JP2020020612A - Distance measuring device, method for measuring distance, program, and mobile body - Google Patents

Abstract

To provide a distance measuring device which can increase the accuracy of detecting an object when an undershoot is generated in a reflection signal from an object.SOLUTION: The present invention includes: an irradiation unit for emitting a laser beam; a light reception unit for receiving a reflected light as a laser beam reflected by an object; a first processing circuit (a signal amplifier 44b of a processing system 2) for performing undershoot processing on a reflection signal received by the light reception unit; and a filter circuit 44d for performing processing of highlighting a peak showing the object in the reflection signal on which the undershoot processing has been performed.SELECTED DRAWING: Figure 28

Description

  The present invention relates to a distance measuring device, a distance measuring method, a program, and a moving object.

  A parallax calculation algorithm for calculating a parallax value for each pixel in a stereo camera is known. The disparity calculation algorithm is a method of voting a cost in a search disparity space, obtaining a disparity value that gives the minimum cost as an integer disparity value d, estimating a sub-pixel disparity from the integer disparity value d, and then calculating the integer disparity value d and the sub-pixel. The disparity is obtained as a disparity estimated value D. Then, a distance corresponding to each pixel is calculated by a corresponding expression (Z = BF / D) between the distance Z and the estimated parallax value D (B is a base line length, and F is a focal length).

  In such a disparity space cost voting method, it is difficult to secure a distance resolution in a distant region where the integer disparity value d is small (the distance Z is large), and the dispersion (dispersion) of the disparity calculation result is large in the dispersion of the distance. There is a problem of affecting. Further, in the industry using stereo cameras (for example, in the vehicle industry), it is expected that the distance measurement performance at a long distance is improved, and that the stereo camera is reduced in cost, miniaturized, and improved in environmental robust performance.

  In order to compensate for the problem of the stereo camera, there is an attempt to use LiDAR (Light Detection and Ranging, Laser Imaging Detection and Ranging) having a low spatial resolution but a high distance resolution. LiDAR is a measurement method that calculates the distance to an object by irradiating the object with a laser beam and returning a reflected signal. As an example of a method of specifying a reflection signal of an object from a time-series reflection signal, there is a method of detecting a peak of a reflection signal. By merging the measurement results of a stereo camera with a high spatial resolution but a low distance resolution at a long distance and a LiDAR with a low spatial resolution but a high distance resolution (this may be called fusion), the advantages of both can be utilized. .

  A technology for improving the ranging accuracy in LiDAR is known (for example, see Patent Document 1). Patent Literature 1 discloses a method in which undershoot is intentionally generated in a reflected signal of a laser beam from an object to sharpen the fall of the object reflected signal, thereby improving the ranging accuracy of LiDAR. .

  However, there is a problem that the peak of the reflected signal cannot always be detected simply by generating the undershoot in the reflected signal from the object. For example, the peak of the reflected signal may not be detected unless appropriate processing is performed on the undershooted reflected signal among the reflected signals.

  The present invention has been made in view of the above circumstances, and has as its object to provide a distance measuring device that improves the detection accuracy of an object when an undershoot occurs in a reflected signal from the object.

  In view of the above problems, the present invention provides an irradiation unit that irradiates a laser beam, a light receiving unit that receives a reflected signal of the laser light reflected by an object, and a process of performing an undershoot process on the reflected signal received by the light receiving unit. 1. A distance measuring apparatus, comprising: a first processing circuit; and a filter circuit that performs a process of enhancing a peak representing the object on the reflected signal on which the undershoot process has been performed by the first processing circuit. I will provide a.

  It is possible to provide a distance measuring device that improves the detection accuracy of an object when an undershoot occurs in a reflected signal from the object.

It is an example of a figure explaining the appearance composition and example of attachment of a distance measuring device. FIG. 2 is a diagram illustrating an example of an overall configuration diagram of a distance measuring device. FIG. 3 is a diagram illustrating an example of a hardware configuration of a distance measuring device. FIG. 9 is a diagram illustrating a principle of deriving a parallax value for an object from a comparison image and a reference image by triangulation, and measuring a distance from the ranging device to the object based on the parallax value. It is a figure which shows the correspondence of a parallax value and distance in a graph format. FIG. 4 is an example of a diagram illustrating calculation of integer parallax by block matching. It is an example of a figure showing typically measurement of the time until an irradiated laser beam is reflected by an object and received. FIG. 3 is an example of a functional block diagram showing the function of a laser signal processing unit in a block shape. FIG. 3 is an example of a functional block diagram illustrating functions of a stereo image calculation unit in a block shape. It is a figure explaining an example of the difference of the signal level received near and far. FIG. 3 is an example of a functional block diagram for explaining the function of a stereo image calculation unit in detail. FIG. 4 is an example of a diagram showing details of a functional configuration of a first combined cost calculation unit. It is a figure showing r direction in calculation of the 1st course cost Lr (p, d). FIG. 9 is an example of a diagram for describing a method of determining a processing range by a range determining unit. It is an example of a figure showing the processing range decided by the range decision part. It is an example of a figure showing the details of the functional composition of the 2nd cost calculation part. FIG. 9 is an example of a diagram showing details of a functional configuration of a second combined cost calculation unit. It is an example of a figure showing the calculation result of the 2nd synthesis cost S 'in a reference pixel area. FIG. 3 is a diagram illustrating an example of image data of a distant image and a reflection signal from such a distant object. FIG. 3 is a diagram illustrating a configuration example of a PD output detection unit that emphasizes a reflection signal from an object. FIG. 6 is a diagram illustrating an example of a light reception signal waveform according to a comparative example and the present embodiment. FIG. 3 is a diagram illustrating an example of a negative feedback circuit. FIG. 7 is a diagram illustrating an example of a reflected signal that has been subjected to undershoot processing. It is an example of a figure explaining the reflection signal from the object (black curtain) with low reflectance of laser light. FIG. 3 is a diagram illustrating a configuration example of a PD output detection unit that emphasizes a reflection signal from an object. It is a figure showing an example of a row subject. FIG. 8 is a diagram illustrating an example of a reflected signal of a row subject subjected to undershoot processing. FIG. 3 is a configuration diagram of an example of a PD output detection unit. 9 is a flowchart illustrating a flow of a parallax image generation process performed by a stereo image calculation unit and a laser signal processing unit. FIG. 3 is an example of a diagram illustrating global analysis of a reflected signal converted by an A / D conversion circuit. FIG. 7 is an example of a diagram illustrating a method of calculating parallax based on a metric space. FIG. 7 is an example of a diagram for explaining a method for simply calculating parallax based on a metric space.

  Hereinafter, as an example of an embodiment for carrying out the present invention, a distance measuring device and a distance measuring method performed by the distance measuring device will be described with reference to the drawings.

<About terms>
Undershoot means that when a signal that has become larger than before detection due to detection of some detection target returns to a value before detection, the signal becomes smaller than the value before detection. If the value before detection is zero, it becomes a negative value due to undershoot. It may be called inversion processing.

  The distance measuring device may be referred to as a distance measuring system because it has a laser radar distance measuring unit and a stereo camera unit to be described later. In addition, it may be called a distance measuring device, a distance measuring unit, or the like.

<Appearance configuration and mounting example of distance measuring device>
First, an external configuration and an example of attachment of the distance measuring device will be described with reference to FIG. FIG. 1 is an example of a diagram for explaining an external configuration and an example of attachment of a distance measuring device.

  The distance measuring apparatus 100 includes a stereo camera unit 110 and a laser radar distance measuring unit 120. The stereo camera unit 110 includes a right camera (first imaging device) 111 and a left camera (second imaging device) 112, and the laser radar ranging unit 120 is provided between the right camera 112 and the left camera 111. Be placed. By integrating (fusion) the measurement results of the laser radar distance measurement unit 120 and the stereo camera unit 110, it becomes possible to acquire three-dimensional information of the surrounding environment with higher accuracy.

  The right camera 112 and the left camera 111 synchronize with each other at a predetermined frame period, perform image capturing, and generate a captured image.

  The laser radar distance measurement unit 120 measures the distance to the irradiation position (object in the irradiation direction) by irradiating the laser beam and receiving the reflected light, using the TOF (Time Of Flight) method.

  The distance measuring apparatus 100 is attached, for example, to a central portion of the vehicle 140 from the ceiling inside the front window. At this time, the stereo camera unit 110 and the laser radar distance measuring unit 120 are both mounted toward the front of the vehicle 140. That is, in the vehicle 140, the distance measuring apparatus 100 is mounted such that the optical axis of the stereo camera unit 110 and the center of the laser beam irradiation direction of the laser radar distance measuring unit 120 are in the same direction.

  The mounting position in FIG. 1 is merely an example, and may be mounted on a dashboard, a roof, a bumper, or the like of a vehicle. The attachment position in FIG. 1 is for acquiring three-dimensional information in front of the vehicle, and may be attached so that three-dimensional information on the right side, left side, or back of the vehicle can be acquired.

<Hardware configuration of distance measuring device>
Subsequently, an overall configuration example of the distance measuring apparatus 100 will be described with reference to FIG. FIG. 2 is a diagram illustrating an example of an overall configuration diagram of the distance measuring apparatus 100.

  The ranging device 100 is configured so that the laser radar ranging unit 120 and the stereo camera unit 110 can mutually transmit and receive necessary information. The stereo camera unit 110 has the stereo camera unit 110 that processes the reference image and the comparison image and outputs a distance image in addition to the right camera and the left camera as described above.

  The laser radar ranging unit 120 acquires a processing range from the stereo camera unit 110 as described later, narrows down objects in the processing range, and outputs distance information for each irradiation direction obtained by the narrowing down to the stereo camera unit 110. The stereo camera unit 110 generates a detailed distance image using the distance information for each irradiation direction, and outputs it to the ECU 190 (Electronic Control Unit). As described above, the fusion of the laser radar distance measurement unit 120 and the stereo camera unit 110 enables acquisition of three-dimensional information with higher accuracy.

