Image processing device, image processing method, and program

Abstract

To distinguish, by a simple structure, an obstacle and a floor surface by a depth camera.SOLUTION: An image processing device 1 includes a distance information obtaining unit 2 and a reference value calculating unit 3. The distance information obtaining unit 2 obtains, from an imaging device, distance information which is calculated from an image picked up by the imaging device and including a floor surface, and which represents a distance to an imaging object corresponding to each pixel of the image. The reference value calculating unit 3 calculates, on the basis of the furthest distance among the distances with respective frequencies taking the maximum value in the frequency distribution of distance relating to each pixel contained in the obtained distance information, the reference value for the distance to distinguish whether the imaging object is the floor surface or not.SELECTED DRAWING: Figure 13

Description

[Technical field]

[0001] The present disclosure relates to an image processing device, an image processing method, and a program. [Background technology]

[0002] The introduction of mobile objects such as robots that perform tasks in facilities such as factories and warehouses is progressing. Such mobile objects move along paths on floor surfaces within the facility while avoiding obstacles such as luggage and equipment. For this reason, information showing paths on floor surfaces within the facility that can be ridden is required.

[0003] To create route information, for example, a depth camera may be installed on the ceiling of a facility, and the floor surface and obstacles may be distinguished from the distance to the object captured by the depth camera. If the floor surface and obstacles can be distinguished in this way, for example, a control system of a moving object or the like may be able to generate information about the route.

[0004] An example of a method for measuring the distance to an imaging target using a depth camera is disclosed in, for example, Patent Document 1. [Prior art documents] [Patent documents]

[0005] [Patent Document 1] JP 2018-25906 A Summary of the Invention [Problem to be solved by the invention]

[0006] In general, it is known that in a depth camera that measures distance by parallax, the error becomes large depending on the distance to the imaged object. In addition, each depth camera has individual differences at the time of manufacture, and needs to be calibrated individually. To deal with this, it is possible to use a laser rangefinder or the like to measure the exact distance for each installed depth camera and correct the error. However, when a large number of depth cameras are set, it is difficult to perform correction for each depth camera.

[0007] One object of the present disclosure is to solve the above-mentioned problems and provide an image processing device, an image processing method, and a program that can distinguish between an obstacle and a floor surface using a depth camera with a simple configuration. [Means for solving the problem]

[0008] An image processing device in one aspect of the present disclosure includes an acquisition means for acquiring distance information calculated from an image including a floor surface captured by an imaging device, the distance information representing the distance from the imaging device to an imaging object corresponding to each pixel of the image, and a calculation means for calculating a reference distance value for distinguishing whether the imaging object is the floor surface or not, based on the farthest distance among the distances at which the frequency distribution of the distances related to each pixel included in the acquired distance information has a maximum frequency value.

[0009] An image processing method according to one aspect of the present disclosure includes acquiring distance information calculated from an image including a floor surface captured by an imaging device, the distance information representing a distance from the imaging device to an imaging target corresponding to each pixel of the image; A reference distance value for determining whether the object to be imaged is the floor surface is calculated based on the farthest distance among the distances at which the frequency distribution of the distances for each pixel included in the acquired distance information takes a maximum value.

[0010] A program according to an aspect of the present disclosure includes causing a computer to acquire distance information calculated from an image including a floor surface captured by an imaging device, the distance information representing a distance from the imaging device to an imaging target corresponding to each pixel of the image; A process is executed to calculate a reference distance value for distinguishing whether the object to be imaged is the floor surface or not, based on the farthest distance among the distances at which the frequency distribution of distances related to each pixel contained in the acquired distance information takes a maximum value. Effect of the Invention

[0011] An effect of the present disclosure is that an obstacle can be distinguished from the floor surface using a depth camera with a simple configuration. [Brief description of the drawings]

