.g. traffic signs, traffic lights or roads
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/50—Context or environment of the image
- G06V20/56—Context or environment of the image exterior to a vehicle by using sensors mounted on the vehicle
- G06V20/588—Recognition of the road, e.g. of lane markings; Recognition of the vehicle driving pattern in relation to the road
-
- G—PHYSICS
- G08—SIGNALLING
- G08G—TRAFFIC CONTROL SYSTEMS
- G08G1/00—Traffic control systems for road vehicles
- G08G1/16—Anti-collision systems
- G08G1/167—Driving aids for lane monitoring, lane changing, e.g. blind spot detection
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B60—VEHICLES IN GENERAL
- B60R—VEHICLES, VEHICLE FITTINGS, OR VEHICLE PARTS, NOT OTHERWISE PROVIDED FOR
- B60R2300/00—Details of viewing arrangements using cameras and displays, specially adapted for use in a vehicle
- B60R2300/60—Details of viewing arrangements using cameras and displays, specially adapted for use in a vehicle characterised by monitoring and displaying vehicle exterior scenes from a transformed perspective
- B60R2300/607—Details of viewing arrangements using cameras and displays, specially adapted for use in a vehicle characterised by monitoring and displaying vehicle exterior scenes from a transformed perspective from a bird's eye viewpoint
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B60—VEHICLES IN GENERAL
- B60R—VEHICLES, VEHICLE FITTINGS, OR VEHICLE PARTS, NOT OTHERWISE PROVIDED FOR
- B60R2300/00—Details of viewing arrangements using cameras and displays, specially adapted for use in a vehicle
- B60R2300/80—Details of viewing arrangements using cameras and displays, specially adapted for use in a vehicle characterised by the intended use of the viewing arrangement
- B60R2300/802—Details of viewing arrangements using cameras and displays, specially adapted for use in a vehicle characterised by the intended use of the viewing arrangement for monitoring and displaying vehicle exterior blind spot views
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B60—VEHICLES IN GENERAL
- B60R—VEHICLES, VEHICLE FITTINGS, OR VEHICLE PARTS, NOT OTHERWISE PROVIDED FOR
- B60R2300/00—Details of viewing arrangements using cameras and displays, specially adapted for use in a vehicle
- B60R2300/80—Details of viewing arrangements using cameras and displays, specially adapted for use in a vehicle characterised by the intended use of the viewing arrangement
- B60R2300/804—Details of viewing arrangements using cameras and displays, specially adapted for use in a vehicle characterised by the intended use of the viewing arrangement for lane monitoring
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10004—Still image; Photographic image
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20048—Transform domain processing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30248—Vehicle exterior or interior
- G06T2207/30252—Vehicle exterior; Vicinity of vehicle
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30248—Vehicle exterior or interior
- G06T2207/30252—Vehicle exterior; Vicinity of vehicle
- G06T2207/30256—Lane; Road marking
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30248—Vehicle exterior or interior
- G06T2207/30252—Vehicle exterior; Vicinity of vehicle
- G06T2207/30261—Obstacle
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N7/00—Television systems
- H04N7/18—Closed-circuit television [CCTV] systems, i.e. systems in which the video signal is not broadcast
Definitions
- the present invention relates to a three-dimensional object detection device.
- a vehicle periphery monitoring device that detects a three-dimensional object around a vehicle and detects whether the three-dimensional object is stationary or moving has been proposed.
- this vehicle periphery monitoring device two images taken at different times are converted into bird's-eye views, the two bird's-eye views are aligned, the difference between the two images after alignment is taken, and the inconsistent portion is stereoscopically converted. It detects as a thing (refer patent document 1).
- the problem to be solved by the present invention is to provide a three-dimensional object detection device capable of increasing the detection accuracy of a three-dimensional object regardless of the brightness of the detection environment.
- the present invention solves the above problem by increasing the weight of the detection result based on edge information when the detection environment is dark, and increasing the weight of the detection result based on differential waveform information when the detection environment is bright.
- Increasing the weight of the detection result based on edge information when the detection environment is dark has little effect on the edge information even if there is a change in the light and darkness of the road surface that causes false detection in the differential waveform information. Can be increased. Also, when the detection environment is bright and the weight of the detection result based on the differential waveform information is increased, even if there is a road surface pattern that causes false detection in the edge information, the differential waveform information has little effect. Detection accuracy can be increased.
- FIG. 1 is a schematic configuration diagram of a vehicle according to an embodiment to which a three-dimensional object detection device of the present invention is applied. It is a top view (three-dimensional object detection by difference waveform information) which shows the driving state of the vehicle of FIG. It is a block diagram which shows the detail of the computer of FIG. 4A and 4B are diagrams for explaining the outline of processing of the alignment unit in FIG. 3, in which FIG. 3A is a plan view showing a moving state of the vehicle, and FIG. It is the schematic which shows the mode of the production
- FIG. 4 is a flowchart (No. 1) illustrating a three-dimensional object detection method using differential waveform information executed by the viewpoint conversion unit, the alignment unit, the smear detection unit, and the first three-dimensional object detection unit of FIG. 3.
- FIG. 2 It is a flowchart (the 2) which shows the solid object detection method using the difference waveform information performed by the viewpoint conversion part of FIG. 3, a position alignment part, a smear detection part, and a 1st solid object detection part.
- FIG. 3 It is a figure (three-dimensional object detection by edge information) which shows the running state of vehicles of Drawing 1, (a) is a top view showing the positional relationship of a detection field etc., and (b) shows the positional relationship of a detection field etc. in real space. It is a perspective view shown.
- 4A and 4B are diagrams for explaining the operation of the luminance difference calculation unit in FIG. 3, in which FIG.
- 3A is a diagram illustrating a positional relationship among attention lines, reference lines, attention points, and reference points in a bird's eye view image; It is a figure which shows the positional relationship of the attention line, reference line, attention point, and reference point.
- 4A and 4B are diagrams for explaining the detailed operation of the luminance difference calculation unit in FIG. 3, in which FIG. 3A is a diagram illustrating a detection region in a bird's-eye view image, and FIG. It is a figure which shows the positional relationship of a reference point.
- FIG. 6 is a flowchart (part 1) illustrating a three-dimensional object detection method using edge information executed by a viewpoint conversion unit, a luminance difference calculation unit, an edge line detection unit, and a second three-dimensional object detection unit in FIG. 3;
- FIG. 1 is a flowchart (part 1) illustrating a three-dimensional object detection method using edge information executed by a viewpoint conversion unit, a luminance difference calculation unit, an edge line detection unit, and a second three-dimensional object detection unit in FIG. 3;
- FIG. 6 is a flowchart (part 2) illustrating the three-dimensional object detection method using edge information executed by the viewpoint conversion unit, the luminance difference calculation unit, the edge line detection unit, and the second three-dimensional object detection unit of FIG. 3. It is a figure which shows the example of an image for demonstrating edge detection operation
- control map which shows the relationship of ratio (alpha) / (beta) of the 1st threshold value and 2nd threshold value with respect to brightness. It is another example of the control map which shows the relationship of ratio (alpha) / (beta) of the 1st threshold value and 2nd threshold value with respect to brightness. It is a control map which shows the example of a setting of 1st threshold value (alpha) and 2nd threshold value (beta) with respect to brightness. It is a control map which shows the example of a setting of the weight X of the detection result by the 1st three-dimensional object detection part with respect to brightness, and the weight Y of the detection result by the 2nd three-dimensional object detection part.
- FIG. 10 is a diagram (No.
- FIG. 10 is a diagram (No. 2) for describing another example of the processing of the alignment unit in FIG. 3;
- FIG. 10 is a diagram (No. 3) for describing another example of the process of the alignment unit in FIG. 3;
- FIG. 1 is a schematic configuration diagram of a vehicle according to an embodiment to which a three-dimensional object detection device 1 of the present invention is applied.
- the three-dimensional object detection device 1 of the present example is in contact with the host vehicle V when changing lanes. The purpose is to detect other vehicles with potential and calculate the travel distance. For this reason, the example demonstrated below shall show the example which mounts the solid-object detection apparatus 1 in the vehicle V, and makes a succeeding vehicle the solid object of a detection target.
- the three-dimensional object detection device 1 of the present example includes a camera 10, a vehicle speed sensor 20, a calculator 30, and a brightness detection unit 40.
- the camera 10 is attached to the vehicle V so that the optical axis is at an angle ⁇ from the horizontal to the lower side at the height h at the rear of the vehicle V.
- the camera 10 captures an image of a predetermined area in the surrounding environment of the vehicle V from this position.
- the vehicle speed sensor 20 detects the traveling speed of the vehicle V, and calculates the vehicle speed from the wheel speed detected by, for example, a wheel speed sensor that detects the rotational speed of the wheel.
- the computer 30 detects a three-dimensional object behind the vehicle, and calculates a moving distance and a moving speed for the three-dimensional object in this example.
- the brightness detection unit 40 detects the brightness of a predetermined area imaged by the camera 10, and details thereof will be described later.
- FIG. 2 is a plan view showing a traveling state of the vehicle V in FIG.
- the camera 10 images the vehicle rear side at a predetermined angle of view a.
- the angle of view a of the camera 10 is set to an angle of view at which the left and right lanes can be imaged in addition to the lane in which the host vehicle V travels.
- FIG. 3 is a block diagram showing details of the computer 30 of FIG. In FIG. 3, the camera 10, the vehicle speed sensor 20, and the brightness detection unit 40 are also illustrated in order to clarify the connection relationship.
- the computer 30 includes a viewpoint conversion unit 31, an alignment unit 32, a first three-dimensional object detection unit 33, a smear detection unit 34, a luminance difference calculation unit 35, and an edge line detection unit 36. And a second three-dimensional object detection unit 37, a three-dimensional object determination unit 38, and a threshold setting unit 39.
- the viewpoint conversion unit 31, the smear detection unit 34, the alignment unit 32, and the first three-dimensional object detection unit 33 are components related to a three-dimensional object detection block using differential waveform information described later, and the viewpoint conversion unit 31.
- the luminance difference calculation unit 35, the edge line detection unit 36, and the second three-dimensional object detection unit 37 are components related to a three-dimensional object detection block using edge information described later.
- each component will be described first.
- the viewpoint conversion unit 31 inputs captured image data of a predetermined area obtained by imaging with the camera 10 and converts the input captured image data into a bird's-eye image data in a bird's-eye view state.
- the state viewed from a bird's-eye view is a state viewed from the viewpoint of a virtual camera looking down from above, for example, vertically downward.
- This viewpoint conversion can be executed as described in, for example, Japanese Patent Application Laid-Open No. 2008-219063.
- the viewpoint conversion of captured image data to bird's-eye view image data is based on the principle that a vertical edge peculiar to a three-dimensional object is converted into a straight line group passing through a specific fixed point by viewpoint conversion to bird's-eye view image data. This is because a planar object and a three-dimensional object can be distinguished if used. Note that the result of the image conversion processing by the viewpoint conversion unit 31 is also used in detection of a three-dimensional object by edge information described later.
- the alignment unit 32 sequentially inputs the bird's-eye image data obtained by the viewpoint conversion of the viewpoint conversion unit 31, and aligns the positions of the inputted bird's-eye image data at different times.
- 4A and 4B are diagrams for explaining the outline of the processing of the alignment unit 32, where FIG. 4A is a plan view showing the moving state of the vehicle V, and FIG. 4B is an image showing the outline of the alignment.
- the host vehicle V at the current time is located at V1, and the host vehicle V one hour before is located at V2.