  In FIG. 2, as an example, the distance image and the reference image are transmitted to the ECU 190. The ECU 190 is an electronic control unit of the vehicle. Note that the ranging device 100 mounted on a vehicle is referred to as a vehicle-mounted device. The ECU 190 performs various driving supports using the distance image and the reference image output from the distance measuring device 100. For the reference image, various patterns are matched to recognize the state of the preceding vehicle, pedestrian, white line, traffic light, and the like.

  Driving support varies depending on the vehicle. For example, when the lateral position of the target object overlaps with the width of the own vehicle, a warning or braking is performed according to TTC (Time To Collision) calculated from the distance and the relative speed. . When it is difficult to stop the vehicle until the collision, the steering is steered in a direction to avoid the collision.

  Further, the ECU 190 performs inter-vehicle distance control at all vehicle speeds following the preceding vehicle at an inter-vehicle distance corresponding to the vehicle speed. When the preceding vehicle stops, the own vehicle also stops, and when the preceding vehicle starts, the own vehicle also starts. Further, when the ECU 190 performs white line recognition or the like, lane keeping control for steering so as to travel in the center of the traveling lane or departure prevention control for changing the traveling direction toward the traveling lane if there is a possibility of departure from the traveling lane. It can be carried out.

  Also, when there is an obstacle in the traveling direction when the vehicle is stopped, sudden start can be suppressed. For example, when there is an obstacle in the traveling direction determined by the operation position of the shift lever and the operation amount of the accelerator pedal is large, the damage is reduced by limiting the engine output or issuing a warning.

  FIG. 3 is a diagram illustrating an example of a hardware configuration of the distance measuring apparatus 100. The distance measuring apparatus 100 includes a sensor stay 201 and a control board storage unit 202. A left camera 111, a right camera 112, and a laser radar distance measuring unit 120 are attached to the sensor stay 201. By arranging the laser radar ranging unit 120 on these straight lines sandwiched between the left camera 111 and the right camera 112, the size and cost of the ranging device 100 are reduced. The distance between the left camera 111 and the right camera 112 is referred to as a base length. The longer the base length, the easier it is to obtain parallax. In order to reduce the size of the distance measuring apparatus 100, it is necessary to shorten the base line length, and it is required that the base line length be kept short while maintaining the accuracy.

  The control board storage unit 202 stores a laser signal processing unit 240, a stereo image calculation unit 250, a memory 260, and an MPU (Micro Processing Unit) 270. By configuring the laser signal processing unit 240 separately from the laser radar distance measuring unit 120, the size of the laser radar distance measuring unit 120 can be reduced. Thus, in the present embodiment, the arrangement of the laser radar distance measurement unit 120 between the left camera 111 and the right camera 112 is realized.

  In the example of FIG. 3, the laser signal processing unit 240 and the stereo image calculation unit 250 are configured as separate circuit boards, but the laser signal processing unit 240 and the stereo image calculation unit 250 share a common circuit. It may be constituted by a substrate. This is because the cost can be reduced by reducing the number of circuit boards.

  Subsequently, each unit on the sensor stay 201 side will be described. As shown in FIG. 3, the left camera 111 includes a camera lens 211, an image sensor 212, and a sensor substrate 213. External light incident through the camera lens 211 is received by the image sensor 212 and is photoelectrically converted at a predetermined frame cycle. The signal obtained by the photoelectric conversion is processed in the sensor substrate 213, and a captured image for each frame is generated. The generated captured images are sequentially transmitted to the stereo image calculation unit 250 as comparison images.

  Note that the right camera 112 has the same configuration as the left camera 111, and captures an image in synchronization with the left camera 111 based on a synchronization control signal. The captured images are sequentially transmitted to the stereo image calculation unit 250 as reference images.

  The laser radar distance measurement unit 120 includes a light source drive circuit 231, a laser light source 232, and an irradiation lens 233. The light source drive circuit 231 operates based on the synchronization control signal from the laser signal processing unit 240, and applies a modulation current (light source emission signal) to the laser light source 232. As a result, the laser light source 232 emits laser light. The laser light emitted from the laser light source 232 is emitted to the outside via the irradiation lens 233.

  In this embodiment, an infrared semiconductor laser diode (LD: Laser Diode) is used as the laser light source 232, and near-infrared light having a wavelength of 800 nm to 950 nm is irradiated as laser light. The laser light source 232 periodically emits laser light having a pulse-like waveform according to the modulation current (light source emission signal) applied by the light source drive circuit 231. Further, the laser light source 232 periodically emits pulsed laser light having a short pulse width of about several nanoseconds to several hundred nanoseconds. However, the wavelength and pulse width of the laser light are not limited to these.

  The pulsed laser light emitted from the laser light source 232 is emitted to the outside as an irradiation beam via the irradiation lens 233, and then is applied to an object in the irradiation direction of the laser light. Since the laser light emitted from the laser light source 232 is collimated into substantially parallel light by the irradiation lens 233, the irradiation range of the irradiated object is suppressed to a predetermined small area.

  The laser radar ranging unit 120 further includes a light receiving lens 234, a light receiving element 235, and a light receiving signal amplifying circuit 236. Laser light applied to an object in the irradiation direction is scattered in a uniform direction. Then, only the light component reflected along the same optical path as the laser light emitted from the laser radar distance measuring unit 120 is guided to the light receiving element 235 via the light receiving lens 234 as reflected light.

  In the present embodiment, a silicon PIN photodiode or an avalanche photodiode is used as the light receiving element 235. The light receiving element 235 generates a reflection signal by photoelectrically converting the reflected light, and the light reception signal amplifying circuit 236 amplifies the generated reflection signal and transmits the amplified signal to the laser signal processing unit 240.

  Subsequently, each unit on the control board storage unit 202 side will be described. The laser signal processing unit 240 calculates the distance to the object in the irradiation direction based on the reflection signal transmitted from the laser radar distance measurement unit 120, and transmits the calculated distance information to the stereo image calculation unit 250.

  The stereo image calculation unit 250 is configured by a dedicated integrated circuit such as a field-programmable gate array (FPGA) or an application specific integrated circuit (ASIC). The stereo image calculation unit 250 outputs a synchronization control signal to the left camera 111, the right camera 112, and the laser signal processing unit 240 for controlling the imaging timing and the timing of emitting and receiving laser light.

  Further, the stereo image calculation unit 250 generates a parallax image based on the comparison image transmitted from the left camera 111, the reference image transmitted from the right camera 112, and the distance information transmitted from the laser signal processing unit 240. The stereo image calculation unit 250 stores the generated parallax image in the memory 260.

  The memory 260 stores the parallax image generated by the stereo image calculation unit 250. Further, the memory 260 provides a work area when the stereo image calculation unit 250 and the MPU 270 execute various processes.

  The MPU 270 controls each unit stored in the control board storage unit 202 and performs an analysis process for analyzing a parallax image stored in the memory 260.

<Ranging principle with stereo camera>
The principle of ranging by a stereo camera will be described with reference to FIG. FIG. 4 is a diagram illustrating the principle of deriving a parallax value for an object from a comparison image and a reference image by triangulation, and measuring the distance from the ranging device 100 to the object based on the parallax value.

First, it is assumed that the right camera 112 and the left camera 111 are installed in parallel equivalence. The S point on the object E in the three-dimensional space is mapped to a position on the same horizontal line of the right camera 112 and the left camera 111. That is, the S point in each image is captured at a point Sa (x, y) in the comparison image Ia and a point Sb (x ', y) in the reference image Ib. At this time, the parallax value d is expressed by Expression (1) using Sa (x, y) and Sb (x ', y).
d = x′−x Equation (1)
As shown in FIG. 4, the distance between the point Sa (x, y) in the comparison image Ia and the intersection of the perpendicular drawn from the right camera 112 onto the imaging surface is set to Δa, and the point Sb (x ′) in the reference image Ib is set. , Y) and the intersection point of the perpendicular drawn from the left camera 111 onto the imaging plane is represented by Δb, the disparity value Δ = Δa + Δb.

Further, the distance Z between the right camera 112 and the left camera 111 and the object E can be derived by using the parallax value Δ. Specifically, the distance Z is a distance from a plane including the focal position of the camera lens 211 and the focal position of the camera lens 221 to the specific point S on the object E. Using the focal length f of the camera lens 211 and the camera lens 221, the base line length B that is the length between the camera lens 211 and the camera lens 221, and the parallax value d, the distance Z is calculated by Expression (2). Can be.
Z = (B × f) / d Equation (2)
According to the equation (2), the distance Z is smaller as the parallax value d is larger, and the distance Z is larger as the parallax value d is smaller. As is clear from equation (2), the smaller the camera (the smaller the base line length B), the greater the distance corresponding to one integer parallax.

  FIG. 5 is a diagram illustrating the correspondence between the disparity value and the distance in a graph format. In FIG. 5, the base length B = 80 mm, the focal length f = 5 mm, and the cell size (pixel pitch) of the imaging elements 212 and 222 is 3 μm. As is clear from FIG. 5, when the parallax value d decreases, the distance sharply increases due to a change of one integer parallax. This means that the distance resolution is rapidly deteriorated as the distance increases (as the parallax value d decreases).

<< Integer disparity calculation by block matching >>
Next, a method of calculating a parallax value will be described with reference to FIG. FIG. 6 is an example of a diagram illustrating calculation of integer parallax by block matching. FIG. 6 shows the SAD (Sum of Absolute Difference) as the cost of the pixel of interest p = (Px3, Py5) in the comparison image Ia captured by the right camera 112 and the reference image Ib captured by the left camera 111. FIG. 5 is a diagram showing an example of calculating the value.

  Since the reference image Ib and the comparison image Ia have different imaging positions, even if the target pixel p = (Px3, Py5) at the same position on the captured image, the reference image Ib does not indicate the same object and is shifted left and right. Point. Therefore, the block size is set to 1 × 1 pixel, and the luminance value of the focused pixel p = (Px3, Py5) on the reference image Ib and the focused pixel p = (Px3, Py5) on the comparison image Ia The difference value from the luminance value of ()) is a large value.