[0012] [Figure 1] FIG. 2 is a schematic diagram showing an overview of distance measurement within a facility in the first embodiment. [Diagram 2] 1 is a block diagram showing a configuration of an image processing device 10 according to a first embodiment. [Diagram 3] FIG. 4 is a diagram illustrating an example of a frequency distribution of distance information in the first embodiment. [Figure 4] 5 is a flowchart showing a reference value calculation process in the first embodiment. [Diagram 5] 3 is a diagram for explaining the correspondence between each pixel of an image captured by a depth camera and the imaging range of the camera in the first embodiment. FIG. [Figure 6] FIG. 4 is a diagram showing an example of distance information in the first embodiment. [Figure 7] 5 is a flowchart showing a maximum value detection process in the first embodiment. [Figure 8] 4 is a diagram illustrating an algorithm for detecting the farthest object distance Ze in the first embodiment. FIG. [Figure 9] 5 is a diagram showing the relationship between the farthest object distance Ze and a reference value in the first embodiment. FIG. [Figure 10]5A to 5C are diagrams illustrating an example of distance information after an imaging target corresponding to each pixel is determined in the first embodiment. [Figure 11] FIG. 4 is a diagram showing an example of a map generated based on a determination result in the first embodiment. [Figure 12] 10 is a flowchart showing a maximum value detection process in the first and second embodiments. [Figure 13] FIG. 1 is a block diagram showing a configuration of an image processing device 1 in the first and third embodiments. [Figure 14] FIG. 5 is a block diagram showing an example of the hardware configuration of a computer 500. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0013] The embodiments will be described in detail with reference to the drawings. In each drawing and each embodiment described in the specification, the same components are given the same reference numerals, and the description will be omitted as appropriate.

[0014] (First embodiment) A first embodiment will be described.

[0015] First, an overview of distance measurement in a facility in the first embodiment will be described. FIG. 1 is a schematic diagram showing an overview of distance measurement in a facility in the first embodiment. FIG. 1 shows, for example, a part of the interior of a factory. Referring to FIG. 1, a mobile object travels on a floor surface in the factory. Objects (hereinafter, described as obstacles) that are obstacles to the travel of the mobile object are arranged on the floor surface. In addition, a depth camera is installed on an upper part of the factory, for example, on the ceiling, and the depth camera captures images of the floor surface and obstacles. The depth camera transmits the captured image and information associated with the image (distance information to be described later) to the image processing device 10, for example, via a communication network. The image processing device 10 performs statistical processing (maximum value detection processing to be described later) on the received distance information, and calculates a reference value for distinguishing between obstacles and the floor surface based on the processing result. The image processing device 10 distinguishes between the floor surface and obstacles based on the calculated reference value, and displays a map showing a route on which the mobile object can travel (floor surface on which obstacles are not arranged) on a display device (not shown). In addition, the depth camera may be installed near the image processing device 10 and connected to each other via a USB cable or the like, and the depth camera may transmit distance information to the image processing device 10.

[0016] (composition) The configuration of an image processing device 10 in the first embodiment will be described.

[0017] Fig. 2 is a block diagram showing the configuration of the image processing device 10 in the first embodiment. With reference to Fig. 2, the image processing device 10 includes a distance information acquisition unit 11, a distance information storage unit 12, a reference value calculation unit 13, a floor surface determination unit 14, and an output control unit 15.

[0018] The distance information acquisition unit 11 acquires distance information calculated from an image including the floor surface captured by the depth camera, and represents the distance from the depth camera to the imaged object corresponding to each pixel of the image (hereinafter also referred to as the object distance).

[0019] The distance information storage unit 12 stores the distance information acquired by the distance information acquisition unit 11. The details of the distance information will be described later.

[0020] The reference value calculation unit 13 calculates a reference distance value for distinguishing whether the object to be imaged is a floor surface or not, based on the farthest distance among the distances at which the frequency takes a maximum value in the frequency distribution of distances related to each pixel included in the distance information acquired from the distance information storage unit 12.

[0021] 3 is a diagram showing an example of a frequency distribution of distance information in the first embodiment. The frequency distribution indicates, for each value of object distance z, the number of pixels (frequency) in an image that indicate that value. In the example of FIG. 3, the horizontal axis indicates object distance z, and the vertical axis indicates frequency N. In the example of FIG. 3, the farthest object distance Z among the object distances z at which the frequency is maximum is e The reference value calculation unit 13 calculates the object distance z corresponding to the maximum value of the convex B, that is, the farthest object distance Z e Based on this, calculate the reference value. e The method of calculating the reference value and the method of calculating the reference value will be described in detail later.

[0022] The floor surface determination unit 14 determines whether the imaging target corresponding to each pixel is a floor surface, based on the distance related to each pixel included in the distance information and the reference value calculated by the reference value calculation unit 13.

[0023] Based on the determination result of the floor surface determination unit 14, the output control unit 15 causes the display device to display a map showing a route along which the mobile object can travel.

[0024] (operation) Next, the operation of the first embodiment will be described.

[0025] (Reference value calculation process) The reference value calculation process in the first embodiment will be described below.