- the other vehicle V is located in the rear direction of the own vehicle V and is in parallel with the own vehicle V, the other vehicle V at the current time is located at V3, and the other vehicle V one hour before is located at V4.
- the host vehicle V has moved a distance d at one time.
- “one hour before” may be a past time for a predetermined time (for example, one control cycle) from the current time, or may be a past time for an arbitrary time.
- the bird's-eye image PB t at the current time is as shown in Figure 4 (b).
- the bird's-eye image PB t becomes a rectangular shape for the white line drawn on the road surface, but a relatively accurate is a plan view state, tilting occurs about the other vehicle V3.
- the white line drawn on the road surface has a rectangular shape and is relatively accurately viewed in plan, but the other vehicle V4 falls down.
- the vertical edges of solid objects are straight lines along the collapse direction by the viewpoint conversion processing to bird's-eye view image data. This is because the plane image on the road surface does not include a vertical edge, but such a fall does not occur even when the viewpoint is changed.
- the alignment unit 32 performs alignment of the bird's-eye images PB t and PB t ⁇ 1 as described above on the data. At this time, the alignment unit 32 is offset a bird's-eye view image PB t-1 before one unit time, to match the position and bird's-eye view image PB t at the current time.
- the image on the left side and the center image in FIG. 4B show a state that is offset by the movement distance d ′.
- This offset amount d ′ is a movement amount on the bird's-eye view image data corresponding to the actual movement distance d of the host vehicle V shown in FIG. It is determined based on the time until the time.
- the alignment unit 32 takes the difference between the bird's-eye images PB t and PB t ⁇ 1 and generates data of the difference image PD t .
- the pixel value of the difference image PD t may be an absolute value of the difference between the pixel values of the bird's-eye images PB t and PB t ⁇ 1 , and the absolute value is predetermined in order to cope with a change in the illuminance environment. “1” may be set when the threshold value is exceeded, and “0” may be set when the threshold value is not exceeded.
- the image on the right side of FIG. 4B is the difference image PD t .
- the alignment unit 32 of this example aligns the positions of the bird's-eye view images at different times on the bird's-eye view, and obtains the aligned bird's-eye view images.
- it can be performed with accuracy according to the required detection accuracy. It may be a strict alignment process such as aligning positions based on the same time and the same position, or may be a loose alignment process that grasps the coordinates of each bird's-eye view image.
- the first three-dimensional object detection unit 33 detects a three-dimensional object based on the data of the difference image PD t shown in FIG. At this time, the first three-dimensional object detection unit 33 of this example also calculates the movement distance of the three-dimensional object in the real space. In detecting the three-dimensional object and calculating the movement distance, the first three-dimensional object detection unit 33 first generates a differential waveform.
- the first three-dimensional object detection unit 33 sets a detection region in the difference image PD t .
- the three-dimensional object detection device 1 of the present example is for the purpose of calculating the travel distance for other vehicles that may be contacted when the host vehicle V changes lanes. For this reason, in this example, rectangular detection areas A1 and A2 are set on the rear side of the host vehicle V as shown in FIG. Such detection areas A1 and A2 may be set from a relative position with respect to the host vehicle V, or may be set based on the position of the white line. When setting the position of the white line as a reference, the movement distance detection device 1 may use, for example, an existing white line recognition technique.
- the first three-dimensional object detection unit 33 recognizes the sides (sides along the traveling direction) of the set detection areas A1 and A2 on the own vehicle V side as the ground lines L1 and L2.
- the ground line means a line in which the three-dimensional object contacts the ground.
- the ground line is set as described above, not a line in contact with the ground. Even in this case, from experience, the difference between the ground wire according to the present embodiment and the ground wire obtained from the position of the other vehicle V is not too large, and there is no problem in practical use.
- FIG. 5 is a schematic diagram illustrating how a differential waveform is generated by the first three-dimensional object detection unit 33 illustrated in FIG. 3.
- the first three-dimensional object detection unit 33 starts from the portion corresponding to the detection areas A1 and A2 in the difference image PD t (right diagram in FIG. 4B) calculated by the alignment unit 32.
- a differential waveform DW t is generated.
- the first three-dimensional object detection unit 33 generates a differential waveform DW t along the direction in which the three-dimensional object falls due to viewpoint conversion.
- the difference waveform DW t is generated for the detection area A2 in the same procedure.
- the first three-dimensional object detection unit 33 defines a line La in the direction in which the three-dimensional object falls on the data of the difference image PD t . Then, the first three-dimensional object detection unit 33 counts the number of difference pixels DP indicating a predetermined difference on the line La.
- the difference pixel DP indicating a predetermined difference exceeds a predetermined threshold when the pixel value of the difference image PDt is an absolute value of the difference between the pixel values of the bird's-eye images PB t and PB t ⁇ 1.
- the pixel value of the difference image PDt is expressed by “0” and “1”, the pixel indicates “1”.
- the first three-dimensional object detection unit 33 counts the number of difference pixels DP and then obtains an intersection CP between the line La and the ground line L1. Then, the first three-dimensional object detection unit 33 associates the intersection point CP with the count number, determines the horizontal axis position based on the position of the intersection point CP, that is, the position on the vertical axis in the right diagram of FIG. The vertical axis position, that is, the position on the horizontal axis in the right diagram of FIG. 5, is determined and plotted as the count number at the intersection CP.
- the first three-dimensional object detection unit 33 defines lines Lb, Lc,... In the direction in which the three-dimensional object falls, counts the number of difference pixels DP, and determines the horizontal axis position based on the position of each intersection CP. And the vertical axis position is determined from the count number (number of difference pixels DP) and plotted.
- the first three-dimensional object detection unit 33 generates the differential waveform DW t as shown in the right diagram of FIG.
- the line La and the line Lb in the direction in which the three-dimensional object collapses have different distances overlapping the detection area A1. For this reason, if the detection area A1 is filled with the difference pixels DP, the number of difference pixels DP is larger on the line La than on the line Lb. For this reason, when the first three-dimensional object detection unit 33 determines the vertical axis position from the count number of the difference pixels DP, the first three-dimensional object detection unit 33 is based on the distance at which the lines La and Lb in the direction in which the three-dimensional object falls and the detection area A1 overlap. Normalize. As a specific example, in the left diagram of FIG.
- the first three-dimensional object detection unit 33 normalizes the count number by dividing it by the overlap distance.
- the difference waveform DW t the line La on the direction the three-dimensional object collapses, the value of the differential waveform DW t corresponding to Lb is substantially the same.
- the first three-dimensional object detection unit 33 calculates the movement distance by comparison with the difference waveform DW t ⁇ 1 one time before. That is, the first three-dimensional object detection unit 33 calculates the movement distance from the time change of the difference waveforms DW t and DW t ⁇ 1 .
- the first three-dimensional object detection unit 33 divides the differential waveform DW t into a plurality of small areas DW t1 to DW tn (n is an arbitrary integer equal to or greater than 2) as shown in FIG.
- FIG. 6 is a diagram illustrating the small areas DW t1 to DW tn divided by the first three-dimensional object detection unit 33.
- the small areas DW t1 to DW tn are divided so as to overlap each other, for example, as shown in FIG. For example, the small area DW t1 and the small area DW t2 overlap, and the small area DW t2 and the small area DW t3 overlap.
- the first three-dimensional object detection unit 33 obtains an offset amount (a movement amount of the differential waveform in the horizontal axis direction (vertical direction in FIG. 6)) for each of the small regions DW t1 to DW tn .
- the offset amount is determined from the difference between the differential waveform DW t in the difference waveform DW t-1 and the current time before one unit time (distance in the horizontal axis direction).
- the first three-dimensional object detection unit 33 for each small area DW t1 ⁇ DW tn, when moving the differential waveform DW t1 before one unit time in the horizontal axis direction, the differential waveform DW t at the current time Is determined as the offset (the position in the horizontal axis direction), and the amount of movement in the horizontal axis between the original position of the differential waveform DW t ⁇ 1 and the position where the error is minimum is determined. Then, the first three-dimensional object detection unit 33 counts the offset amount obtained for each of the small areas DW t1 to DW tn and forms a histogram.
- FIG. 7 is a diagram illustrating an example of a histogram obtained by the first three-dimensional object detection unit 33.
- the offset amount which is the amount of movement that minimizes the error between each of the small areas DW t1 to DW tn and the differential waveform DW t ⁇ 1 one time before, has some variation.
- the first three-dimensional object detection unit 33 forms a histogram of offset amounts including variations, and calculates a movement distance from the histogram.
- the first three-dimensional object detection unit 33 calculates the movement distance of the three-dimensional object from the maximum value of the histogram. That is, in the example illustrated in FIG.
- the first three-dimensional object detection unit 33 calculates the offset amount indicating the maximum value of the histogram as the movement distance ⁇ * .
- the moving distance ⁇ * is a relative moving distance of the other vehicle V with respect to the host vehicle V. For this reason, when calculating the absolute movement distance, the first three-dimensional object detection unit 33 calculates the absolute movement distance based on the obtained movement distance ⁇ * and the signal from the vehicle speed sensor 20. .
- the first three-dimensional object detection unit 33 weights each of the plurality of small areas DW t1 to DW tn and forms an offset amount calculated for each of the small areas DW t1 to DW tn according to the weight.
- a histogram may be formed.
- FIG. 8 is a diagram illustrating weighting by the first three-dimensional object detection unit 33.
- the small area DW m (m is an integer of 1 to n ⁇ 1) is flat. That is, in the small area DW m , the difference between the maximum value and the minimum value of the number of pixels indicating a predetermined difference is small.
- the first three-dimensional object detection unit 33 to reduce the weight for such small area DW m. This is because the flat small area DW m has no characteristics and is likely to have a large error in calculating the offset amount.
- the small region DW m + k (k is an integer equal to or less than nm) is rich in undulations. That is, in the small area DW m , the difference between the maximum value and the minimum value of the number of pixels indicating a predetermined difference is large.
- the first three-dimensional object detection unit 33 increases the weight for such small area DW m. This is because the small region DW m + k rich in undulations is characteristic and there is a high possibility that the offset amount can be accurately calculated. By weighting in this way, the calculation accuracy of the movement distance can be improved.
- the differential waveform DW t is divided into a plurality of small areas DW t1 to DW tn in order to improve the calculation accuracy of the movement distance.
- the small area DW t1 is divided. It is not necessary to divide into ⁇ DW tn .
- the first three-dimensional object detection unit 33 calculates the movement distance from the offset amount of the differential waveform DW t when the error between the differential waveform DW t and the differential waveform DW t ⁇ 1 is minimized. That is, the method for obtaining the offset amount of the difference waveform DW t in the difference waveform DW t-1 and the current time before one unit time is not limited to the above disclosure.
- the computer 30 includes a smear detection unit 34.
- the smear detection unit 34 detects a smear occurrence region from data of a captured image obtained by imaging with the camera 10. Since smear is a whiteout phenomenon that occurs in a CCD image sensor or the like, the smear detector 34 may be omitted when the camera 10 using a CMOS image sensor or the like in which such smear does not occur.
- FIG. 9 is an image diagram for explaining the processing by the smear detection unit 34 and the calculation processing of the differential waveform DW t thereby.
- data of the captured image P in which the smear S exists is input to the smear detection unit 34.
- the smear detection unit 34 detects the smear S from the captured image P.
- There are various methods for detecting the smear S For example, in the case of a general CCD (Charge-Coupled Device) camera, the smear S is generated only in the downward direction of the image from the light source.