  Next, the pixel of interest p on the comparison image Ia is shifted rightward by one pixel. That is, the SAD when the parallax d = 1 is calculated. Specifically, the difference between the luminance value of the pixel of interest p = (Px3 + 1, Py5) on the comparison image Ia and the luminance value of the pixel of interest p = (Px3, Py5) on the reference image Ib. Calculate the value.

  Hereinafter, similarly, d = 2, 3,..., And the SAD is calculated for each. In the example of FIG. 6, when d = 3, the object pointed to by the target pixel p = (Px3, Py5) of the reference image Ib and the target pixel p = (Px3 + 3, Py5) of the comparative image Ia are obtained. It is assumed that the object to be pointed is the same. Therefore, the SAD when d = 3 is smaller than the SAD when d = 3. The calculated SAD is called a cost, and in a certain search width (for example, 64 pixels), a parallax value (a parallax value obtained by a parallax indicating a minimum cost among costs changed as d = 1, 2, 3,...) (Integer parallax). Thereafter, decimal parallax is obtained by parabolic fitting, higher-order polynomial, or the like.

<Time measurement by laser signal processing unit>
Next, the principle of time measurement by the laser signal processing unit 240 will be described with reference to FIG. FIG. 7 is a diagram schematically showing the measurement of the time t until the irradiated laser light is reflected by the object and received. FIG. 7A schematically shows an analog signal, and FIG. 7B schematically shows a digital signal. As shown in FIG. 7A, a pulsed laser beam is emitted at a certain time, and a pulsed reflected light is received after a time t. Therefore, the distance to the object in the irradiation direction can be calculated by multiplying the time t by the speed of the light in the air.

  FIG. 7B illustrates the same principle with binarized irradiation light and reflected light. Since the laser radar ranging unit 120 receives noise other than reflected light, it is rare that only reflected light is clearly obtained as shown in FIG. Therefore, it is common to perform a process of comparing a signal received by the laser radar distance measurement unit 120 with a threshold value and detecting a signal exceeding the threshold value as a reflection signal. Since the laser signal processing section 240 binarizes the signal received based on the threshold value, a signal of 0 or 1 is obtained as shown in FIG.

<Configuration example of laser signal processing unit>
Next, a configuration example of the laser radar distance measurement unit 120 and the laser signal processing unit 240 will be described with reference to FIG. FIG. 8 is an example of a functional block diagram showing the functions of the laser radar distance measuring section 120 and the laser signal processing section 240 in a block shape.

  As an example, the distance measuring apparatus 100 is mounted on a vehicle that is a moving body, irradiates a laser beam, and receives light reflected (scattered) by an object (for example, a preceding vehicle, a stopped vehicle, an obstacle, a pedestrian, and the like). Then, information about the object such as the presence or absence of the object and the distance to the object (hereinafter, may be referred to as “object information”) is detected. The laser signal processing unit receives power supply from, for example, a battery (storage battery) of the vehicle.

  The laser radar ranging unit 120 and the laser signal processing unit 240 include the irradiation system 10, the light receiving optical system 30, the detection system 40, the time measurement unit 45, the synchronization system 50, the measurement control unit 46, and the like. The irradiation system 10 includes an LD (laser diode) as a light source, an LD driving unit 12, and an irradiation optical system 20. In the present embodiment, the LD 11 is used as the light source, but is not limited to this. For example, another light emitting element such as a VCSEL (surface emitting laser), an organic EL element, and an LED (light emitting diode) may be used.

  The LD 11 corresponds to the laser light source 232, is driven by the LD driving unit 12, and periodically emits a pulsed laser beam. The LD driving unit 12 turns on (emits) the LD 11 using the LD driving signal (rectangular pulse signal) output from the measurement control unit 46. The LD drive unit 12 includes, for example, a capacitor connected to be able to supply a current to the LD 11, a transistor for switching conduction / non-conduction between the capacitor and the LD 11, a charging unit capable of charging the capacitor, and the like. .

  In the present embodiment, for example, the light beam is scanned by rotating the reflection mirror. Using this operation method, the synchronization system 50 uses the synchronization detection PD 54 disposed on the optical path of the light beam reflected by the reflection mirror at a certain angle, and the synchronization detection PD 54. And a PD output detector 56 for detecting a voltage signal (light receiving signal) based on the output current (photocurrent) of the PD 54 for use. The angle of the reflection mirror can be detected by the signal of the PD for synchronization detection, and the scanning direction of the system can be synchronized. That is, the irradiation direction of each laser beam is determined based on the angle of the reflection mirror when the synchronization signal is obtained.

  The detection system 40 receives the light emitted from the irradiation optical system 20 corresponding to the irradiation lens 233 and reflected and scattered by the object via the light receiving optical system 30 corresponding to the light receiving lens 234, and receives the time measurement PD 42 (photo). Diode) and a PD output detection unit 44 that detects a voltage signal (light receiving signal) based on the output current (photocurrent) of the time measuring PD 42. As the light receiving element, a photo diode (PD), an avalanche photo diode (APD), a single photon avalanche diode (SPAD) that is a Geiger mode APD, or the like may be used.

  The time measuring PD 42 is driven by the light receiving element driving unit. The time measuring PD 42 corresponds to the light receiving element 235.

  The PD output detection unit 44, the time measurement unit 45, and the measurement control unit 46 correspond to the laser signal processing unit 240. The PD output detection unit 44 amplifies the analog signal (output voltage) from the time measurement PD 42 as needed, binarizes the analog signal (output voltage) based on a threshold voltage, and converts the binarized signal (digital signal) into a time measurement unit. 45. Note that binarization refers to converting a reflected signal into a signal of 1 or 0 by comparing it with a threshold. The PD output detector 44 has a characteristic configuration of the present embodiment, and will be described later in detail.

  The time measuring unit 45 measures the reciprocating time to the object from the binarized signal based on the light receiving timing and the rising timing of the LD drive signal, and outputs the time to the measurement control unit 46 as a time measurement result. That is, the time measuring unit 45 converts the time from the irradiation of the laser beam to the detection of the peak of the reflected signal into distance information to the object.

  The measurement control unit 46 starts measurement or stops measurement in response to a measurement control signal (measurement start signal or measurement stop signal) from the vehicle-mounted device. An LD drive signal is generated based on the synchronization signal and output to the LD drive unit 12 and the time measurement unit 45. In addition, the reciprocating distance to the object is calculated by converting the time measurement result from the time measuring unit 45 into a distance, and の of the reciprocating distance is output to the stereo image calculating unit 250 as distance information to the object.

<About the function of the stereo image calculation unit>
FIG. 9 is an example of a functional block diagram showing the functions of the stereo image calculation unit 250 in a block shape. As shown in FIG. 9, the stereo image calculation unit 250 includes a distortion correction unit 13 and a distance calculation unit 14. The right camera 112 and the left camera 111 form a stereo camera. In the present embodiment, a captured image captured by the right camera 112 is used as a reference image, and a captured image captured by the left camera 111 is used as a comparison image.

  The distortion correction unit 13 and the distance calculation unit 14 may be realized using a dedicated electronic circuit, or may be realized by executing a program for realizing each unit by a CPU (computer). Therefore, stereo image calculation section 250 has the function of an information processing device. It is also an image processing device in that it processes images.

  The distortion correction unit 13 performs general distortion correction on the reference image and the comparison image. By this image correction, the reference image and the comparative image are corrected so that no difference other than parallax occurs. Image correction is enabled by prior calibration. When the left camera 111 and the right camera 112 are installed, they capture, for example, a subject for calibration (for example, a checkered chart). A geometric conversion LUT (Look) for comparing two images and converting image data so as to minimize hardware internal error factors such as camera lens distortion, optical axis deviation, focal length deviation, and image sensor distortion. Up Table) has been generated. The distortion correction unit 13 performs image correction with reference to such an LUT.

  The distance calculation unit 14 calculates parallax by applying an algorithm such as block matching or the SGM propagation method to the reference image and the comparison image. Further, although the details will be described later, when outputting the distance image, the distance calculation unit 14 weights the cost for each pixel calculated from the distance information output by the laser radar distance measurement unit 120 and the stereo matching cost to make the final calculation. Cost is calculated.

  Note that the configuration in FIG. 9 is merely an example, and the laser signal processing unit 240 and the stereo image calculation unit 250 may be integrally configured. Also, the ECU 190 may have the function of the stereo image calculation unit 250.

<Problems of TOF method ranging method>
In the distance measurement method of the TOF method performed by the laser signal processing unit 240, a signal is received in a form as shown in Expression (3). The distance can be measured from the time from when the laser is irradiated to the time when the reflected light is received. As can be seen from Expression (3), the intensity of the signal Pr received by the time measurement PD 42 decreases with the square of the distance. For example, assume that the targets have the same reflectance Rtgt, Ssnd, and the like, and when the distance L is doubled (for example, 10 [m] and 20 [m]), the signal level becomes 1 / It will be 4.

  On the other hand, in order to gain the distance range, measures such as strengthening Po, improving TFG, or increasing the SRcv are taken.However, when Po is strengthened, the signal is too strong in the short distance part and saturation occurs. It may cause errors, errors may occur due to missing peaks, and costs may increase. Increasing the size of Srcv increases the size of the module, and the dramatic improvement in TFG is not expected.

L: Detection distance
S _Rcv : Area of receiver optical system η: Ratio of light that is focused on APD surface and input to APD
S _Snd : Irradiation beam area on target
R _Tgt : Target reflectance
P _O : Light source output
T _FG : Light Efficiency FIG. 10 is a diagram illustrating an example of the difference between the signal levels received near and far. FIG. 10A shows a reflection signal from a nearby object. The reflected signal is a portion of the signal Pr that shows a peak, and the rest is noise. As shown in FIG. 10A, the signal level of a nearby object can be strongly distinguished from noise and detected. The distance is calculated based on the time until the reflection signal is detected. There are various methods for determining the reflection signal from the signal Pr. For example, there is a method of detecting the maximum peak position or detecting a plurality of positions equal to or more than a threshold (multiple detection).