[0026] 4 is a flowchart showing the reference value calculation process in the first embodiment. The reference value calculation process is executed, for example, at a predetermined cycle to obtain a reference value for distinguishing between a floor surface and an obstacle. In addition, the reference value calculation process is executed, for example, when a depth camera is newly installed or relocated, to obtain a reference value as calibration.

[0027] The distance information acquisition unit 11 acquires distance information from the depth camera (step S11). Then, the distance information acquisition unit 11 causes the distance information storage unit 12 to store the acquired distance information.

[0028] Here, the distance information will be described with reference to Fig. 5. Fig. 5 is a diagram for explaining the correspondence between each pixel of an image captured by a depth camera and the imaging range of the camera in the first embodiment.

[0029] Referring to Fig. 5, the imaging targets, which are the floor surface and obstacles, are shown in an imaging range that can be captured by one depth camera. Also, above the imaging targets in the imaging range, each pixel of the image captured by the depth camera is virtually deployed on the camera plane. Here, the camera plane is a virtual plane parallel to the floor surface at the height of the camera. The position (xy coordinates) of each pixel is represented by the x-axis and y-axis (xy plane), and the distance z from the depth camera to a point on the floor surface or obstacle corresponding to each pixel is represented by the position on the z-axis (z coordinate).

[0030] Here, the number of pixels of an image captured by the depth camera is n×n (n is a positive integer). The number of pixels varies depending on the type of depth camera, and is not limited to n×n, and may be n×m (m is a positive integer).

[0031] In the example of Figure 5, the position (xy coordinates) of each pixel is O(x0y0), with the upper left corner of the image as the origin, and the x-axis runs from x0 to x n-1 The y-axis is expressed in coordinates from y0 to y n-1 In this case, for example, pixel A(x i y i)(For i which is a positive integer and i < n), the imaging object within the imaging range is located at point A' on the floor surface. And for pixel A(x i y i ), the object distance z, which is the distance from the depth camera to the floor surface, is assigned as the coordinate Z of the z-axis i,i .

[0032] In this way, each pixel is associated with xy coordinates indicating the position on the image (on the xy plane) and a z coordinate which is the object distance z at that coordinate.

[0033] FIG. 6 is a diagram showing an example of distance information in the first embodiment. The distance information includes the xy coordinates of each pixel, the z coordinate of each pixel, and the determination result. The determination result is the result of determining whether each pixel is the floor surface or an obstacle based on a reference value described later. In the example of the distance information in FIG. 6, the coordinates of the x-axis, y-axis, and z-axis associated with each pixel shown in FIG. 5 and the blank determination result are shown.

[0034] The reference value calculation unit 13 performs a maximum value detection process of detecting the farthest object distance Z among the object distances z at which the frequency takes the maximum value from the acquired distance information (step S12). Here, the maximum value detection process will be described in detail. e

[0035] (Maximum value detection process) FIG. 7 is a flowchart showing the maximum value detection process in the first embodiment. The reference value calculation unit 13 acquires the distance information at a certain time stored in the distance information storage unit 12 (step S121). The reference value calculation unit 13 generates a frequency distribution of the acquired distance information (step S122). Here, the reference value calculation unit 13 generates a frequency distribution of the object distance z for the entire pixels of the image as shown in FIG. 3.

[0036] FIG. 8 is a diagram for explaining an algorithm for detecting the farthest object distance Z e in the first embodiment. Hereinafter, the algorithm will be described with reference to FIG. 8.

[0037] ​The reference value calculation unit 13 calculates the farthest object distance Z e In order to detect the object distance z, an initial value Z0 is set to the variable Z representing the object distance z (step S123). Here, the initial value Z0 is set to a value larger than the possible value of the distance between the depth camera and the floor surface. By setting such an initial value Z0, the variable Z becomes the farthest object distance Z in the following local maximum value detection process. e The possible values ​​of the distance between the depth camera and the floor surface may be measured in advance, for example, by a laser distance measuring device.

[0038] The reference value calculation unit 13 calculates a range [m x ,M x ] is calculated (step S124). Here, x is an integer equal to or greater than 0 that indicates the number of repetitions. m x is the start point of the interval (the point where the object distance z is the minimum value) at the iteration number x, and M x is the end point of the interval (the point where the object distance z is maximum) at the iteration count x. x , end point M x For example, the value Z of variable Z at iteration number x x Using m x =Z x (1-A), M x =Z x It is calculated by (1+B). The constant values ​​A and B are real constants greater than 0 and less than 1, respectively. A and B are values ​​that vary depending on individual differences of depth cameras. Referring to FIG. 8(a), for example, the interval initially calculated using the initial value Z0 of the variable Z is [m0, M0].