- CCD Charge-Coupled Device
- a region having a luminance value equal to or higher than a predetermined value from the lower side of the image to the upper side of the image and continuous in the vertical direction is searched, and this is identified as a smear S generation region.
- the smear detection unit 34 generates smear image SP data in which the pixel value is set to “1” for the place where the smear S is generated and the other place is set to “0”. After the generation, the smear detection unit 34 transmits the data of the smear image SP to the viewpoint conversion unit 31.
- the viewpoint conversion unit 31 to which the data of the smear image SP is input converts the viewpoint into a state of bird's-eye view.
- the viewpoint conversion unit 31 generates data of the smear bird's-eye view image SB t .
- the viewpoint conversion unit 31 transmits the data of the smear bird's-eye view image SB t to the alignment unit 32.
- the viewpoint conversion unit 31 transmits the data of the smear bird's-eye view image SB t ⁇ 1 one hour before to the alignment unit 32.
- the alignment unit 32 aligns the smear bird's-eye images SB t and SB t ⁇ 1 on the data.
- the specific alignment is the same as the case where the alignment of the bird's-eye images PB t and PB t ⁇ 1 is executed on the data.
- the alignment unit 32 performs a logical sum on the smear S generation region of each smear bird's-eye view image SB t , SB t ⁇ 1 . Thereby, the alignment part 32 produces
- the alignment unit 32 transmits the data of the mask image MP to the first three-dimensional object detection unit 33.
- the first three-dimensional object detection unit 33 sets the count number of the frequency distribution to zero for the portion corresponding to the smear S generation region in the mask image MP. That is, when the differential waveform DW t as shown in FIG. 9 is generated, the first three-dimensional object detection unit 33 sets the count number SC by the smear S to zero and generates a corrected differential waveform DW t ′. It will be.
- the first three-dimensional object detection unit 33 obtains the moving speed of the vehicle V (camera 10), and obtains the offset amount for the stationary object from the obtained moving speed. After obtaining the offset amount of the stationary object, the first three-dimensional object detection unit 33 ignores the offset amount corresponding to the stationary object among the maximum values of the histogram and calculates the moving distance of the three-dimensional object.
- FIG. 10 is a diagram illustrating another example of the histogram obtained by the first three-dimensional object detection unit 33.
- two maximum values ⁇ 1 and ⁇ 2 appear in the obtained histogram.
- one of the two maximum values ⁇ 1, ⁇ 2 is the offset amount of the stationary object.
- the first three-dimensional object detection unit 33 obtains the offset amount for the stationary object from the moving speed, ignores the maximum value corresponding to the offset amount, and adopts the remaining one maximum value to move the three-dimensional object movement distance. Is calculated.
- the first three-dimensional object detection unit 33 stops calculating the movement distance.
- 11 and 12 are flowcharts showing the three-dimensional object detection procedure of this embodiment.
- the computer 30 inputs data of an image P captured by the camera 10, and generates a smear image SP by the smear detection unit 34 (S ⁇ b> 1).
- the viewpoint conversion unit 31 generates data of the bird's-eye view image PB t from the data of the captured image P from the camera 10, and also generates data of the smear bird's-eye view image SB t from the data of the smear image SP (S2).
- the alignment unit 32 aligns the data of the bird's-eye view image PB t and the data of the bird's-eye view image PB t ⁇ 1 of the previous time, and the data of the smear bird's-eye view image SB t and the smear bird's-eye view of the previous time.
- the data of the image SB t-1 is aligned (S3).
- the alignment unit 32 generates data for the difference image PD t and also generates data for the mask image MP (S4).
- the first three-dimensional object detection unit 33, the data of the difference image PD t, and a one unit time before the difference image PD t-1 of the data generates a difference waveform DW t (S5).
- the first three-dimensional object detection unit 33 After generating the differential waveform DW t , the first three-dimensional object detection unit 33 sets the count number corresponding to the smear S generation region in the differential waveform DW t to zero, and suppresses the influence of the smear S (S6).
- the first three-dimensional object detection unit 33 determines whether or not the peak of the differential waveform DW t is greater than or equal to the first threshold value ⁇ (S7).
- the first threshold value ⁇ is set by the three-dimensional object determination unit 38 shown in FIG. 3, and details thereof will be described later.
- the peak of the difference waveform DW t is not equal to or greater than the first threshold value ⁇ , that is, when there is almost no difference, it is considered that there is no three-dimensional object in the captured image P.
- the first three-dimensional object detection unit 33 determines that there is no three-dimensional object and no other vehicle exists. (FIG. 12: S16). Then, the processes shown in FIGS. 11 and 12 are terminated.
- the first three-dimensional object detection unit 33 determines that a three-dimensional object exists, and sets the difference waveform DW t as the difference waveform DW t .
- the area is divided into a plurality of small areas DW t1 to DW tn (S8).
- the first three-dimensional object detection unit 33 performs weighting for each of the small areas DW t1 to DW tn (S9).
- the first three-dimensional object detection unit 33 calculates an offset amount for each of the small areas DW t1 to DW tn (S10), and generates a histogram with the weights added (S11).
- the first three-dimensional object detection unit 33 calculates a relative movement distance that is a movement distance of the three-dimensional object with respect to the host vehicle V based on the histogram (S12). Next, the first three-dimensional object detection unit 33 calculates the absolute movement speed of the three-dimensional object from the relative movement distance (S13). At this time, the first three-dimensional object detection unit 33 calculates the relative movement speed by differentiating the relative movement distance with respect to time, and calculates the absolute movement speed by adding the own vehicle speed detected by the vehicle speed sensor 20.
- the first three-dimensional object detection unit 33 determines whether the absolute movement speed of the three-dimensional object is 10 km / h or more and the relative movement speed of the three-dimensional object with respect to the host vehicle V is +60 km / h or less (S14). ). When both are satisfied (S14: YES), the first three-dimensional object detection unit 33 determines that the three-dimensional object is the other vehicle V (S15). Then, the processes shown in FIGS. 11 and 12 are terminated. On the other hand, when neither one is satisfied (S14: NO), the first three-dimensional object detection unit 33 determines that there is no other vehicle (S16). Then, the processes shown in FIGS. 11 and 12 are terminated.
- the rear side of the host vehicle V is set as the detection areas A1 and A2, and emphasis is placed on whether or not there is a possibility of contact when the host vehicle V changes lanes. For this reason, the process of step S14 is performed. That is, assuming that the system according to this embodiment is operated on a highway, if the speed of a three-dimensional object is less than 10 km / h, even if another vehicle V exists, Since it is located far behind the vehicle V, there are few problems. Similarly, when the relative moving speed of the three-dimensional object with respect to the own vehicle V exceeds +60 km / h (that is, when the three-dimensional object is moving at a speed higher than 60 km / h than the speed of the own vehicle V), the lane is changed. In some cases, since the vehicle is moving in front of the host vehicle V, there is little problem. For this reason, it can be said that the other vehicle V which becomes a problem at the time of lane change is judged in step S14.
- step S14 it is determined whether the absolute moving speed of the three-dimensional object is 10 km / h or more and the relative moving speed of the three-dimensional object with respect to the vehicle V is +60 km / h or less.
- the absolute moving speed of the stationary object may be detected to be several km / h. Therefore, the possibility of determining that the stationary object is the other vehicle V can be reduced by determining whether the speed is 10 km / h or more.
- the relative speed of the three-dimensional object with respect to the host vehicle V may be detected at a speed exceeding +60 km / h. Therefore, the possibility of erroneous detection due to noise can be reduced by determining whether the relative speed is +60 km / h or less.
- step S14 it may be determined that the absolute movement speed is not negative or not 0 km / h. Further, in the present embodiment, since emphasis is placed on whether or not there is a possibility of contact when the host vehicle V changes lanes, when the other vehicle V is detected in step S15, the driver of the host vehicle is notified. A warning sound may be emitted or a display corresponding to a warning may be performed by a predetermined display device.
- the number of pixels indicating a predetermined difference is counted on the data of the difference image PD t along the direction in which the three-dimensional object falls by viewpoint conversion.
- the difference waveform DW t is generated by frequency distribution.
- the pixel indicating the predetermined difference on the data of the difference image PD t is a pixel that has changed in an image at a different time, in other words, a place where a three-dimensional object exists.
- the difference waveform DW t is generated by counting the number of pixels along the direction in which the three-dimensional object collapses and performing frequency distribution at the location where the three-dimensional object exists.
- the differential waveform DW t is generated from the information in the height direction for the three-dimensional object. Then, the moving distance of the three-dimensional object is calculated from the time change of the differential waveform DW t including the information in the height direction. For this reason, compared with the case where only one point of movement is focused on, the detection location before the time change and the detection location after the time change are specified including information in the height direction. The same location is likely to be obtained, and the movement distance is calculated from the time change of the same location, so that the calculation accuracy of the movement distance can be improved.
- the count number of the frequency distribution is set to zero for the portion corresponding to the smear S generation region in the differential waveform DW t .
- the waveform portion generated by the smear S in the differential waveform DW t is removed, and a situation in which the smear S is mistaken as a three-dimensional object can be prevented.
- the moving distance of the three-dimensional object is calculated from the offset amount of the differential waveform DW t when the error of the differential waveform DW t generated at different times is minimized. For this reason, the movement distance is calculated from the offset amount of the one-dimensional information called the waveform, and the calculation cost can be suppressed in calculating the movement distance.
- the differential waveform DW t generated at different times is divided into a plurality of small regions DW t1 to DW tn .
- a plurality of waveforms representing respective portions of the three-dimensional object are obtained.
- weighting is performed for each of the plurality of small areas DW t1 to DW tn , and the offset amount obtained for each of the small areas DW t1 to DW tn is counted according to the weight to form a histogram. For this reason, the moving distance can be calculated more appropriately by increasing the weight for the characteristic area and decreasing the weight for the non-characteristic area. Therefore, the calculation accuracy of the moving distance can be further improved.
- the weight is increased as the difference between the maximum value and the minimum value of the number of pixels indicating a predetermined difference increases. For this reason, the characteristic undulation region having a large difference between the maximum value and the minimum value has a larger weight, and the flat region having a small undulation has a smaller weight.
- the moving distance is calculated by increasing the weight in the area where the difference between the maximum value and the minimum value is large. The accuracy can be further improved.
- the moving distance of the three-dimensional object is calculated from the maximum value of the histogram obtained by counting the offset amount obtained for each of the small areas DW t1 to DW tn . For this reason, even if there is a variation in the offset amount, a more accurate movement distance can be calculated from the maximum value.
- the offset amount for a stationary object is obtained and this offset amount is ignored, it is possible to prevent a situation in which the calculation accuracy of the moving distance of the three-dimensional object is lowered due to the stationary object.
- the calculation of the moving distance of the three-dimensional object is stopped. For this reason, it is possible to prevent a situation in which an erroneous movement distance having a plurality of maximum values is calculated.
- the vehicle speed of the host vehicle V is determined based on a signal from the speed sensor 20, but the present invention is not limited to this, and the speed may be estimated from a plurality of images at different times. In this case, a vehicle speed sensor becomes unnecessary, and the configuration can be simplified.
- the captured image at the current time and the image one hour before are converted into a bird's-eye view, the converted bird's-eye view is aligned, the difference image PD t is generated, and the generated difference image PD
- t is evaluated along the falling direction (the falling direction of the three-dimensional object when the captured image is converted into a bird's eye view)
- the differential waveform DW t is generated, but the present invention is not limited to this.
- the differential waveform DW t may be generated by evaluating along the direction corresponding to the falling direction (that is, the direction in which the falling direction is converted into the direction on the captured image).