  FIG. 10B shows a reflected signal from a distant object. In the situation where the signal level of the distant object is weak and the reflected signal is almost the same as the noise intensity, either the method of detecting the maximum peak position or the method of detecting the signal level above the threshold, Difficult to detect reflected signals.

  In general, a method is used in which a signal Pr having a threshold or more is used as a reflection signal in order from a short distance, or a peak position is detected. However, in the example of FIG. 10B, the intensity of the substantial reflected signal is at the same level as the noise detected before that, and it is difficult to detect the reflected signal as a whole. is there. The laser radar distance measuring section 120 has such a problem in principle. Therefore, the stereo image calculation unit 250 calculates the processing range, and enables the laser radar distance measurement unit 120 to detect a reflected signal from a distant object.

<Function of stereo image calculation unit>
FIG. 11 is an example of a functional block diagram illustrating the function of the stereo image calculation unit 250 in detail. Note that FIG. 11 shows only a functional configuration for realizing processing related to parallax calculation, among functional configurations of the stereo image calculation unit 250, and other functional configurations (for example, a function for transmitting a synchronization control signal and the like). Has been omitted.

  The stereo image calculation unit 250 includes a processing range calculation unit 710 and a parallax image generation unit 720. The processing range calculation unit 710 further includes a first cost calculation unit 711, a first combined cost calculation unit 712, a first parallax calculation unit 713, and a range determination unit 714.

  In addition, the parallax image generation unit 720 further includes a second cost calculation unit 721, a second combined cost calculation unit 722, and a second parallax calculation unit 723. Hereinafter, details of each unit of the processing range calculation unit 710 and the parallax image generation unit 720 will be described.

<<< Details of Each Unit of Processing Range Calculation Unit >>>
First, details of each unit of the processing range calculation unit 710 will be described. First, the first cost calculation unit 711 calculates a cost by block matching as described with reference to FIG. This cost is called a first cost.

  The first combined cost calculator 712 calculates the first combined cost S by combining the costs C (p, d) of the respective pixel regions notified from the first cost calculator 711, and obtains a combined result. The first combined cost calculation unit 712 calculates a plurality of first path costs Lr using a processing method such as SGM (Semi-Global Matching), and calculates each first path cost Lr as the reference pixel area p. , The first combined cost S is calculated.

  FIG. 12 is a diagram illustrating details of the functional configuration of the first combined cost calculation unit. As shown in FIG. 12, the first combined cost calculator 712 includes a first route cost calculator 1001 and a first combined cost S calculator 1002. When acquiring the cost C (p, d) from the first cost calculating unit 711, the first route cost calculating unit 1001 calculates the first route cost Lr (p, d) based on the following equation (4).

Here, the above equation (4) is a general equation of the path cost Lr using SGM. In the above equation (4), P1 and P2 are fixed parameters.

  Based on the above equation (4), the first path cost calculation unit 1001 adds the first path cost in each pixel area in the r direction shown in FIG. 13 to the cost C (p, d) of the reference pixel area p. The first route cost Lr (p, d) is obtained by adding the minimum value of Lr. FIG. 13 is a diagram illustrating the r direction in the calculation of the first route cost Lr (p, d).

  As illustrated in FIG. 13, the first path cost calculation unit 1001 determines the first path cost Lr (for example, Lr135 (p−2r) in the endmost pixel area in the r direction (for example, r135 direction) of the reference pixel area p. , D)). Subsequently, the first route cost calculation unit 1001 obtains a first route cost Lr (Lr135 (pr, d)) along the r direction. In the present embodiment, the first route cost calculation unit 1001 calculates a first route cost Lr (for example, Lr135 (p, d)) obtained by repeating these processes in eight directions, and calculates the first route cost. Lr0 (p, d) to Lr315 (p, d) are obtained.

  Based on the first route costs Lr0 (p, d) to Lr315 (p, d) in eight directions obtained by the first route cost calculating unit 1001, the first combined cost S calculating unit 1002 calculates the following formula (5). The first combined cost S (p, d) is calculated.

The first synthesis cost S calculation unit 1002 notifies the first parallax calculation unit 713 of the calculated first synthesis cost S (p, d).

The first parallax calculation unit 713 extracts a corresponding pixel area in the comparison image Ia corresponding to the reference pixel area p based on the first synthesis cost S (p, d) calculated by the first synthesis cost calculation unit 712. Then, the parallax of the reference pixel area p is calculated. Note that the first cost calculator 711 and the first combined cost calculator 712 perform the same processing for other reference pixel regions in the reference image Ib. Then, the first parallax calculation unit 713 calculates each parallax (first parallax) for each reference pixel region, and notifies the range determination unit 714 of the calculation result.

  The range determination unit 714 extracts the parallax of the reference pixel area p and the surrounding pixel areas from the calculation result of the parallax (first parallax) notified from the first parallax calculation unit 713, and determines the processing range. This will be specifically described with reference to FIG.

  FIG. 14 is a diagram for explaining a method of determining the processing range by the range determining unit. In the reference image Ib, the irradiation position (point 530) where the laser beam is irradiated by the laser radar ranging unit 120 and its surroundings 2 shows a pixel region of FIG.

  As shown in FIG. 14, it is assumed that a laser beam is irradiated on an object in the real space corresponding to the position specified by the coordinates (xl, yl) of the point 530 in the reference image Ib. Note that the coordinates (xl, yl) may coincide with the coordinates (x, y) of the reference pixel area p, or may be slightly shifted.

  In such a case, the range determination unit 714 extracts a pixel area of ± 1 / 2a pixels in the horizontal direction and ± b pixels in the vertical direction with the irradiation position (xl, yl) as the center.

  The range determining unit 714 further extracts the parallax calculated for the extracted pixel region from the parallax calculated by the first parallax calculating unit 713. Note that, in the example of FIG. 14, the upper left pixel area, the upper right pixel area, the lower left pixel area, and the lower right pixel area among the parallaxes calculated for the pixel areas extracted by the range determining unit 714 due to space limitations. Only parallax calculated for each is shown.

  The range determination unit 714 further extracts a parallax with the highest calculation frequency from the extracted parallaxes. Then, the range determination unit 714 determines a distance corresponding to the parallax ± 1 pixel having the maximum calculation frequency as a processing range when the laser signal processing unit 240 processes the laser light reception signal.

  Specifically, the range determining unit 714 determines a range from the minimum distance corresponding to the parallax having the highest calculation frequency + 1 pixel to the maximum distance corresponding to the parallax having the highest calculation frequency−1 pixel among the extracted parallaxes. , The processing range for processing the laser light reception signal. Expression (2) is used for conversion from parallax to distance.

  Further, the range determining unit 714 notifies the laser signal processing unit 240 of the determined processing range. Accordingly, the laser signal processing unit 240 detects a signal indicating reflection on the object in the notified processing range among the laser reception signals, and calculates, for example, distance information of the point 530. The laser signal processing unit 240 converts a signal indicating reflection on an object into a processing range converted into a time range by dividing each of the minimum distance and the maximum distance defining the notified processing range by the speed of light. To detect.

  FIG. 15 is a diagram illustrating the processing range determined by the range determining unit 714. In FIG. 15, the horizontal axis represents the response time from the irradiation of the laser beam to the reception of the laser beam, and the vertical axis represents the signal intensity of the laser light reception signal. In FIG. 15, processing ranges 1310 and 1320 are processing ranges (processing ranges converted into time ranges) determined by the range determining unit 714.

  FIG. 15A shows an example of a laser light reception signal when the distance L to an object irradiated with laser light is short. As shown in FIG. 15A, when the distance L to the object irradiated with the laser beam is short, the signal strength (611) of the signal indicating the reflection on the object and the reflection on the object other than the object are determined in the processing range 1310. The difference between the signal strength and the signal strength (1312) is further increased. For this reason, detection of a signal indicating reflection from the object is further facilitated.

  FIG. 15B shows an example of the laser light reception signal when the distance L to the object irradiated with the laser light is long. As shown in FIG. 15B, even when the distance L to the object irradiated with the laser beam is long (for example, also at the point 530), according to the processing range 1320, the signal intensity (621) and The difference from the signal strength (1322) can be increased. That is, it is possible to increase the difference between the signal strength of the signal indicating the reflection on the object and the signal strength of the signal indicating the reflection on the object other than the object. For this reason, it is possible to easily detect the signal indicating the reflection on the object, and to reduce the possibility of erroneously detecting the signal indicating the reflection on the object other than the signal indicating the reflection on the object. it can.

  In the above description, among the parallaxes extracted by the range determining unit 714, a case has been described in which a distance equivalent to the maximum parallax ± 1 pixel of the calculation frequency is determined as the processing range. It is not limited to. For example, the range determination unit 714 may determine a processing range corresponding to parallax calculated based on the following equation (6).

In the equation (6), dmode is the parallax with the highest calculation frequency among the parallaxes calculated for the reference pixel area p and the surrounding pixel areas. w is a coefficient indicating the width of the standard deviation from the parallax having the highest calculation frequency. n ′ represents the number of parallaxes within ± 1 pixel with respect to the parallax with the highest calculation frequency among the integer parallaxes calculated for the reference pixel area p and the surrounding pixel areas. d ′ represents a parallax that is within ± 1 pixel of the parallax with the highest calculation frequency among integer parallaxes calculated for the reference pixel area p and the surrounding pixel areas.

  According to Equation (6), the processing range can be widened when the variation in parallax is large, and the processing range can be narrowed when the variation in parallax is small.

<< Details of each unit of parallax image generation unit >>
Subsequently, details of each unit of the parallax image generation unit 720 will be described.

  FIG. 16 is a diagram illustrating details of the functional configuration of the second cost calculator. The second cost calculation unit 721 includes a reference image acquisition unit 1401, a comparison image acquisition unit 1402, a cost C calculation unit 1403, and a cost C adjustment unit 1404. Further, the second cost calculation unit 721 includes a distance information acquisition unit 1411, a cost Cl calculation unit 1412, and a weighting addition unit 1420.