[0039] The reference value calculation unit 13 calculates the average value of the object distance z within the calculated section, and updates the variable Z with the calculated average value (step S125). x+1 is in the interval [m x ,M x ], the average value of object distance z in

[0040]

number

[0041] Here, Δz is the bin width of the horizontal axis z of the frequency distribution, and N(m x ), N(m x +Δz), and N(M x ) are the object distance m x , Nm x +Δz and M x is the frequency (number of pixels) of

[0042] For example, the reference value calculation unit 13 calculates the value Z1 of the variable Z for the interval [m0, M0] based on the above-mentioned equation 1.

[0043] Next, the reference value calculation unit 13 judges whether the variable Z has converged (step S126). Here, the reference value calculation unit 13 judges that the variable Z has converged when the change in the variable Z satisfies a predetermined condition. As the predetermined condition, for example, the value Z of the variable Z calculated in the previous iteration x and the value of variable Z calculated in the current iteration, Z x+1 The predetermined condition may be that the difference between and falls within a predetermined range continuously for a predetermined number of repetitions.

[0044] If the variable Z has not converged (NO in step S126), the reference value calculation unit 13 repeats the interval calculation process in step S124 and the average value calculation process in step S125.

[0045] Referring to FIG. 8B, by repeating the interval calculation process in step S124 and the average value calculation process in step S125, the value Z of the variable Z at the number of repetitions k (where k=0, 1, 2, . . .) is calculated as follows: k is the farthest object distance Z e is approaching.

[0046] If the variable Z has converged (step S126 YES), the reference value calculation unit 13 calculates the converged variable Z by multiplying the farthest object distance Z e (step S127).

[0047] Referring to FIG. 8(c), the value Z of the variable Z at the iteration number l l is the farthest object distance Z e It converges to a neighborhood.

[0048] In this way, the reference value calculation unit 13 detects the maximum value from the distance information and calculates the farthest object distance Z e Calculate.

[0049] Referring to FIG. 4, the reference value calculation unit 13 calculates the farthest object distance Z e Here, the reference value calculation unit 13 calculates a reference value based on, for example, the farthest object distance Z e The reference value calculation unit 13 sets the object distance that is a predetermined distance closer to the farthest object distance Z e From the farthest object distance Z e Alternatively, the reference value may be set to a target distance that is a predetermined percentage closer than the farthest target distance Z e From the frequency distribution, the farthest object distance Z e Alternatively, an object distance that is closer than the width of the peak corresponding to the peak (for example, the object distance at which the frequency is a predetermined percentage of the maximum value) may be set as the reference value.

[0050] FIG. 9 shows the farthest object distance Z e 9 is a diagram showing the relationship between the farthest object distance Z e The distance to the object closest by a is the reference value Z s It states that:

[0051] The floor surface determination unit 14 determines whether the imaged object is a floor surface or an obstacle for each pixel included in the distance information based on the reference value calculated by the reference value calculation unit 13 (step S14). For example, in the example of the distance information shown in FIG. 6, the object distance z and the reference value Zs Compare the object distance z with the reference value Z s The object with smaller pixels is judged as an obstacle, and the object distance z is equal to or smaller than the reference value Z s The imaged object of the above pixels is determined to be the floor surface. Then, the floor surface determination unit 14 sets the determination result as distance information. In addition, since the distance information of each pixel fluctuates due to the unclearness of the image caused by, for example, the spherical aberration of the lens, the floor surface determination unit 14 sets the determination result as distance information. s Pixels located within a predetermined range of object distances in the vicinity may be determined to be the floor. In this way, the floor determination unit 14 can reduce fluctuations in the distance information of each pixel and reduce erroneous determinations.

[0052] Fig. 10 is a diagram showing an example of distance information after the imaging target corresponding to each pixel is determined in the first embodiment. The distance information in Fig. 10 is set with the determination result based on the relationship between each pixel and the imaging target shown in Fig. 5.

[0053] The output control unit 15 generates a map showing a route that the moving object can travel based on the determination result set in the distance information, and displays the map on the display device (step S15). Fig. 11 is a diagram showing an example of a map generated based on the determination result in the first embodiment. For example, the output control unit 15 generates the map in Fig. 11 based on the determination result of the distance information in Fig. 10. In the map in Fig. 11, the positions of pixels determined to be the floor surface are displayed as a route that can be traveled.