- the difference image PD t is generated from the difference between the two images subjected to the alignment, and the difference image PD t is converted into a bird's eye view
- the bird's-eye view does not necessarily have to be clearly generated as long as the evaluation can be performed along the direction in which the user falls.
- a difference image between them is obtained. It generates a PD t, but generates a difference waveform DW t by evaluating along a direction corresponding to the direction collapsing the difference image PD t as shown in FIG. 5, the current time and one time before the bird's-eye image data PB t 1 and PB t ⁇ 1 are respectively evaluated along a direction corresponding to the falling direction as shown in FIG. 5 to generate a differential waveform DW t at the current time and one hour before, respectively.
- the waveforms may be aligned as shown in FIG. 4B, and final difference waveform information may be generated from the difference between these two difference waveforms.
- FIGS. 13A and 13B are diagrams illustrating an imaging range and the like of the camera 10 in FIG. 3.
- FIG. 13A is a plan view, and FIG. Show.
- the camera 10 has a predetermined angle of view a, and images the rear side from the host vehicle V1 included in the predetermined angle of view a.
- the angle of view a of the camera 10 is set so that the imaging range of the camera 10 includes the adjacent lane in addition to the lane in which the host vehicle V1 travels.
- the detection areas A1 and A2 in this example are trapezoidal in a plan view (when viewed from a bird's eye), and the positions, sizes, and shapes of the detection areas A1 and A2 are determined based on the distances d 1 to d 4. Is done.
- the detection areas A1 and A2 in the example shown in the figure are not limited to a trapezoidal shape, and may be other shapes such as a rectangle when viewed from a bird's eye view as shown in FIG.
- the distance d1 is a distance from the host vehicle V1 to the ground lines L1 and L2.
- the ground lines L1 and L2 mean lines on which a three-dimensional object existing in the lane adjacent to the lane in which the host vehicle V1 travels contacts the ground.
- an object is to detect other vehicles V2 and the like (including two-wheeled vehicles and the like) traveling in the left and right lanes adjacent to the lane of the host vehicle V1 on the rear side of the host vehicle V1.
- a distance d1 which is a position to be the ground lines L1 and L2 of the other vehicle V2 is calculated from a distance d11 from the own vehicle V1 to the white line W and a distance d12 from the white line W to a position where the other vehicle V2 is predicted to travel. It can be determined substantially fixedly.
- the distance d1 is not limited to being fixedly determined, and may be variable.
- the computer 30 recognizes the position of the white line W with respect to the host vehicle V1 by a technique such as white line recognition, and determines the distance d11 based on the recognized position of the white line W.
- the distance d1 is variably set using the determined distance d11.
- the distance d1 is It shall be fixedly determined.
- the distance d2 is a distance extending in the vehicle traveling direction from the rear end portion of the host vehicle V1.
- the distance d2 is determined so that the detection areas A1 and A2 are at least within the angle of view a of the camera 10.
- the distance d2 is set so as to be in contact with the range divided into the angle of view a.
- the distance d3 is a distance indicating the length of the detection areas A1, A2 in the vehicle traveling direction. This distance d3 is determined based on the size of the three-dimensional object to be detected. In the present embodiment, since the detection target is the other vehicle V2 or the like, the distance d3 is set to a length including the other vehicle V2.
- the distance d4 is a distance indicating a height set so as to include a tire such as the other vehicle V2 in the real space, as shown in FIG. 13B.
- the distance d4 is a length shown in FIG. 13A in the bird's-eye view image.
- the distance d4 may be a length that does not include a lane that is further adjacent to the left and right adjacent lanes in the bird's-eye view image (that is, a lane that is adjacent to two lanes). If the lane adjacent to the two lanes is included from the lane of the own vehicle V1, there is another vehicle V2 in the adjacent lane on the left and right of the own lane that is the lane in which the own vehicle V1 is traveling. This is because it becomes impossible to distinguish whether the other vehicle V2 exists.
- the distances d1 to d4 are determined, and thereby the positions, sizes, and shapes of the detection areas A1 and A2 are determined. More specifically, the position of the upper side b1 of the detection areas A1 and A2 forming a trapezoid is determined by the distance d1. The starting point position C1 of the upper side b1 is determined by the distance d2. The end point position C2 of the upper side b1 is determined by the distance d3. The side b2 of the detection areas A1 and A2 having a trapezoidal shape is determined by a straight line L3 extending from the camera 10 toward the starting point position C1.
- a side b3 of trapezoidal detection areas A1 and A2 is determined by a straight line L4 extending from the camera 10 toward the end position C2.
- the position of the lower side b4 of the detection areas A1 and A2 having a trapezoidal shape is determined by the distance d4.
- the areas surrounded by the sides b1 to b4 are set as the detection areas A1 and A2.
- the detection areas A1 and A2 are true squares (rectangles) in the real space on the rear side from the host vehicle V1.
- the viewpoint conversion unit 31 inputs captured image data of a predetermined area obtained by imaging with the camera 10.
- the viewpoint conversion unit 31 performs viewpoint conversion processing on the input captured image data to the bird's-eye image data in a bird's-eye view state.
- the bird's-eye view is a state seen from the viewpoint of a virtual camera looking down from above, for example, vertically downward (or slightly obliquely downward).
- This viewpoint conversion process can be realized by a technique described in, for example, Japanese Patent Application Laid-Open No. 2008-219063.
- the luminance difference calculation unit 35 calculates a luminance difference with respect to the bird's-eye view image data subjected to viewpoint conversion by the viewpoint conversion unit 31 in order to detect the edge of the three-dimensional object included in the bird's-eye view image. For each of a plurality of positions along a vertical imaginary line extending in the vertical direction in the real space, the brightness difference calculating unit 35 calculates a brightness difference between two pixels in the vicinity of each position.
- the luminance difference calculation unit 35 can calculate the luminance difference by either a method of setting only one vertical virtual line extending in the vertical direction in the real space or a method of setting two vertical virtual lines.
- the brightness difference calculation unit 35 applies a first vertical imaginary line corresponding to a line segment extending in the vertical direction in the real space and a vertical direction in the real space different from the first vertical imaginary line with respect to the bird's-eye view image that has undergone viewpoint conversion.
- a second vertical imaginary line corresponding to the extending line segment is set.
- the luminance difference calculation unit 35 continuously obtains a luminance difference between a point on the first vertical imaginary line and a point on the second vertical imaginary line along the first vertical imaginary line and the second vertical imaginary line.
- the operation of the luminance difference calculation unit 35 will be described in detail.
- the luminance difference calculation unit 35 corresponds to a line segment extending in the vertical direction in the real space and passes through the detection area A1 (hereinafter referred to as the attention line La).
- the luminance difference calculation unit 35 corresponds to a line segment extending in the vertical direction in the real space and also passes through the second vertical virtual line Lr (hereinafter referred to as a reference line Lr) passing through the detection area A1.
- the reference line Lr is set at a position separated from the attention line La by a predetermined distance in the real space.
- the line corresponding to the line segment extending in the vertical direction in the real space is a line that spreads radially from the position Ps of the camera 10 in the bird's-eye view image.
- This radially extending line is a line along the direction in which the three-dimensional object falls when converted to bird's-eye view.
- the luminance difference calculation unit 35 sets the attention point Pa (point on the first vertical imaginary line) on the attention line La.
- the luminance difference calculation unit 35 sets a reference point Pr (a point on the second vertical plate) on the reference line Lr.
- the attention line La, the attention point Pa, the reference line Lr, and the reference point Pr have the relationship shown in FIG. 14B in the real space.
- the attention line La and the reference line Lr are lines extending in the vertical direction in the real space, and the attention point Pa and the reference point Pr are substantially the same height in the real space. This is the point that is set.
- the attention point Pa and the reference point Pr do not necessarily have the same height, and an error that allows the attention point Pa and the reference point Pr to be regarded as the same height is allowed.
- the luminance difference calculation unit 35 obtains a luminance difference between the attention point Pa and the reference point Pr. If the luminance difference between the attention point Pa and the reference point Pr is large, it is considered that an edge exists between the attention point Pa and the reference point Pr. Therefore, the edge line detection unit 36 shown in FIG. 3 detects an edge line based on the luminance difference between the attention point Pa and the reference point Pr.
- FIG. 15 is a diagram illustrating a detailed operation of the luminance difference calculation unit 35, in which FIG. 15 (a) shows a bird's-eye view image in a bird's-eye view state, and FIG. 15 (b) is shown in FIG. 15 (a). It is the figure which expanded a part B1 of the bird's-eye view image. Although only the detection area A1 is illustrated and described in FIG. 15, the luminance difference is calculated in the same procedure for the detection area A2.
- the other vehicle V2 When the other vehicle V2 is reflected in the captured image captured by the camera 10, the other vehicle V2 appears in the detection area A1 in the bird's-eye view image as shown in FIG. As shown in the enlarged view of the area B1 in FIG. 15A in FIG. 15B, it is assumed that the attention line La is set on the rubber part of the tire of the other vehicle V2 on the bird's-eye view image.
- the luminance difference calculation unit 35 first sets the reference line Lr.
- the reference line Lr is set along the vertical direction at a position away from the attention line La by a predetermined distance in the real space.
- the reference line Lr is set at a position separated from the attention line La by 10 cm in real space.
- the reference line Lr is set on the wheel of the tire of the other vehicle V2 that is separated from the rubber of the tire of the other vehicle V2 by, for example, about 10 cm on the bird's eye view image.
- the luminance difference calculation unit 35 sets a plurality of attention points Pa1 to PaN on the attention line La.
- attention point Pai when an arbitrary point is indicated
- the number of attention points Pa set on the attention line La may be arbitrary.
- N attention points Pa are set on the attention line La.
- the luminance difference calculation unit 35 sets the reference points Pr1 to PrN so as to be the same height as the attention points Pa1 to PaN in the real space. Then, the luminance difference calculation unit 35 calculates the luminance difference between the attention point Pa and the reference point Pr having the same height. Thereby, the luminance difference calculation unit 35 calculates the luminance difference between the two pixels for each of a plurality of positions (1 to N) along the vertical imaginary line extending in the vertical direction in the real space. For example, the luminance difference calculating unit 35 calculates a luminance difference between the first attention point Pa1 and the first reference point Pr1, and the second difference between the second attention point Pa2 and the second reference point Pr2. Will be calculated.
- the luminance difference calculation unit 35 continuously calculates the luminance difference along the attention line La and the reference line Lr. That is, the luminance difference calculation unit 35 sequentially obtains the luminance difference between the third to Nth attention points Pa3 to PaN and the third to Nth reference points Pr3 to PrN.
- the luminance difference calculation unit 35 repeatedly executes the above-described processing such as setting the reference line Lr, setting the attention point Pa and the reference point Pr, and calculating the luminance difference while shifting the attention line La in the detection area A1. That is, the luminance difference calculation unit 35 repeatedly executes the above processing while changing the positions of the attention line La and the reference line Lr by the same distance in the extending direction of the ground line L1 in the real space. For example, the luminance difference calculation unit 35 sets the reference line Lr as the reference line Lr in the previous processing, sets the reference line Lr for the attention line La, and sequentially obtains the luminance difference. It will be.
- the edge line detection unit 36 detects an edge line from the continuous luminance difference calculated by the luminance difference calculation unit 35.
- the first attention point Pa ⁇ b> 1 and the first reference point Pr ⁇ b> 1 are located in the same tire portion, and thus the luminance difference is small.