  The reference image acquisition unit 1401 acquires the reference image Ib from the right camera 112. Further, a reference pixel area p is extracted from the acquired reference image Ib. The comparison image acquisition unit 1402 acquires the comparison image Ia from the left camera 111.

  The cost C calculator 1403 calculates the cost C (p, d) of the reference pixel area p. The method of calculating the cost C (p, d) has been described with reference to FIG.

  The cost C adjusting unit 1404 adjusts the cost C (p, d) of the reference pixel area p calculated by the cost C calculating unit 1403 based on the reliability. The cost C adjustment unit 1404 obtains the adjusted cost C ′ (p, d) by performing adjustment using the following equation (7).

Here, D represents the maximum value of the shift amount (parallax). k represents a count value of the shift amount. Q (p) indicates the reliability of the cost C (p, d) of the reference pixel area p. The reliability Q (p) is calculated using, for example, the following equation (8).

Here, Cmin1 and Cmin2 represent the lowest cost and the second lowest cost among the respective costs C (p, d) calculated by changing the shift amount d within a predetermined range (0 to D). ing. The reliability Q (p) calculated based on the above equation (6) is obtained by normalizing the value calculated based on Cmin1 and Cmin2 to 0 to less than 1.0. Is corrected so as to approach, and is used in the above equation (7).

  The adjusted cost C ′ (p, d) adjusted by the cost C adjusting unit 1404 is, for example, a reference pixel area p in an area with a small texture (an area in which a change in pixel value between adjacent pixel areas is small). When the reliability Q (p) is low, the value becomes larger.

  The distance information acquisition unit 1411 acquires distance information from the laser signal processing unit 240. The distance information obtained from the laser signal processing unit 240 is distance information in which the possibility of erroneous detection is reduced by limiting the processing range. Here, the distance information acquired from the laser signal processing unit 240 is set to Zl. The distance information acquiring unit 1411 notifies the acquired distance information Zl to the cost Cl calculating unit 1412.

  The cost Cl calculation unit 1412 calculates the cost Cl based on the distance information Zl notified from the distance information acquisition unit 1411. The cost Cl is a parameter indicating the degree of dissimilarity between the pixel region in the comparison image Ia at the position derived based on the acquired distance information Zl and the reference pixel region p.

  Specifically, the cost Cl calculation unit 1412 first calculates the shift amount dl using the following equation (9) based on the distance information Zl. As a result, a pixel region in the comparison image Ia at a position derived based on the distance information Zl is extracted.

In the above equation (9), B is a base line length between the camera lens 211 and the camera lens 221. f is the focal length of the camera lens 211 and the camera lens 221.

  Subsequently, the cost Cl calculation unit 1412 calculates the cost Cl (p, dl) for the shift amount dl. Similarly to the above-described calculation of the cost C (p, d), the cost Cl calculating unit 1412 calculates the cost as the dissimilarity between the pixel value of the pixel region 511 at the position of the shift amount dl and the pixel value of the reference pixel region p. Calculate Cl (p, dl).

  The weighting addition unit 1420 uses the adjusted cost C ′ (p, d) adjusted by the cost C adjustment unit 1404 and the cost Cl (p, dl) calculated by the cost Cl calculation unit 1412 to calculate The weighted addition is performed based on the equation (10) to calculate the weighted cost.

Here, wd gives priority to the adjusted cost C ′ (p, d) adjusted by the cost C adjusting unit 1404 or the cost Cl (p, dl) calculated by the cost Cl calculating unit 1412. Is a weighting factor that indicates When giving priority to the adjusted cost C ′ (p, d) adjusted by the cost C adjusting unit 1404, the value of wd is increased. On the other hand, when giving priority to the cost Cl (p, dl) calculated by the cost Cl calculation unit 1412, the value of wd is reduced.

  Specifically, when the shift amount d ≠ dl, the value of wd is increased. This makes it possible to further increase the weighting cost of the pixel area of the shift amount d ≠ dl in the pixel area 511 in the comparison image Ia. Note that the adjusted cost C ′ (p, d) has a larger value in, for example, an area with less texture, and the value of wd is increased to reduce the adjusted cost C ′ (p, d). By giving priority, the weighting cost of the pixel area with the shift amount d ≠ dl becomes a larger value.

  On the other hand, when the shift amount d = dl, the value of wd is reduced. This makes it possible to further reduce the weighting cost of the pixel area with the shift amount d = dl among the pixel areas 511 in the comparison image Ia. The cost Cl (p, dl) calculated by the cost Cl calculating unit 1412 is a value smaller than the adjusted cost C ′ (p, d) calculated by the cost C adjusting unit 1404. For this reason, by making the value of wd smaller and giving priority to the cost Cl (p, dl) calculated by the cost Cl calculator 1412, the weighting cost of the pixel area with the shift amount d = dl becomes smaller.

  That is, according to the above equation (10), the difference in cost between the pixel area with the shift amount d = dl and the other pixel areas can be made more apparent as the weighting cost.

  As a result, when extracting the corresponding pixel region from the second combined cost S ′ calculated by the second combined cost calculation unit 722, the pixel region of d = dl in the comparison image Ia is easily extracted. That is, it is possible to accurately extract the position 512 of the corresponding pixel area corresponding to the reference pixel area p.

  In the above equation (10), the value of wd may be a fixed value, or may be changed according to the value of the distance information Zl. Alternatively, it may be configured to change according to the surrounding environment (for example, during the day or at night).

  The weighting addition unit 1420 notifies the second combined cost calculation unit 722 of the weighted cost calculated based on the above equation (10).

  FIG. 17 is a diagram illustrating details of the functional configuration of the second combined cost calculation unit. As shown in FIG. 17, the second combined cost calculator 722 includes a second route cost calculator 1501 and a second combined cost S ′ calculator 1502.

  The second combined cost calculator 722 calculates the second combined cost S ′ by combining the weighted costs of the respective pixel regions notified from the second cost calculator 721, and obtains a combined result. The second combined cost calculator 722 calculates a plurality of second route costs Lr ′ using a processing method such as SGM (Semi-Global Matching), and calculates each second route cost Lr ′ as a reference pixel. The second combined cost S ′ is calculated by consolidating in the area p.

  Upon acquiring the weighted cost, the second combined cost calculation unit 722 calculates the second route cost Lr ′ (p, d) based on the following equation (11).

Here, equation (11) is obtained by replacing the cost C (p, d) with a weighted cost in a general equation of the path cost using SGM. In the above equation (11), P1 and P2 are fixed parameters.

  According to Expression (11), the second combined cost calculation unit 722 adds the minimum value of the second path cost Lr ′ in each pixel region in the r direction to the weighted cost of the reference pixel region p, and Of the second path Lr ′ (p, d) of The second combined cost calculation unit 722 calculates the second route cost for each of the r0 to r315 directions, thereby obtaining the second route costs Lr0 ′ (p, d) to Lr315 ′ (p, d).

  Based on the second route costs Lr0 ′ (p, d) to Lr315 ′ (p, d) in eight directions, the second combined cost calculation unit 722 calculates the second combined cost S ′ (p, d) is calculated.

The second parallax calculation unit 723 extracts a corresponding pixel area in the comparison image Ia corresponding to the reference pixel area p based on the second synthesis cost S ′ calculated by the second synthesis cost calculation unit 722, and Recalculate the parallax of the area p.

  FIG. 18 is a diagram illustrating a calculation result of the second combined cost S ′ in the reference pixel area p. The second parallax calculation unit 723 calculates a shift amount dmin that minimizes the second combined cost S ′ (p, d) in the predetermined range (0 to D), thereby extracting a corresponding pixel region from the comparison image Ia. I do. Thereby, the second parallax calculating unit 723 can acquire the shift amount dmin as the parallax (second parallax) between the extracted corresponding pixel region and the reference pixel region.

  Note that the second parallax calculating unit 723 generates a parallax image by performing the same processing on other reference pixel regions in the reference image Ib and obtaining a recalculation result of the parallax (second parallax). The obtained parallax image is stored in the memory 260.

<Undershoot processing in case the reflected signal is extremely weak>
By detecting the reflected signal from the processing range, the laser signal processing unit 240 can increase the detection accuracy of the reflected signal in the received light signal. However, if the reflected signal is extremely weak in the first place, it becomes difficult to detect the reflected signal from the processing range. This will be described with reference to FIG.

  FIG. 19 is a diagram illustrating an example of image data captured at a distant place and a reflection signal from such a distant object. FIG. 19A is an entire reference image captured by the right camera 112, for example, and FIG. 19B is an image obtained by enlarging a part of FIG. 19A. In FIG. 19A, an object having a low reflectance of laser light (black curtain) is placed 143 [m] ahead, and FIG. 19B shows an object having a low reflectance of laser light (black curtain). .

FIG. 19C shows a reflection signal of a laser beam whose irradiation direction is an object (black curtain) having a low reflectance of the laser beam. Since the object (black curtain) having a low laser light reflectance is 143 [m] ahead, the peak of the reflected signal appears at about 1 × 10 ⁻⁶ seconds after irradiation, considering the round-trip time. In FIG. 19C, a portion surrounded by a circle 701 indicates a reflection signal from an object having a low reflectance of laser light.

  As shown in FIG. 19 (c), since the reflected signal is mixed with the surrounding noise, the signal level becomes so small that even if it is seen by a human, it cannot be judged even if a person does not know in advance where the peak is. ing.

  In addition to the attenuation due to the distance, in a backlight situation where sunlight shines from the front in the traveling direction (which easily occurs in an in-vehicle application), the noise becomes very large due to the sunlight, and the situation is further worsened. A technique for enhancing a reflected signal from an object even in such a situation will be described.

  FIG. 20 shows a configuration example of the PD output detection unit 44 that emphasizes a reflection signal from an object. 20 includes a current-voltage converter 44a, a signal amplifier 44b, and a binarization circuit 44c. The reflection signal from the object received by the time measurement PD 42 is converted by the time measurement PD 42 into a current value, and the current value is input to the PD output detection unit 44 as an output current.

  The current-voltage converter 44a converts the input current value into a voltage value. The voltage value is increased by the signal amplifier 44b and output as an output voltage (reflection signal). At this time, the signal amplifier 44b uses a non-inverting amplifier circuit (negative feedback circuit) having a negative feedback function, for example, including an operational amplifier OA. And overshoot occurs. The reflected light pulse from the object has substantially the same waveform as the emission pulse, and the pulse width is 10 ns to several tens ns as described above. It is desired for the PD output detection unit 44 to be able to respond to this reflected light pulse sufficiently quickly (high-speed response).

  The principle that undershoot and overshoot occur is that in a position control system composed of a mass point, a spring, and an attenuator, if the spring constant is too large (too much feedback is applied), the overshoot may occur from the steady position (overshoot). This is the same principle as the phenomenon of returning to (undershoot) repeatedly and settling to a steady position while performing damped vibration.

  Note that the configuration in FIG. 20 is merely an example, and the PD output detection unit 44 may include a high-pass filter that outputs a high-frequency component of the output of the signal amplifier 44b.

  FIG. 21A shows a light receiving signal waveform of Comparative Example 1, and FIG. 21B shows a light receiving signal waveform of the present embodiment. The light receiving signal waveform of Comparative Example 1 shown in FIG. 21 (A) shows the waveform of five light receiving signals with noise and the light receiving time 2 (the light receiving time) measured for each light receiving signal when undershoot does not occur. This is a simulation result of the timing at which the falling of the signal coincides with the threshold voltage). When the undershoot is not generated, the falling of the light receiving signal (output voltage) is slow (slow), and the measurement error at the light receiving time 2 increases due to the fluctuation caused by the noise.

  In an amplifier circuit using negative feedback, if the amplification factor is too large or the band of the circuit is too narrow for an input signal, the feedback becomes weak and no undershoot occurs. If the amplification factor is too small and the bandwidth of the circuit is widened, the ringing after undershoot increases and the threshold value is exceeded again after the signal falls. If the light receiving signal exceeds the threshold value again after the signal falls, the light receiving signal is recognized as a second object by the time measuring circuit, and a problem of erroneous detection of a pulse occurs. If the ringing itself is suppressed to prevent this erroneous pulse detection, the effect of the undershoot may be weakened.

  The light receiving signal waveform of Example 1 shown in FIG. 21B is a simulation of the waveform of five light receiving signals with noise and the light receiving time 2 measured for each light receiving signal when an undershoot occurs. The result. As a result, it can be seen that in Example 1, the falling of the light receiving signal is faster (sharper), the fluctuation caused by noise can be suppressed, and the measurement error at the light receiving time 2 can be reduced as compared with Comparative Example 1. In Comparative Example 1, the measurement error in five measurements was 11 ns (1.6 m in terms of distance), whereas in Example 1, the measurement error in five measurements was 3 ns (distance 0.5m).

  The light receiving signal waveforms of Comparative Example 1 and Example 1 are results of simulating the case where the negative feedback circuits 44b-1 and 44b-2 shown in FIGS. 22A and 22B are used as the signal amplifier 44b. is there. Each of the negative feedback circuits 44b-2 is connected in parallel to a resistor R1 having one end grounded or a bias voltage applied to one end thereof, and the other end connected to a negative input terminal of the operational amplifier OA, and the operational amplifier OA. It has a connected resistor R2 and capacitor C1, a capacitor C3 on the output side of the operational amplifier, and a resistor R3 for installing the capacitor C3.

  The voltage value from the current-voltage converter 44a is input to the + input terminal of the operational amplifier OA. In each negative feedback circuit, the capacitor C1 is not essential. In the first embodiment, C1 = 1 pF, R1 = 1 kΩ, R2 = 3 kΩ, C2 = 30 pF, and R3 = 500Ω, but the numerical values are merely examples.

  FIG. 23 shows one reflected signal subjected to the undershoot processing. In the present embodiment, like the reflection signal shown in FIG. 23, the time required for the ringing to settle after the undershoot (settling time R: the time required for the amplitude of the ringing to become 5% or less of the maximum amplitude of the received light signal). By realizing a waveform that takes a longer time to return to the ground level (zero) than that, prevention of erroneous detection of a pulse and improvement in distance measurement accuracy due to undershoot are realized at the same time.

  Such a received light waveform is realized by, for example, inputting a bell-shaped pulse having a full width at half maximum of 11 [ns] using the signal amplifier 44b shown in FIG. The light receiving waveform of FIG. 23 can be realized by preventing the discharge to the ground by a high-pass filter having a large time constant after the undershoot occurs in the negative feedback circuit.

<Detection of peak from reflected signal after undershoot processing>
An example of the undershoot-processed reflected signal will be described with reference to FIG. FIG. 24 is an example of a diagram illustrating a reflection signal from an object (black curtain) having a low reflectance of laser light. FIG. 24A shows the reflected signal subjected to the undershoot processing, and FIG. 24B shows a part of the reflected signal (enlarged) corresponding to the distance from the object shown in FIG.

  As shown in FIG. 24 (b), the first 20 [ns] of the reflected signal in which the undershoot has occurred is the period of the reflected signal indicating that the laser light has been reflected by the object. The valley portion of the second half 40 [ns] is the undershoot width (cycle of the undershoot portion).

  The period due to reflection from the object is determined by the relationship between the light emission pulse width and the response speed of the light receiving element driving unit. That is, the pulse width of the laser light becomes equal to the pulse width of the laser light if the light receiving element driving section can respond sufficiently fast, and becomes longer than the response speed of the light receiving element driving section if the response speed of the light receiving element driving section is slow. Further, the undershoot width can be designed by changing the values of the elements of the signal amplifier 44b.

The present embodiment is characterized in that the width Δt_u of the undershoot is set to be larger than the width Δt_r of the reflected signal. In other words, the frequency f_u of the undershoot is set lower than the frequency f_r of the reflected signal. Thereby, erroneous detection of the pulse can be prevented, and the pulse can be relatively large due to undershoot.
Δt_u> Δt_r
f_u <f_r
The frequency having a cycle of 20 [ns] of the reflected signal is 25 [Mhz], and the frequency having a cycle of 40 [ns] of the undershoot is 12.5 [Mhz]. Considering that signals of these periods are extracted from the reflected signal, a low-pass filter that passes signals of 12.5 to 25 [Mhz] or less is used.

FIG. 25 shows a configuration example of a PD output detection unit 44 that emphasizes a reflection signal from an object. The PD output detection unit 44 in FIG. 25 has an LPF 44d in addition to the configuration in FIG. The cutoff frequency fc of the LPF 44d is
f_u <f_r <f_c
Satisfy the relationship. As shown in FIG. 25, the LPF 44d whose cutoff frequency fc is set to be equal to or higher than the frequency f_r of the reflected signal is passed through the reflected signal.

  According to the LPF 44d, first, high-frequency components that become noise can be cut (filtering processing). Also, since the LPF 44d is adjusted so that the cutoff frequency fc satisfies the above relationship, an undershoot whose frequency is lower than the frequency of the peak corresponding to the object can be added. With the benefit of this undershoot, the amplitude of the reflected signal can be relatively increased. That is, the LPF 44d can enhance the reflected signal.

  In Patent Literature 1 (Japanese Patent Laid-Open No. 2017-062169), the purpose of ensuring a large amount of undershoot is not to “emphasize peaks”, but to “make a fall of a peak earlier and obtain a fall time without delay. That's why. The purpose of the LPF 44d of the present embodiment is to “emphasize peaks”, and is different in that an LPF 44d (filter circuit) is provided separately from the first processing circuit that performs undershoot processing.

  The processing range sent from the stereo camera unit 110 is input to the binarization circuit 44c, and the binarization circuit 44c binarizes the reflected signal in the processing range. This makes it possible to detect the peak of the reflected signal at the time corresponding to the object even if the reflected signal is extremely weak.

  FIG. 24C shows a reflected signal in which the low frequency component of the reflected signal of FIG. 24A is cut by the LPF 44d. In FIG. 24C, the LPF 44d is cut with the object light receiving signal width Δt_r = 20 [ns] (frequency f_r = 25 [MHz]) and the undershoot width Δt_u = 40 [ns] (frequency f_u = 12.5 [MHz]). An example in which the off frequency f_c is set to 30 [MHz] is shown.

  As shown in FIG. 24C, by applying the LPF 44d having an appropriate cutoff frequency, a clear peak can be confirmed at a time corresponding to the distance to the object. Therefore, if the PD output detection unit 44 narrows down the time corresponding to the distance to the object using the processing range from the stereo image calculation unit 250, the peak of the reflection signal can be detected even if the reflection signal is extremely weak. Become.

  Note that the cycle and frequency used in FIG. 24 are merely examples, and it is sufficient that “f_u, f_r, f_c” satisfies the relationship of f_u <f_r <f_c.

  For example, in the case where the light emission pulse width from the LD is 12 [ns], even if the response of the drive unit of the PD is not sufficient and the object light receiving signal width expands to 48 [ns]. , 48 [ns] is set to f_r, the undershoot frequency is set lower, and the cutoff frequency is set higher, the same effect can be obtained.

In FIG. 24, the effect obtained by passing through the LPF 44d has been described. However, the present embodiment is not limited to the LPF 44d, and a frequency lower than f_r may be cut. For example, a so-called BPF (bandpass filter) can be employed. In this case, the cutoff frequency on the high frequency side is f_cH, and the cutoff frequency on the low frequency side is f_cL,
f_cL <f_u <f_r <f_cH
Filtering processing that satisfies the relationship may be performed.

  For example, by cutting low-frequency components of shot noise and white noise due to backlight, the object reflection signal can be more easily detected, and some low-frequency characteristic noise can be effectively removed. Can be expected.