- the second to sixth attention points Pa2 to Pa6 are located in the rubber part of the tire, and the second to sixth reference points Pr2 to Pr6 are located in the wheel part of the tire. Therefore, the luminance difference between the second to sixth attention points Pa2 to Pa6 and the second to sixth reference points Pr2 to Pr6 becomes large. Therefore, the edge line detection unit 36 may detect that an edge line exists between the second to sixth attention points Pa2 to Pa6 and the second to sixth reference points Pr2 to Pr6 having a large luminance difference. it can.
- the edge line detection unit 36 firstly follows the following Equation 1 to determine the i-th attention point Pai (coordinate (xi, yi)) and the i-th reference point Pri (coordinate ( xi ′, yi ′)) and the i th attention point Pai are attributed.
- Equation 1 t represents a threshold value
- I (xi, yi) represents the luminance value of the i-th attention point Pai
- I (xi ′, yi ′) represents the luminance value of the i-th reference point Pri.
- the attribute s (xi, yi) of the attention point Pai is “1”.
- the attribute s (xi, yi) of the attention point Pai is “ ⁇ 1”.
- the edge line detection unit 36 determines whether or not the attention line La is an edge line from the continuity c (xi, yi) of the attribute s along the attention line La based on Equation 2 below.
- the continuity c (xi, yi) is “1”.
- the attribute s (xi, yi) of the attention point Pai is not the same as the attribute s (xi + 1, yi + 1) of the adjacent attention point Pai + 1
- the continuity c (xi, yi) is “0”.
- the edge line detection unit 36 obtains the sum for the continuity c of all the points of interest Pa on the line of interest La.
- the edge line detection unit 36 normalizes the continuity c by dividing the obtained sum of continuity c by the number N of points of interest Pa.
- the edge line detection unit 36 determines that the attention line La is an edge line when the normalized value exceeds the threshold ⁇ .
- the threshold value ⁇ is a value set in advance through experiments or the like.
- the edge line detection unit 36 determines whether or not the attention line La is an edge line based on Equation 3 below. Then, the edge line detection unit 36 determines whether or not all the attention lines La drawn on the detection area A1 are edge lines.
- Equation 3 >> ⁇ c (xi, yi) / N> ⁇
- the second three-dimensional object detection unit 37 detects a three-dimensional object based on the amount of edge lines detected by the edge line detection unit 36.
- the three-dimensional object detection device 1 detects an edge line extending in the vertical direction in real space. The fact that many edge lines extending in the vertical direction are detected means that there is a high possibility that a three-dimensional object exists in the detection areas A1 and A2. For this reason, the second three-dimensional object detection unit 37 detects a three-dimensional object based on the amount of edge lines detected by the edge line detection unit 36. Furthermore, prior to detecting the three-dimensional object, the second three-dimensional object detection unit 37 determines whether or not the edge line detected by the edge line detection unit 36 is correct.
- the second three-dimensional object detection unit 37 determines whether or not the luminance change along the edge line of the bird's-eye view image on the edge line is larger than a predetermined threshold value. When the luminance change of the bird's-eye view image on the edge line is larger than the threshold value, it is determined that the edge line is detected by erroneous determination. On the other hand, when the luminance change of the bird's-eye view image on the edge line is not larger than the threshold value, it is determined that the edge line is correct.
- This threshold value is a value set in advance by experiments or the like.
- FIG. 16 is a diagram showing the luminance distribution of the edge line
- FIG. 16A shows the edge line and luminance distribution when another vehicle V2 as a three-dimensional object exists in the detection area A1
- FIG. Indicates an edge line and a luminance distribution when there is no solid object in the detection area A1.
- the attention line La set in the tire rubber part of the other vehicle V2 is determined to be an edge line in the bird's-eye view image.
- the luminance change of the bird's-eye view image on the attention line La is gentle. This is because the tire of the other vehicle V2 is extended in the bird's-eye view image by converting the image captured by the camera 10 into a bird's-eye view image.
- the attention line La set in the white character portion “50” drawn on the road surface in the bird's-eye view image is erroneously determined as an edge line.
- the brightness change of the bird's-eye view image on the attention line La has a large undulation. This is because a portion with high brightness in white characters and a portion with low brightness such as a road surface are mixed on the edge line.
- the second three-dimensional object detection unit 37 determines whether or not the edge line is detected by erroneous determination. When the luminance change along the edge line is larger than a predetermined threshold, the second three-dimensional object detection unit 37 determines that the edge line has been detected by erroneous determination. And the said edge line is not used for the detection of a solid object. Thereby, white characters such as “50” on the road surface, weeds on the road shoulder, and the like are determined as edge lines, and the detection accuracy of the three-dimensional object is prevented from being lowered.
- the second three-dimensional object detection unit 37 calculates the luminance change of the edge line according to any one of the following mathematical formulas 4 and 5.
- the luminance change of the edge line corresponds to the evaluation value in the vertical direction in the real space.
- Equation 4 evaluates the luminance distribution by the sum of the squares of the differences between the i-th luminance value I (xi, yi) on the attention line La and the adjacent i + 1-th luminance value I (xi + 1, yi + 1).
- Equation 5 evaluates the luminance distribution by the sum of the absolute values of the differences between the i-th luminance value I (xi, yi) on the attention line La and the adjacent i + 1-th luminance value I (xi + 1, yi + 1).
- the threshold value t2 is used to binarize the attribute b of the adjacent luminance value, and the binarized attribute b is summed for all the attention points Pa. Also good.
- the attribute b (xi, yi) of the attention point Pa (xi, yi) is “1”. Become. If the relationship is other than that, the attribute b (xi, yi) of the attention point Pai is '0'.
- This threshold value t2 is set in advance by an experiment or the like in order to determine that the attention line La is not on the same three-dimensional object. Then, the second three-dimensional object detection unit 37 sums up the attributes b for all the attention points Pa on the attention line La, obtains an evaluation value in the vertical equivalent direction, and determines whether the edge line is correct.
- 17 and 18 are flowcharts showing details of the three-dimensional object detection method according to the present embodiment.
- FIG. 17 and FIG. 18 for the sake of convenience, the processing for the detection area A1 will be described, but the same processing is executed for the detection area A2.
- step S21 the camera 10 images a predetermined area specified by the angle of view a and the attachment position.
- step S22 the viewpoint conversion unit 31 inputs the captured image data captured by the camera 10 in step S21, performs viewpoint conversion, and generates bird's-eye view image data.
- step S23 the luminance difference calculation unit 35 sets the attention line La on the detection area A1. At this time, the luminance difference calculation unit 35 sets a line corresponding to a line extending in the vertical direction in the real space as the attention line La.
- luminance difference calculation part 35 sets the reference line Lr on detection area
- step S25 the luminance difference calculation unit 35 sets a plurality of attention points Pa on the attention line La.
- the luminance difference calculation unit 35 sets the attention points Pa as many as not causing a problem at the time of edge detection by the edge line detection unit 36.
- step S26 the luminance difference calculation unit 35 sets the reference point Pr so that the attention point Pa and the reference point Pr are substantially the same height in the real space. Thereby, the attention point Pa and the reference point Pr are arranged in a substantially horizontal direction, and it becomes easy to detect an edge line extending in the vertical direction in the real space.
- step S27 the luminance difference calculation unit 35 calculates the luminance difference between the attention point Pa and the reference point Pr that have the same height in the real space.
- the edge line detection unit 36 calculates the attribute s of each attention point Pa in accordance with Equation 1 above.
- step S28 the edge line detection unit 36 calculates the continuity c of the attribute s of each attention point Pa in accordance with Equation 2 above.
- step S29 the edge line detection unit 36 determines whether or not the value obtained by normalizing the total sum of continuity c is greater than the threshold value ⁇ according to the above formula 3.
- the edge line detection unit 36 detects the attention line La as an edge line in step S30. Then, the process proceeds to step S31.
- the edge line detection unit 36 does not detect the attention line La as an edge line, and the process proceeds to step S31.
- step S31 the computer 30 determines whether or not the processing in steps S23 to S30 has been executed for all the attention lines La that can be set on the detection area A1. If it is determined that the above processing has not been performed for all the attention lines La (S31: NO), the processing returns to step S23, a new attention line La is set, and the processing up to step S31 is repeated. On the other hand, when it is determined that the above process has been performed for all the attention lines La (S31: YES), the process proceeds to step S32 in FIG.
- the second three-dimensional object detection unit 37 calculates a luminance change along the edge line for each edge line detected in step S30 of FIG.
- the second three-dimensional object detection unit 37 calculates the luminance change of the edge line according to any one of the above formulas 4, 5, and 6.
- Step S33 the second three-dimensional object detection unit 37 excludes edge lines whose luminance change is larger than a predetermined threshold from the edge lines. That is, it is determined that an edge line having a large luminance change is not a correct edge line, and the edge line is not used for detecting a three-dimensional object. As described above, this is to prevent characters on the road surface, roadside weeds, and the like included in the detection area A1 from being detected as edge lines. Therefore, the predetermined threshold value is a value set based on a luminance change generated by characters on the road surface, weeds on the road shoulder, or the like obtained in advance by experiments or the like.
- the second three-dimensional object detection unit 37 determines whether or not the amount of the edge line is equal to or greater than the second threshold value ⁇ .
- the second threshold value ⁇ is a value set in advance by experiments or the like, and is set by the three-dimensional object determination unit 38 described later. For example, when a four-wheeled vehicle is set as the three-dimensional object to be detected, the second threshold value ⁇ is set based on the number of edge lines of the four-wheeled vehicle that have appeared in the detection region A1 in advance through experiments or the like.
- the second three-dimensional object detection unit 37 detects that a three-dimensional object exists in the detection area A1 in step S35.
- the second three-dimensional object detection unit 37 determines that there is no three-dimensional object in the detection area A1. Thereafter, the processing illustrated in FIGS. 17 and 18 ends.
- the vertical direction in the real space with respect to the bird's-eye view image A vertical imaginary line is set as a line segment extending to. Then, for each of a plurality of positions along the vertical imaginary line, a luminance difference between two pixels in the vicinity of each position can be calculated, and the presence or absence of a three-dimensional object can be determined based on the continuity of the luminance difference.
- the attention line La corresponding to the line segment extending in the vertical direction in the real space and the reference line Lr different from the attention line La are set for the detection areas A1 and A2 in the bird's-eye view image. Then, a luminance difference between the attention point Pa on the attention line La and the reference point Pr on the reference line Lr is continuously obtained along the attention line La and the reference line La. In this way, the luminance difference between the attention line La and the reference line Lr is obtained by continuously obtaining the luminance difference between the points. In the case where the luminance difference between the attention line La and the reference line Lr is high, there is a high possibility that there is an edge of the three-dimensional object at the set position of the attention line La.
- a three-dimensional object can be detected based on a continuous luminance difference.
- the detection accuracy of a three-dimensional object can be improved.
- the luminance difference between two points of approximately the same height near the vertical imaginary line is obtained.
- the luminance difference is obtained from the attention point Pa on the attention line La and the reference point Pr on the reference line Lr, which are substantially the same height in the real space, and thus the luminance when there is an edge extending in the vertical direction. The difference can be detected clearly.
- FIG. 19 is a diagram illustrating an example of an image for explaining the processing of the edge line detection unit 36.
- 102 is an adjacent image.
- a region where the brightness of the first striped pattern 101 is high and a region where the brightness of the second striped pattern 102 is low are adjacent to each other, and a region where the brightness of the first striped pattern 101 is low and the second striped pattern 102. Is adjacent to a region with high brightness.