<Inconvenience caused by undershoot processing>
However, when an undershoot signal is generated in the reflected signal, there is a problem in detecting a row subject. A row subject is a subject in which a plurality of objects (for example, a person and a vehicle) are arranged in a line in the same direction as viewed from the laser radar distance measuring unit 120.

  FIG. 26 shows an example of a row subject. In FIG. 26, since a pedestrian is present in front of the vehicle, the pedestrian 811 and the vehicle 810 form a line when viewed from the laser radar distance measurement unit 120. When the distance between the plurality of objects in the platoon subject is short, the undershoot portion of the reflection signal of the object in front and the reflection signal of the object in the back overlap each other, and as a result, the shape of the reflection signal is distorted.

  FIG. 27 shows an example of a reflected signal of a row subject subjected to undershoot processing. FIG. 27 is a signal example for explaining the inconvenience of the reflected signal subjected to the undershoot processing instead of the actual reflected signal. As shown in FIG. 27, the peak 703 of the reflected signal of the car and the undershoot portion of the reflected signal 704 of the person cancel each other, and the peak 703 of the reflected signal of the car disappears. In other words, the peak 703 of the reflected signal of the car, which is a deep object, disappears completely, and the shape of the actually obtained reflected signal 702 is greatly different from what is expected.

  In order to avoid this inconvenience, the present embodiment employs a configuration that acquires both a reflected signal with an undershoot and a reflected signal without an undershoot. This will be described with reference to FIG.

FIG. 28 is a configuration diagram of an example of the PD output detection unit 44. The PD output detection unit 44 in FIG. 28 has three processing systems.
Processing system 1: Without undershoot Processing system 2: With undershoot + LPF
Processing system 3: No undershoot Processing system 1 is not suitable for processing to detect the distance to an object from an extremely weak reflection signal because there is no undershoot processing. Peaks corresponding to each object can be detected.

  The processing system 2 is not suitable for detecting the distance of a plurality of objects included in the row subject by the undershoot process and the LPF 44d, but can detect the distance to the object even when the reflected signal is extremely weak. The signal amplifier 44b of the processing system 1 and the signal amplifier 44b of the processing system 2 have different resistance and capacitor values. The signal amplifier 44b of the processing system 1 has an amplifying function but does not perform undershoot processing. The signal amplifier 44b of the processing system 2 has an amplifying function and performs an undershoot process. The processing system 3 will be described later.

  It should be noted that “without undershoot” is not necessary until there is no undershoot at all, and may include a slight undershoot of the reflected signal. In other words, if it is possible to separate the objects of the row subject, a slight undershoot is allowed.

  The time measuring unit 45 measures the time from each of the signals (digital signals) of the three processing systems to the object. Therefore, it is possible to measure the time (distance) to a distant object (an object whose reflected signal is extremely weak) and each of the platoon subjects.

<Operation procedure>
FIG. 29 is a flowchart illustrating the flow of a parallax image generation process performed by the stereo image calculation unit and the laser signal processing unit. The flowchart illustrated in FIG. 29 illustrates a process of calculating a parallax for one reference pixel region p corresponding to an irradiation position where the laser beam is irradiated by the laser radar distance measurement unit 120. Therefore, when generating a parallax image, the stereo image calculation unit 250 and the laser signal processing unit 240 execute the flowchart shown in FIG. 29 for each similar reference pixel region.

  First, the processing of the stereo image calculation unit will be described. The reference image acquisition unit 801 acquires the reference image Ib and extracts the reference pixel area p (Step S-10).

  Next, the cost C calculation unit 803 calculates the pixel value of the pixel region 511 at the position of the shift amount d in the comparison image Ia acquired by the comparison image acquisition unit 802 and the pixel value of the reference pixel region p. , And calculate the cost C (p, d) (step S-20).

  Next, the first route cost calculation unit 1001 calculates each first route cost Lr (p, d) based on the cost C (p, d) (step S-30).

  The first combined cost S calculation unit 1002 calculates the first combined cost S (p, d) based on each first route cost Lr (p, d) (Step S-40).

  The first parallax calculation unit 713 calculates a shift amount (first parallax) at which the first combined cost S (p, d) is minimized (Step S-50). Accordingly, the first parallax calculation unit 713 extracts the corresponding pixel region from the comparison image Ia and acquires the calculation result of the parallax (first parallax) between the extracted corresponding pixel region and the reference pixel region p.

  It should be noted that only the block matching may be sufficient to calculate at higher speed or rougher (the process up to step S-20 is stopped). Alternatively, machine learning (deep learning) may be used to calculate at high speed or roughly.

  The range determining unit 714 determines a processing range when the laser signal processing unit 240 processes the laser light reception signal based on the calculation result of the parallax (Step S-60). Further, the range determining unit 714 notifies the laser signal processing unit 240 of the determined processing range.

  Thereby, the laser signal processing unit 240 detects the signal indicating the reflection on the object, and calculates the distance information Zl.

  Next, the cost C calculation unit 1403 calculates the pixel value of the pixel region 511 at the position of the shift amount d in the comparison image Ia acquired by the comparison image acquisition unit 1402 and the pixel value of the reference pixel region p. , And calculate the cost C (p, d) (step S-70).

  The cost C adjustment unit 1404 adjusts the calculated cost C (p, d) based on the reliability Q (p), and calculates the adjusted cost C ′ (p, d) (Step S-80). .

  The distance information acquisition unit 1411 acquires, from the laser signal processing unit 240, distance information Zl indicating the distance to the object in the real space corresponding to the position of the reference pixel area p (step S-90).

  The cost Cl calculation unit 1412 calculates a cost Cl based on the distance information Zl acquired by the distance information acquisition unit 1411 (step S-100).

  The weighting and adding unit 1420 weights and adds the adjusted cost C ′ (p, d) adjusted by the cost C adjusting unit 1404 and the cost Cl (p, d) calculated by the cost Cl calculating unit 1412, The weighting cost is calculated (step S-110).

  The second route cost calculation unit 1501 calculates each second route cost Lr ′ (p, d) using the weighted cost (Step S-120).

  The second combined cost S ′ calculation unit 1502 calculates a second combined cost S ′ (p, d) based on each second route cost Lr ′ (p, d) (Step S-130).

  The second parallax calculation unit 723 extracts a corresponding pixel region from the comparison image Ia by calculating a shift amount (dmin) that minimizes the second synthesis cost S ′ (p, d) (Step S-140). . Thereby, the second parallax calculation unit 723 obtains a recalculation result of the parallax (second parallax) between the extracted corresponding pixel region and the reference pixel region p.

  Next, the processing of the stereo image calculation unit 250 will be described. First, the irradiation optical system 20 irradiates a pulse of a laser beam (step L-10). The light receiving optical system 30 receives the reflected signal (Step L-20).

  Next, the processing systems 1 and 2 of the PD output detection unit 44 process the reflected signal in parallel. The current-voltage converter 44a of the processing system 1 converts the current of the reflected signal into a voltage, and the signal amplifier 44b amplifies the reflected signal (Step L-30). The signal amplifier 44b of the processing system 1 does not cause the undershoot or causes only a slight undershoot.

  The current-voltage converter 44a of the processing system 2 converts the current of the reflected signal into a voltage, and the signal amplifier 44b amplifies and undershoots the reflected signal (step L-40).

  The LPF 44d performs a filtering process on the undershoot-processed reflected signal to reduce high frequency components (step L-50).

  Next, the laser signal processing unit 240 acquires the processing range from the stereo image calculation unit 250, and the binarization circuit 44c binarizes the processing ranges of the reflected signals of the processing system 1 and the processing system 2 (step L-). 60).

  Next, the time measuring unit 45 converts the measured time into distance information Z1 (step L-70). The laser signal processing unit 240 sends the distance information Z1 and the irradiation direction of the laser light to the stereo image calculation unit 250.

  Note that a method other than the method described in FIG. 29 may be adopted as a method of reflecting the LiDAR detection result in the parallax. These will be described with reference to FIGS. 31 and 32.

<Detection of road surface unevenness, road shoulder, etc.>
A description will be given of detection of road surface irregularities and road shoulders using the processing system 3. In the processing system 3, a reflected signal without undershoot is input to the A / D conversion circuit 44e. The A / D conversion circuit 44e may be a relatively slow circuit. By observing the global slope of the digital data sampled by the A / D conversion circuit 44e, the measurement control unit 46 performs road surface estimation (detects road surface unevenness and slope). Looking at a local mountain makes it possible to detect a road shoulder or a white line.

  FIG. 30 is an example of a diagram illustrating global analysis of a reflected signal converted by the A / D conversion circuit 44e. FIG. 30A shows a situation in front of the vehicle detected by the distance measuring apparatus 100, and FIG. 30B shows a reflected signal detected in the situation of FIG. As shown in FIG. 30A, a concave portion 852, a convex portion 853, and an object 854 exist in front of the distance measuring apparatus 100. The laser light 860 was irradiated in a vertically long shape including a concave portion 852, a convex portion 853, and an object 854.

  The reflected signal of FIG. 30A will be described. The closer the road surface, the more the laser light is reflected, so the reflected signal has a rising portion 850a. Thereafter, a reflected signal from a distant road surface is received with time, but the reflected signal is weakened as the distance increases, so the reflected signal attenuates with a gradient 851a that is inversely proportional to the square of the distance according to equation (3). Therefore, the unevenness generated on the slope 851a indicates the unevenness of the road surface or the presence of an object.

  In FIG. 30, a concave portion 852a and a convex portion 853a are confirmed at the inclination 851a. When the road surface is flat, the irradiation light continuously hits the road surface. However, when the road surface has the concave portion 852, the irradiation surface changes discontinuously. At the discontinuous point, the reflected signal drops for a moment, so it is estimated that there is a concave portion 852 due to the concave portion 852a of the reflected signal. The convex portion 853a is presumed to have a high road surface because the convex component has a large facing component to LiDAR and a large reflection signal. That is, it is estimated that there is a convex portion 853 on the road surface. Further, it is estimated that the object portion 854a in which the reflection signal sharply increases has the object 854 on the road surface. As described above, by the A / D conversion circuit 44e converting the reflection signal including the road surface into digital data, the measurement control unit 46 can perform the road surface estimation (detecting the unevenness and inclination of the road surface).

  In order to perform such a road surface estimation, the laser radar distance measuring unit 120 irradiates the laser so as to have a large diameter and hit the road surface widely. Since the laser may have a large diameter and the A / D conversion circuit 44e may have a relatively low sampling rate, the cost for narrowing down the laser and the cost for the A / D conversion circuit can be reduced. An increase in the cost of the radar ranging unit 120 can be suppressed.

<Another example of a method for reflecting the detection result of LiDAR in parallax>
FIG. 31 is an example of a diagram illustrating a method of calculating parallax based on a metric space.
(1) After determining the processing range in step S-60 in FIG. 29, the laser signal processing unit 240 artificially increases the peak of the LiDAR analog signal existing in the corresponding range.
(2) The laser signal processing unit 240 inverts the analog signal.
(3) The stereo image calculation unit 250 converts the cost C (p, d) of the parallax space calculated from the stereo image into the cost C (p, z) of the distance space.
(4) The stereo image calculation unit 250 obtains a synthesis cost obtained by adding the analog signal of (2) and the cost C (p, z) of (3).

  Parallax can be obtained by searching for the minimum value of the synthesis cost.

FIG. 32 is an example of a diagram illustrating a method for simply calculating parallax based on a metric space.
(1) The stereo image calculation unit 250 converts the cost C (p, d) of the parallax space calculated from the stereo image into the cost C (p, z) of the distance space.
(2) After determining the processing range in step S60 in FIG. 29, the laser signal processing unit 240 specifies the peak portion of the LiDAR analog signal in the corresponding range.
(3) The stereo image calculation unit 250 reduces the cost value of the cost C (p, z) corresponding to the peak portion by a predetermined value to obtain the modification cost.

  Parallax can be obtained by searching for the minimum value of the modification cost.

<Summary>
As described above, the distance measuring apparatus 100 of the present embodiment sets the processing range using the distance information to the object obtained from the stereo camera, narrows the region of interest in the reflected signal, and reduces the noise of the reflected signal. It is easy to suppress taking as an object.

  Since the reflected signal is so weak that it becomes less than the noise due to disturbance light, etc., it is difficult to detect the peak corresponding to the object even if the processing range is narrowed. By intentionally generating and passing through an LPF or BPF designed according to the frequency band of the reflected signal and the frequency band of the undershoot width, the peak of the reflected signal can be emphasized (the signal strength is relatively large due to the undershoot). it can). Therefore, it is possible to “improve the detection accuracy of a distant weak signal”.

  In addition, if the undershoot is intentionally generated, there is a disadvantage that it is difficult to separate a plurality of objects in the platoon subject. To cope with this, the distance measuring apparatus 100 of the present embodiment can prepare a processing system that performs undershoot processing on the reflected signal and a processing system that does not perform undershoot processing to enable separation of a plurality of objects in the platoon subject. it can. In other words, the reflected signal with the undershoot is used for detecting a distant weak signal, and the reflected signal without the undershoot is used for calculating the distance between a plurality of objects in the row. Therefore, it is possible to "improve the detection accuracy of a distant weak signal while enabling separation of the row subjects".

  In addition, the distance measuring apparatus 100 according to the present embodiment irradiates a large-diameter laser, observes a global inclination of a reflected signal, performs road surface estimation (roughness and inclination of the road surface), and looks at a local mountain. Thus, the road shoulder and the white line can be detected.

  A large-diameter laser beam is applied, so that the cost of narrowing down the laser can be reduced. Also, the road surface estimation requires A / D conversion at a relatively slow sampling rate, so the cost of the A / D converter can be reduced. it can.

<Other application examples>
As described above, the best mode for carrying out the present invention has been described using the embodiment. However, the present invention is not limited to such embodiment at all, and various modifications may be made without departing from the gist of the present invention. And substitutions can be added.

  In the above embodiment, the case where the cost C is calculated for the pixel area has been described. However, the calculation of the cost C may be performed for each pixel. That is, the “pixel region” described in the above embodiment includes one or more pixels.

  In the above embodiment, the case where the stereo camera unit 110 and the laser radar distance measuring unit 120 are integrally configured has been described. However, the stereo camera unit 110 and the laser radar distance measuring unit 120 are configured separately. May be done.

  In the above embodiment, the laser signal processing unit 240 mainly extracts the peak from the analog signal, and the stereo image calculation unit 250 handles the digital signal. However, the laser signal processing unit immediately executes the laser signal processing immediately after the light receiving element 235 receives the reflected signal. The unit 240 may convert the digital signal into a digital signal and extract a peak from the digital signal.

  In the above embodiment, the laser signal processing unit 240 and the stereo image calculation unit 250 have been described as being configured by a dedicated integrated circuit. However, for example, the functions of the present embodiment may be realized by the information processing apparatus executing a storage medium in which software program codes for realizing the functions of the laser signal processing unit 240 and the stereo image calculation unit 250 are recorded. .

  In the above embodiment, the distance information calculated by the laser signal processing unit 240 is input to the stereo image calculation unit 250 and used for recalculation of parallax, but is used for purposes other than recalculation of parallax. You may.

  In the above embodiment, the case where the distance measuring apparatus 100 is attached to the vehicle 140 has been described. However, the attachment destination of the distance measuring apparatus 100 is not limited to the vehicle 140, and may be a motorcycle, a bicycle, a wheelchair, a cultivator for agriculture, or the like. Alternatively, a moving body such as an autonomous mobile robot may be used. Alternatively, a flying object such as a drone may be used. Alternatively, an industrial robot or the like fixedly installed in an FA (Factory Automation) may be used.

  Note that the stereo image calculation unit 250 is an example of an image processing unit, the irradiation system 10 is an example of an irradiation unit, the time measurement PD 42 is an example of a light receiving unit, and the signal amplifier 44b of the processing system 2 is the first The LPF 44d is an example of a filter circuit, the binarization circuit 44c is an example of an acquisition unit, and the time measurement unit 45 is an example of a conversion unit. The signal amplifier 44b of the processing system 1 is an example of a second processing circuit, and the second parallax calculating unit 723 is an example of a recalculating unit. The distance for the range determination unit 714 to determine the processing range is an example of first distance information, and the distance calculated by the measurement control unit 46 is an example of second distance information. The cost Cl calculated by the cost Cl calculation unit 1412 is an example of a first cost, the cost C (p, d) calculated by the cost C calculation unit 1403 is an example of a second cost, and the weighting addition unit 1420 The calculated weighting cost is an example of a third cost.

1-3 Processing system 44 PD output detector 44a Current-voltage converter 44b Signal amplifier 44c Binarization circuit 44d LPF
45 Time measuring unit 46 Measurement control unit 100 Distance measuring device 110 Stereo camera unit 120 Laser radar distance measuring unit 140 Vehicle 190 ECU
240 laser signal processing unit 250 stereo image calculation unit

JP 2017-062169 A

Claims (10)

An irradiation unit that irradiates a laser beam;
A light receiving unit that receives a reflected signal of the laser light reflected by the object,
A first processing circuit that performs undershoot processing on the reflected signal received by the light receiving unit;
A filter circuit that performs a process of enhancing a peak representing the object on the reflection signal on which the undershoot process has been performed by the first processing circuit;
A distance measuring device comprising: An image processing unit that calculates first distance information from images obtained by the plurality of imaging devices,
An acquisition unit configured to acquire the peak from the processing range of the reflection signal determined by the image processing unit based on the first distance information;
Conversion means for converting the time from when the irradiation unit irradiates the laser beam to when the acquisition means detects the peak into second distance information to the object,
The distance measuring apparatus according to claim 1, further comprising:   The range finder according to claim 1, wherein the filter circuit is a low-pass filter or a band-pass filter. When the frequency corresponding to the undershoot portion is f_u, the frequency corresponding to the peak of the reflected signal is f_r, the cutoff frequency on the high frequency side is f_cH, and the cutoff frequency on the low frequency side is f_cL,
The distance measuring apparatus according to claim 1, wherein the filter circuit performs a filtering process that satisfies a relationship of f_cL <f_u <f_r <f_cH. A second processing circuit that does not perform undershoot processing on the reflected signal received by the light receiving unit;
The acquisition unit acquires the peak and the undershoot portion from the processing range of the reflection signal on which the first processing circuit has performed the undershoot process;
3. The distance measuring apparatus according to claim 2, wherein the second processing circuit acquires a peak representing the object from the processing range of the reflection signal on which the undershoot processing is not performed.   6. The distance measurement according to claim 5, further comprising a recalculating unit that recalculates parallax of images obtained by a plurality of imaging devices based on the second distance information converted by the converting unit. apparatus. The image processing unit converts the second distance information calculated by the conversion unit into a first cost,
Calculating a second cost obtained by block matching of the plurality of images obtained by the plurality of imaging devices, and calculating a distance from a third cost calculated by weighting the first cost and the second cost; The distance measuring apparatus according to claim 6, wherein the distance measurement is performed.   A moving object comprising the distance measuring device according to claim 1. Irradiating a laser beam by the irradiating unit;
A light receiving unit for receiving a reflected signal of the laser beam reflected by the object,
A first processing circuit performing an undershoot process on the reflected signal received by the light receiving unit;
A filter circuit for performing a process of enhancing a peak representing the object on the reflected signal on which the undershoot process has been performed by the first processing circuit;
A distance measuring method comprising: An irradiation unit that irradiates a laser beam;
A light receiving unit that receives a reflected signal of the laser light reflected by the object,
A first processing circuit that performs undershoot processing on the reflected signal received by the light receiving unit;
A filter circuit that performs a process of enhancing a peak representing the object on the reflection signal on which the undershoot process has been performed by the first processing circuit;
Program to function as