- the portion 103 located at the boundary between the first striped pattern 101 and the second striped pattern 102 tends not to be perceived as an edge depending on human senses.
- the edge line detection unit 36 determines the part 103 as an edge line only when there is continuity in the attribute of the luminance difference in addition to the luminance difference in the part 103, the edge line detection unit 36 An erroneous determination of recognizing a part 103 that is not recognized as an edge line as a sensation as an edge line can be suppressed, and edge detection according to a human sensation can be performed.
- the edge line detection unit 36 when the luminance change of the edge line detected by the edge line detection unit 36 is larger than a predetermined threshold value, it is determined that the edge line has been detected by erroneous determination.
- the captured image acquired by the camera 10 is converted into a bird's-eye view image, the three-dimensional object included in the captured image tends to appear in the bird's-eye view image in a stretched state.
- the luminance change of the bird's-eye view image in the stretched direction tends to be small.
- the bird's-eye view image includes a high luminance region such as a character portion and a low luminance region such as a road surface portion.
- the brightness change in the stretched direction tends to increase in the bird's-eye view image. Therefore, by determining the luminance change of the bird's-eye view image along the edge line as in this example, the edge line detected by the erroneous determination can be recognized, and the detection accuracy of the three-dimensional object can be improved.
- the three-dimensional object detection device 1 of this example uses the detection result by the first three-dimensional object detection unit 33 and the second three-dimensional object detection unit.
- the first three-dimensional object detection is performed according to the brightness of the detection areas A1 and A2 detected by the brightness detection unit 40 and the three-dimensional object determination unit 38 that finally determines whether or not the object is a three-dimensional object from the detection result 37.
- a threshold setting unit 39 that sets the first threshold value ⁇ of the unit 33 and the second threshold value ⁇ of the second three-dimensional object detection unit 37.
- the brightness detection unit 40 is a camera such as a shutter speed, an aperture value, and a gain value controlled by the camera 10 itself based on the light amounts of the detection areas A1 and A2 detected by the camera 10 or the light amounts detected by the camera 10.
- 10 control values are read from the camera 10 and output to the threshold setting unit 39.
- threshold setting control according to the brightness of the actual environment can be executed. Further, by detecting the control value of the camera 10 based on the amount of light, it is possible to set a threshold value according to the characteristics of the three-dimensional object detection in accordance with the actual captured image.
- the brightness detection unit 40 can calculate the altitude of the sun from the time and the longitude of the current position in place of the light amounts, control values, and the like of the detection areas A1 and A2, and use this as the brightness.
- the solar altitude as the brightness detection value
- the road surface detection situation becomes close to reality, and a threshold value can be set according to the characteristics of the three-dimensional object detection.
- the threshold setting unit 39 sets the weight of the detection result of the second three-dimensional object detection unit 37 to be larger as the brightness detected by the brightness detection unit 40 is darker, and the first three-dimensional object detection unit 33 as the brightness is brighter.
- the detection result weight is set small. That is, when the detection environment is dark, the weight of the detection result based on the edge information is increased, and when the detection environment is bright, the weight of the detection result based on the differential waveform information is increased.
- the ratio ⁇ / ⁇ between the first threshold value ⁇ and the second threshold value ⁇ is set to be smaller as the brightness detected by the brightness detection unit 40 is darker, and the ratio ⁇ / ⁇ is set as the brightness is brighter. Set larger.
- FIGS. 21 to 24 are control maps showing setting examples of the ratio ⁇ / ⁇ between the first threshold value ⁇ and the second threshold value ⁇ with respect to brightness.
- FIG. 21 is an example in which the ratio ⁇ / ⁇ is increased stepwise as the brightness becomes brighter.
- FIG. 22 is an example in which the ratio ⁇ / ⁇ is increased stepwise as the brightness is increased. This is an example in which hysteresis is set to prevent control hunting.
- FIG. 23 shows an example in which the ratio ⁇ / ⁇ is proportionally increased as the brightness becomes brighter.
- FIG. 24 similarly shows an example in which the ratio ⁇ / ⁇ is increased in proportion to the brighter brightness. However, this is an example in which hysteresis is set to prevent control hunting.
- FIG. 21 to 24 show examples of setting the ratio ⁇ / ⁇ between the first threshold value ⁇ and the second threshold value ⁇ , but the first threshold value ⁇ and the second threshold value ⁇ are independent of each other. It may be set.
- FIG. 25 is a control map in which the first threshold value ⁇ and the second threshold value ⁇ are set independently of each other, and the first threshold value ⁇ is decreased and the second threshold value ⁇ is increased as the brightness becomes brighter. In this case, only one of the first threshold value ⁇ and the second threshold value ⁇ may be controlled.
- the second threshold value ⁇ is set to a constant value, and the first threshold value ⁇ is decreased as the brightness increases.
- the setting value of the second threshold value ⁇ may be set larger than the detection upper limit value, and the detection of the three-dimensional object may be executed only by the first three-dimensional object detection unit 33.
- FIG. 26 is a control map showing a setting example of the weight X of the detection result by the first three-dimensional object detection unit 33 and the weight Y of the detection result by the second three-dimensional object detection unit 37.
- the weight X of the detection result by the three-dimensional object detection unit 33 is reduced, and the weight Y of the detection result by the second three-dimensional object detection unit 37 is set large.
- the detection result weight X by the first three-dimensional object detection unit 33 is set to be larger, and the detection result weight Y by the second three-dimensional object detection unit 37 is decreased.
- the weight Y is set to zero. That is, in this range, the three-dimensional object is detected only by the first three-dimensional object detection unit 33.
- the detection environment When the detection environment is dark, such as at night, the lighting condition of the road surface greatly changes due to the influence of street lights, headlights, etc., so there is a possibility that a three-dimensional object detection method using differential waveform information will make a false detection. Even if the light / dark fluctuation of the road surface due to such illumination occurs, an edge does not occur on the road surface, and thus the solid object detection method using the edge information has little influence. Therefore, when the detection environment is dark, the detection accuracy of the three-dimensional object can be increased by increasing the weight of the detection result based on the edge information. On the other hand, when the detection environment is bright as in the daytime, the road pattern and the edge of the object outside the road are clearly detected.
- the three-dimensional object detection method using the edge information is erroneously detected.
- such road surface patterns and edges of road objects have little influence on the three-dimensional object detection method using the differential waveform information. Therefore, when the detection environment is bright, the detection accuracy of the three-dimensional object can be increased by increasing the weight of the detection result based on the differential waveform information.
- step S 41 the brightness of the detection areas A 1 and A 2 is detected by the brightness detection unit 40 and output to the threshold setting unit 39.
- step S42 the threshold value setting unit 39 calculates the first threshold value ⁇ and the second threshold value ⁇ using the detected brightness and any one of the previously stored control maps shown in FIGS. It outputs to the object detection part 33 and the 2nd three-dimensional object detection part 37.
- FIG. The first threshold value ⁇ and the second threshold value ⁇ can be obtained, for example, by appropriately dividing the initial values ⁇ 0 and ⁇ 0 so that the ratio ⁇ / ⁇ obtained from the control maps shown in FIGS. .
- step S43 the three-dimensional object is detected based on the difference waveform information in the above-described procedure.
- step S44 the detection of the three-dimensional object based on the edge information is performed according to the procedure described above.
- the first threshold value ⁇ and the second threshold value ⁇ corresponding to the brightness are set to the first three-dimensional object detection unit 33 and the second three-dimensional object detection unit 37, respectively. Is set to
- step S45 it is detected whether it is a three-dimensional object in step S43, and it is determined whether it is detected as a three-dimensional object in step S44, and it is detected that it is a three-dimensional object in any step S43, S44. In that case, the process proceeds to step S46, and it is finally determined that the object is a three-dimensional object. If it is detected that the object is not a three-dimensional object in any of steps S43 and S44, the process proceeds to step S47, and it is finally determined that the object is not a three-dimensional object.
- the three-dimensional object detection device 1 of the present example when the detection environment is dark such as at night, a detection result based on differential waveform information that may be erroneously detected due to the influence of a streetlight, a headlight, or the like.
- the detection accuracy of the three-dimensional object can be increased.
- the detection environment is bright as in the daytime, the detection result weight by the differential waveform information is added to the detection result by the edge information that may be erroneously detected by the road pattern or the edge of the object outside the road. Therefore, the detection accuracy of the three-dimensional object can be increased.
- the differential waveform information is generated by the alignment unit 32 and the first three-dimensional object detection unit 33 in FIG. 3, in the above-described embodiment, based on the moving speed of the own vehicle, as shown in FIG.
- the bird's-eye view image and the bird's-eye view image one hour ago are aligned by shifting the position by the moving distance in the real space of the bird's-eye view image, the difference image in this state is obtained, and the difference waveform information is generated from this. It is also possible to use this method.
- the pixel value (edge amount) of the difference image of the captured image with different offset timing is compared with the pixel value (edge amount) of the difference image of the captured image with different timing not offset. It is determined whether the object is a stationary object or a moving object.
- a solid object image Q (T0) is detected in the detection areas A1 and A2 at the past timing T0, and at the current timing T1 after the timing of T0, the detection area A1.
- the subject vehicle V which is the detection subject, moves along the direction B, so that the three-dimensional object detected at the past timing T0 on the image.
- the image Q (T0) moves to the position of the image Q (T1) of the three-dimensional object on the upper side in the drawing of the detection areas A1 and A2.
- FIG. 27 A case where the three-dimensional object is a moving object will be described based on FIG. 28, and a case where the three-dimensional object is a stationary object will be described based on FIG.
- both the host vehicle V and the other vehicle VX move. There is a tendency to maintain a predetermined positional relationship. That is, when the captured image is offset, the position of the other vehicle VX tends to shift, and many pixels (edges) that can be characteristic are detected in the difference image PDt.
- FIG. 28B when the captured image is not offset, the positions of the host vehicle V and the other vehicle VX tend to approach each other, and the difference image PDt has pixels (edges) that can be characteristic. Less detected. If the number of pixels (edges) in the difference image PDt is large, the integrated value tends to be high. If the number of pixels (edges) in the difference image PDt is small, the integrated value in the difference waveform information tends to be low.
- the detected three-dimensional object is a stationary still object Q1
- the own vehicle V moves while the stationary object Q1 is stationary.
- the stationary object Q1 tend to be separated. That is, when the captured image is offset, the positions of the host vehicle V and the stationary object Q1 tend to approach, and a small number of pixels (edges) that can be characteristic are detected in the difference image PDt.
- FIG. 29B if the captured image is not offset, the position of the stationary object Q1 tends to be different from the previous captured image as the host vehicle V moves, and the difference image PDt is characteristic. Many possible pixels (edges) are detected.
- the integrated value in the luminance distribution information tends to be high, and if there are few pixels (edges) in the difference image PDt, the integrated value in the luminance distribution information tends to be low.
- the position of the first bird's-eye view image obtained at the first time T0 when the three-dimensional object is detected, and the position of the second bird's-eye view image obtained at the second time T1 after the first time. are obtained by performing frequency distribution by counting the number of pixels in which the brightness difference between adjacent image areas is equal to or greater than a predetermined threshold on the difference image of the aligned bird's-eye view images.
- a first integrated value of one luminance distribution information is obtained. That is, the offset difference image is generated in consideration of the movement amount of the host vehicle V.
- the offset amount d ′ corresponds to the movement amount on the bird's-eye view image data corresponding to the actual movement distance of the host vehicle V shown in FIG. 4A, and the signal from the vehicle speed sensor 20 and the current amount from one hour before. It is determined based on the time until the time.
- the first integrated value is the total of values plotted as the first luminance distribution information or a predetermined area.
- the first bird's-eye view image obtained at the first time T0 and the second bird's-eye view image obtained at the second time T1 after the first time T0 are obtained without shifting the positions.
- the second integrated value of the second luminance distribution information generated by counting the number of pixels in which the luminance difference between the adjacent image regions is equal to or greater than a predetermined threshold and performing frequency distribution is obtained. That is, a difference image that is not offset is generated, and its integrated value (second integrated value) is calculated.
- the second integrated value is all of the values plotted as the second luminance distribution information or the total value of the predetermined area.
- the evaluation value corresponding to the number of times that the second integrated value is determined to be greater than the first integrated value is equal to or greater than a predetermined evaluation threshold
- the three-dimensional object detected by the first three-dimensional object detection unit 33 is detected. Judged as “moving object”.
- the calculation method of the evaluation value is not limited, in this embodiment, every time it is determined that the second integrated value is larger than the first integrated value in the process repeatedly executed at a predetermined cycle, the evaluation point is counted up. The total value is obtained as an “evaluation value”.
- the pixel amount (edge amount) extracted from the difference image between the past captured image that has been offset and the current captured image, the past captured image that is not offset, and the current captured image Based on the magnitude relationship with the pixel amount (edge amount) extracted from the difference image with the captured image, the image transition feature of the moving object and the image transition feature of the stationary object are identified, and the three-dimensional object is the moving object. Whether the object is a stationary object or not can be determined with high accuracy.
- the second integrated value of the pixels (edge amount) indicating the predetermined difference in the difference image from the image not offset is the first pixel (edge amount) indicating the predetermined difference in the difference image from the offset image.
- the evaluation value is calculated by adding the first count value. That is, as the determination that the second integrated value is larger than the first integrated value is accumulated, the evaluation value is increased. If the evaluation value is equal to or greater than a predetermined evaluation threshold, it is determined that the three-dimensional object is a stationary object.
- the first count value is set higher as the number of consecutive determinations increases. As described above, when the determination that the second integrated value is larger than the first integrated value continues, it is determined that there is an increased possibility that the detected three-dimensional object is a stationary object, and the evaluation value becomes larger. Since the first count value is increased as described above, it is possible to determine with high accuracy whether or not the three-dimensional object is a moving object based on the continuous observation result.
- the first count value is added, and when it is determined that the second integrated value is smaller than the first integrated value, The evaluation value may be calculated by subtracting the second count value.
- the stationary object detection unit 38 determines that the second integrated value is smaller than the first integrated value after determining that the second integrated value is larger than the first integrated value. Further, after that, when it is determined that the second integrated value is larger than the first integrated value, the first count value is set high.
- the detected three-dimensional object is a stationary object. Since it is determined that there is a high possibility, and the first count value is increased so that the evaluation value is increased, it is possible to determine a stationary object with high accuracy based on the continuous observation result. Incidentally, the detection state of the feature of the moving object tends to be observed stably. If the detection result is unstable and the determination result that the three-dimensional object is a stationary object is discretely detected, it can be determined that the detected three-dimensional object is likely to be a stationary object. It is.
- the second count value is subtracted to calculate the evaluation value. In this case, if the determination that the second integrated value is smaller than the first integrated value continues for a predetermined number of times, the second count value is set high.
- the second integrated value is smaller than the first integrated value
- it is determined that the detected three-dimensional object is likely to be a moving object (another vehicle VX)
- a stationary object is determined.
- the second count value related to the subtraction is increased so that the evaluation value for performing the reduction becomes smaller, so that the stationary object can be determined with high accuracy based on the continuous observation result.
- the camera 10 corresponds to an imaging unit according to the present invention
- the viewpoint conversion unit 31 corresponds to an image conversion unit according to the present invention
- the alignment unit 32 and the first three-dimensional object detection unit 33 are the first according to the present invention.
- the three-dimensional object detection unit corresponds to the luminance difference calculation unit 35, the edge line detection unit 36, and the second three-dimensional object detection unit 37, and corresponds to the second three-dimensional object detection unit according to the present invention.
- the brightness detection unit 40 corresponds to a brightness detection unit according to the present invention
- the threshold setting unit 39 corresponds to a threshold setting unit according to the present invention.
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Multimedia (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Mechanical Engineering (AREA)
- Image Processing (AREA)
- Image Analysis (AREA)
- Traffic Control Systems (AREA)
- Closed-Circuit Television Systems (AREA)
Abstract
Description
視点変換部31は、カメラ10による撮像にて得られた所定領域の撮像画像データを入力し、入力した撮像画像データを鳥瞰視される状態の鳥瞰画像データに視点変換する。鳥瞰視される状態とは、上空から例えば鉛直下向きに見下ろす仮想カメラの視点から見た状態である。この視点変換は、例えば特開2008-219063号公報に記載されるようにして実行することができる。撮像画像データを鳥瞰視画像データに視点変換するのは、立体物に特有の鉛直エッジは鳥瞰視画像データへの視点変換により特定の定点を通る直線群に変換されるという原理に基づき、これを利用すれば平面物と立体物とを識別できるからである。なお、視点変換部31による画像変換処理の結果は、後述するエッジ情報による立体物の検出においても利用される。
次に、図3に示す視点変換部31、輝度差算出部35、エッジ線検出部36及び第2立体物検出部37で構成されるエッジ情報を利用した立体物の検出ブロックについて説明する。図13は、図3のカメラ10の撮像範囲等を示す図であり、図13(a)は平面図、図13(b)は、自車両V1から後側方における実空間上の斜視図を示す。図13(a)に示すように、カメラ10は所定の画角aとされ、この所定の画角aに含まれる自車両V1から後側方を撮像する。カメラ10の画角aは、図2に示す場合と同様に、カメラ10の撮像範囲に自車両V1が走行する車線に加えて、隣接する車線も含まれるように設定されている。
《数1》
I(xi,yi)>I(xi’,yi’)+tのとき
s(xi,yi)=1
I(xi,yi)<I(xi’,yi’)-tのとき
s(xi,yi)=-1
上記以外のとき
s(xi,yi)=0
《数2》
s(xi,yi)=s(xi+1,yi+1)のとき(且つ0=0を除く)、
c(xi,yi)=1
上記以外のとき、
c(xi,yi)=0
《数3》
Σc(xi,yi)/N>θ
《数4》
鉛直相当方向の評価値=Σ[{I(xi,yi)-I(xi+1,yi+1)}2]
《数5》
鉛直相当方向の評価値=Σ|I(xi,yi)-I(xi+1,yi+1)|
《数6》
鉛直相当方向の評価値=Σb(xi,yi)
但し、|I(xi,yi)-I(xi+1,yi+1)|>t2のとき、
b(xi,yi)=1
上記以外のとき、
b(xi,yi)=0
図3に戻り、上述した2つの立体物検出部33,37による立体物の検出にあたり、本例の立体物検出装置1は、第1立体物検出部33による検出結果と第2立体物検出部37の検出結果から立体物であるか否かを最終的に判断する立体物判断部38と、明るさ検出部40により検出された検出領域A1,A2の明るさに応じて第1立体物検出部33の第1閾値αと第2立体物検出部37の第2閾値βとを設定する閾値設定部39とを備える。
つまり、立体物が検出された第1の時刻T0において得られた第1鳥瞰視画像の位置と、第1の時刻の後の第2の時刻T1において得られた第2鳥瞰視画像の位置とを鳥瞰視上で位置合わせし、この位置合わせされた鳥瞰視画像の差分画像上において、互いに隣接する画像領域の輝度差が所定閾値以上である画素数をカウントして度数分布化して生成した第1輝度分布情報の第1積算値を求める。つまり、自車両Vの移動量を考慮して、オフセットした差分画像を生成する。オフセットする量d’は、図4(a)に示した自車両Vの実際の移動距離に対応する鳥瞰視画像データ上の移動量に対応し、車速センサ20からの信号と一時刻前から現時刻までの時間に基づいて決定される。第1積算値は、第1輝度分布情報としてプロットされた値の全部又は所定領域の合計値である。
10…カメラ
20…車速センサ
30…計算機
31…視点変換部
32…位置合わせ部
33…第1立体物検出部
34…スミア検出部
35…輝度差算出部
36…エッジ検出部
37…第2立体物検出部
38…立体物判断部
39…閾値設定部
40…明るさ検出部
a…画角
A1,A2…検出領域
CP…交点
DP…差分画素
DWt,DWt’…差分波形
DWt1~DWm,DWm+k~DWtn…小領域
L1,L2…接地線
La,Lb…立体物が倒れ込む方向上の線
P…撮像画像
PBt…鳥瞰視画像
PDt…差分画像
MP…マスク画像
S…スミア
SP…スミア画像
SBt…スミア鳥瞰視画像
V…自車両、他車両
Claims (9)
- 所定領域を撮像する撮像手段と、
前記撮像手段により得られた画像を鳥瞰視画像に視点変換する画像変換手段と、
前記所定領域の明るさを検出する明るさ検出手段と、
前記画像変換手段により得られた異なる時刻の鳥瞰視画像の差分画像から差分波形情報を生成し、少なくとも前記明るさ検出手段により検出された明るさが所定値以上の明るさである場合に、前記差分波形情報が第1閾値α以上であることで立体物を検出する第1立体物検出手段と、
前記画像変換手段により得られた鳥瞰視画像からエッジ情報を検出し、少なくとも前記明るさ検出手段により検出された明るさが所定値未満の明るさである場合に、前記エッジ情報が第2閾値β以上であることで立体物を検出する第2立体物検出手段と、を備える立体物検出装置。 - 前記第1立体物検出手段の検出結果と前記第2立体物検出手段の検出結果とから、立体物であるか否かを最終的に判断する立体物判断手段と、
前記明るさ検出手段により検出された明るさに応じて前記第1閾値及び前記第2閾値を設定する閾値設定手段と、を備え、
前記立体物判断手段は、前記差分波形情報が前記第1閾値α以上であって、前記エッジ情報が前記第2閾値β以上の場合に、立体物であると判断し、
前記閾値設定手段は、前記明るさ検出手段により検出された明るさが暗いほど、前記第1閾値αと前記第2閾値βとの比α/βを小さく設定し、前記明るさが明るいほど前記比α/βを大きく設定する請求項1に記載の立体物検出装置。 - 前記第1立体物検出手段は、前記画像変換手段により得られた異なる時刻の鳥瞰視画像の位置を鳥瞰視上で位置合わせし、鳥瞰視画像に視点変換した際に立体物が倒れ込む方向に沿って前記位置合わせされた鳥瞰視画像の差分画像上において所定の差分を示す画素数をカウントして度数分布化することで、一次元の差分波形情報を生成する請求項1又は2に記載の立体物検出装置。
- 前記第1立体物検出手段は、前記画像変換手段により得られた異なる時刻の鳥瞰視画像のそれぞれについて、鳥瞰視画像に視点変換した際に立体物が倒れ込む方向に沿う所定の差分を示す画素数をカウントして度数分布化することで、一次元の差分波形情報をそれぞれ生成し、これら差分波形情報を位置合わせし、前記位置合わせされた差分波形情報の差分から前記差分波形情報を生成する請求項1又は2に記載の立体物検出装置。
- 前記第1立体物検出手段は、前記撮像手段が装着された移動体の移動速度に基づいて、前記異なる時刻の鳥瞰視画像の実空間における移動距離を算出し、当該算出された移動距離だけ前記異なる時刻の鳥瞰視画像をずらして位置合わせし、前記差分波形情報を生成する請求項1~4のいずれか一項に記載の立体物検出装置。
- 前記第1立体物検出手段は、
前記撮像手段が装着された移動体の移動速度に基づいて、前記異なる時刻の鳥瞰視画像の実空間における移動距離を算出し、当該算出された移動距離だけ前記異なる時刻の鳥瞰視画像をずらして位置合わせし、当該位置合わせされた鳥瞰視画像の差分画像上において、所定の差分を示す画素数をカウントして度数分布化して生成した第1差分波形情報の第1積算値を求めるとともに、
前記異なる時刻の鳥瞰視画像の位置をずらさずにこれらの差分画像を求め、当該差分画像上において所定の差分を示す画素数をカウントして度数分布化して生成した第2差分波形情報の第2積算値を求め、
前記第2積算値が前記第1積算値よりも大きいと判断された回数に応じた評価値が所定の評価閾値以上であるか否かによって前記立体物を検出する請求項1~4のいずれか一項に記載の立体物検出装置。 - 前記第2立体物検出手段は、鳥瞰視画像に視点変換した際に立体物が倒れ込む方向に沿って、前記エッジ情報を検出する請求項1~6のいずれか一項に記載の立体物検出装置。
- 前記明るさ検出手段は、前記撮像手段が検知する前記所定領域の光量又はこれに基づく撮像手段の制御値を検出する請求項1~7のいずれか一項に記載の立体物検出装置。
- 前記明るさ検出手段は、太陽の位置に基づく明るさの特性値を検出する請求項1~7のいずれか一項に記載の立体物検出装置。
Priority Applications (4)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
EP12860558.1A EP2797320B1 (en) | 2011-12-19 | 2012-07-27 | Object detection device |
US14/363,882 US9740943B2 (en) | 2011-12-19 | 2012-07-27 | Three-dimensional object detection device |
CN201280063075.3A CN104012081B (zh) | 2011-12-19 | 2012-07-27 | 三维物体检测装置 |
JP2013550148A JP5776795B2 (ja) | 2011-12-19 | 2012-07-27 | 立体物検出装置 |
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
JP2011-276683 | 2011-12-19 | ||
JP2011276683 | 2011-12-19 |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2013094242A1 true WO2013094242A1 (ja) | 2013-06-27 |
Family
ID=48668151
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/JP2012/069095 WO2013094242A1 (ja) | 2011-12-19 | 2012-07-27 | 立体物検出装置 |
Country Status (5)
Country | Link |
---|---|
US (1) | US9740943B2 (ja) |
EP (1) | EP2797320B1 (ja) |
JP (1) | JP5776795B2 (ja) |
CN (1) | CN104012081B (ja) |
WO (1) | WO2013094242A1 (ja) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
EP3009983A1 (en) * | 2014-10-13 | 2016-04-20 | Conti Temic microelectronic GmbH | Obstacle detection apparatus and method |
CN115327529A (zh) * | 2022-09-05 | 2022-11-11 | 中国科学技术大学 | 一种融合毫米波雷达和激光雷达的3d目标检测与追踪方法 |
Families Citing this family (13)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2014061123A1 (ja) * | 2012-10-17 | 2014-04-24 | 富士通株式会社 | 画像処理装置、画像処理プログラムおよび画像処理方法 |
TW201441581A (zh) * | 2013-04-29 | 2014-11-01 | Hon Hai Prec Ind Co Ltd | 車輛輔助系統及車輛輔助方法 |
JP2015170174A (ja) * | 2014-03-07 | 2015-09-28 | ソニー株式会社 | 情報処理装置、情報処理システム、情報処理方法及びプログラム |
KR101637716B1 (ko) * | 2014-11-03 | 2016-07-07 | 현대자동차주식회사 | 차량의 장애물 위치 인식 장치 및 방법 |
DE102014019420A1 (de) * | 2014-12-22 | 2016-06-23 | GM Global Technology Operations LLC (n. d. Ges. d. Staates Delaware) | Kraftfahrzeugscheinwerfersystem, Kraftfahrzeug, Verfahren zum Betrieb eines Kraftfahrzeugscheinwerfersystems sowie Computerprogrammprodukt |
FR3047932B1 (fr) * | 2016-02-19 | 2018-03-16 | Peugeot Citroen Automobiles Sa | Dispositif et procede d'estimation du niveau d'attention d'un conducteur d'un vehicule |
DE102016109027A1 (de) * | 2016-05-17 | 2017-11-23 | Dr. Ing. H.C. F. Porsche Aktiengesellschaft | Verfahren zur Lageüberprüfung charakteristischer Punkte in Lichtverteilungen |
JP6782433B2 (ja) * | 2017-03-22 | 2020-11-11 | パナソニックIpマネジメント株式会社 | 画像認識装置 |
CN110793564A (zh) * | 2018-08-02 | 2020-02-14 | 昆山博威泰克电子科技有限公司 | 视觉检测设备和视觉检测方法 |
WO2020170835A1 (ja) * | 2019-02-18 | 2020-08-27 | ソニー株式会社 | 情報処理装置、情報処理方法及び情報処理プログラム |
JP7327992B2 (ja) * | 2019-05-10 | 2023-08-16 | フォルシアクラリオン・エレクトロニクス株式会社 | 立体物検出装置、及び立体物検出方法 |
CN111709923B (zh) * | 2020-06-10 | 2023-08-04 | 中国第一汽车股份有限公司 | 一种三维物体检测方法、装置、计算机设备和存储介质 |
EP3968274A1 (en) * | 2020-09-14 | 2022-03-16 | Tata Consultancy Services Limited | Method and system for asset inspection using unmanned aerial vehicles |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2008219063A (ja) | 2007-02-28 | 2008-09-18 | Sanyo Electric Co Ltd | 車両周辺監視装置及び方法 |
JP2008229063A (ja) | 2007-03-20 | 2008-10-02 | Daiichi Shokai Co Ltd | 遊技機 |
JP2009265783A (ja) * | 2008-04-23 | 2009-11-12 | Sanyo Electric Co Ltd | 運転支援システム及び車両 |
JP2010226449A (ja) * | 2009-03-24 | 2010-10-07 | Aisin Seiki Co Ltd | 障害物検出装置 |
Family Cites Families (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
KR100866450B1 (ko) * | 2001-10-15 | 2008-10-31 | 파나소닉 주식회사 | 차량 주위 감시 장치 및 그 조정 방법 |
JP2006339960A (ja) * | 2005-06-01 | 2006-12-14 | Nissan Motor Co Ltd | 物体検出装置、および物体検出方法 |
JP2009129001A (ja) * | 2007-11-20 | 2009-06-11 | Sanyo Electric Co Ltd | 運転支援システム、車両、立体物領域推定方法 |
JP5556077B2 (ja) * | 2009-07-28 | 2014-07-23 | 日産自動車株式会社 | 走行支援装置 |
JP4970516B2 (ja) * | 2009-09-30 | 2012-07-11 | 日立オートモティブシステムズ株式会社 | 周囲確認支援装置 |
JP5251927B2 (ja) * | 2010-06-21 | 2013-07-31 | 日産自動車株式会社 | 移動距離検出装置及び移動距離検出方法 |
JP5809785B2 (ja) * | 2010-07-30 | 2015-11-11 | 日立オートモティブシステムズ株式会社 | 車両用外界認識装置およびそれを用いた配光制御システム |
-
2012
- 2012-07-27 US US14/363,882 patent/US9740943B2/en active Active
- 2012-07-27 EP EP12860558.1A patent/EP2797320B1/en active Active
- 2012-07-27 WO PCT/JP2012/069095 patent/WO2013094242A1/ja active Application Filing
- 2012-07-27 CN CN201280063075.3A patent/CN104012081B/zh active Active
- 2012-07-27 JP JP2013550148A patent/JP5776795B2/ja active Active
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2008219063A (ja) | 2007-02-28 | 2008-09-18 | Sanyo Electric Co Ltd | 車両周辺監視装置及び方法 |
JP2008229063A (ja) | 2007-03-20 | 2008-10-02 | Daiichi Shokai Co Ltd | 遊技機 |
JP2009265783A (ja) * | 2008-04-23 | 2009-11-12 | Sanyo Electric Co Ltd | 運転支援システム及び車両 |
JP2010226449A (ja) * | 2009-03-24 | 2010-10-07 | Aisin Seiki Co Ltd | 障害物検出装置 |
Non-Patent Citations (1)
Title |
---|
See also references of EP2797320A4 |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
EP3009983A1 (en) * | 2014-10-13 | 2016-04-20 | Conti Temic microelectronic GmbH | Obstacle detection apparatus and method |
WO2016058893A1 (en) * | 2014-10-13 | 2016-04-21 | Conti Temic Microelectronic Gmbh | Obstacle detection apparatus and method |
US10417507B2 (en) | 2014-10-13 | 2019-09-17 | Conti Temic Microelectronic Gmbh | Freespace detection apparatus and freespace detection method |
CN115327529A (zh) * | 2022-09-05 | 2022-11-11 | 中国科学技术大学 | 一种融合毫米波雷达和激光雷达的3d目标检测与追踪方法 |
Also Published As
Publication number | Publication date |
---|---|
CN104012081B (zh) | 2017-04-19 |
EP2797320A1 (en) | 2014-10-29 |
EP2797320B1 (en) | 2016-09-14 |
US9740943B2 (en) | 2017-08-22 |
US20140368656A1 (en) | 2014-12-18 |
EP2797320A4 (en) | 2015-05-27 |
CN104012081A (zh) | 2014-08-27 |
JPWO2013094242A1 (ja) | 2015-04-27 |
JP5776795B2 (ja) | 2015-09-09 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
JP5776795B2 (ja) | 立体物検出装置 | |
JP5997276B2 (ja) | 立体物検出装置及び異物検出装置 | |
JP5804180B2 (ja) | 立体物検出装置 | |
WO2013157301A1 (ja) | 立体物検出装置及び立体物検出方法 | |
JP5874831B2 (ja) | 立体物検出装置 | |
JP5682735B2 (ja) | 立体物検出装置 | |
WO2014017600A1 (ja) | 立体物検出装置および立体物検出方法 | |
JP5743020B2 (ja) | 立体物検出装置 | |
JP5794378B2 (ja) | 立体物検出装置及び立体物検出方法 | |
JP5783319B2 (ja) | 立体物検出装置及び立体物検出方法 | |
JP5835459B2 (ja) | 立体物検出装置 | |
JP5871069B2 (ja) | 立体物検出装置及び立体物検出方法 | |
JP5794379B2 (ja) | 立体物検出装置及び立体物検出方法 | |
JP6003987B2 (ja) | 立体物検出装置及び立体物検出方法 | |
JP5768927B2 (ja) | 立体物検出装置 | |
WO2014017602A1 (ja) | 立体物検出装置および立体物検出方法 | |
JP5817913B2 (ja) | 立体物検出装置及び立体物検出方法 | |
JP5668891B2 (ja) | 立体物検出装置 |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 12860558 Country of ref document: EP Kind code of ref document: A1 |
|
ENP | Entry into the national phase |
Ref document number: 2013550148 Country of ref document: JP Kind code of ref document: A |
|
WWE | Wipo information: entry into national phase |
Ref document number: 14363882 Country of ref document: US |
|
NENP | Non-entry into the national phase |
Ref country code: DE |
|
REEP | Request for entry into the european phase |
Ref document number: 2012860558 Country of ref document: EP |
|
WWE | Wipo information: entry into national phase |
Ref document number: 2012860558 Country of ref document: EP |