[0054] This completes the operation of the first embodiment.

[0055] In the above-described maximum value detection process, the frequency distribution of distance information is generated (step S121) and then the initial value is set (step S122). However, the frequency distribution of distance information may be generated after the initial value is set.

[0056] In the above-mentioned reference value calculation process, a frequency distribution is generated, the frequency is aggregated for each bin width, and the average value of the set interval is calculated. However, this is not limited to this. For example, the z coordinate value of each pixel of the distance information may be calculated within the interval [m x ,M x ] and divide it by the number of pixels in the range [m x ,M x ] may be calculated as the average value of the object distance z.

[0057] Next, the effects of the first embodiment will be described.

[0058] According to the first embodiment, the depth camera can distinguish between an obstacle and a floor surface with a simple configuration. The reason is that the distance information acquisition unit 11 acquires distance information calculated from an image including a floor surface captured by the depth camera, the distance information indicating the distance from the depth camera to an imaged object corresponding to each pixel of the image, and the reference value calculation unit 13 calculates a reference value of distance for distinguishing whether the imaged object is a floor surface or not, based on the farthest distance among the distances having a maximum frequency in the frequency distribution of the distances related to each pixel included in the acquired distance information.

[0059] (Modification of the first embodiment) In the reference value calculation process of the first embodiment described above, the initial value Z0 of the variable Z is set to a value larger than the possible value of the distance between the depth camera and the floor surface, and the interval is set based on the initial value Z0. In addition, constant values ​​A and B are used when calculating the interval for each repetitive process. However, this is not limited to the above, and the entire range of the object distance z of the frequency distribution may be set as the initial value of the interval, and variable values ​​A' and B' that become smaller at a predetermined rate for each repetition may be used instead of the constant values ​​A and B. In this case, the starting point m x、 End point M n are respectively, m x =Z x (1-A'), M x =Z x The range calculation process and the average calculation process are then repeated until the variable Z converges, thereby obtaining the farthest object distance Z eis calculated.

[0060] For example, if the floor area occupies more than half of the imaging range of the depth camera, the farthest object distance Z can be calculated by this process without setting the initial value Z0 of the variable Z. e can be calculated.

[0061] Second embodiment A second embodiment will now be described.

[0062] In the first embodiment, the maximum value detection process is performed based on distance information based on one image. In the second embodiment, the maximum value detection process is performed based on distance information based on multiple images taken at different times, and the furthest object distance z among the object distances z at which the frequency has a maximum value is detected. e Calculate.

[0063] 12 is a flowchart showing the maximum value detection process in the second embodiment. Here, the points different from the maximum value detection process in the first embodiment will be described.

[0064] The reference value calculation unit 13 sets an initial value Z0 of the variable Z representing the object distance z (step S221). The distance information acquisition unit 11 acquires distance information from the distance information storage unit 12 (step S222). Here, each time the distance information acquisition unit 11 repeats steps S222 to S226, it acquires distance information calculated from images captured at different times.

[0065] The reference value calculation unit 13 generates a frequency distribution for the acquired distance information (step S223). The interval calculation process (step S224) and the average value calculation process (step S225) are similar to the interval calculation process (step S123) and the average value calculation process (step S124) of the maximum value detection process in the first embodiment.

[0066] If the variable Z has not converged (NO in step S226), the reference value calculation unit 13 repeats the process from step S222, which is to obtain the distance information.

[0067] If the variable Z has converged (step S226 YES), the reference value calculation unit 13 calculates the converged variable Z by multiplying the farthest object distance Z e (step S227).

[0068] (Effects of the second embodiment) According to the second embodiment, even if there is a temporal fluctuation in the distance information of the depth camera due to the usage environment, etc., the reference value can be calculated correctly. This is because the reference value calculation unit 13 uses a frequency distribution related to the distance information calculated from images captured at different times for each repetition of the reference value calculation process to obtain the farthest distance among the distances at which the frequency is at a maximum value, and calculates a reference value for distinguishing between the floor surface and an obstacle. This reduces the influence of the temporal fluctuation in the distance information of the depth camera, and allows the calculation of a correct reference value.

[0069] (Third embodiment) A third embodiment will now be described.

[0070] 13 is a block diagram showing a configuration of an image processing device 1 in the third embodiment. The image processing device 1 includes a distance information acquisition unit 2 and a reference value calculation unit 3. The distance information acquisition unit 2 and the reference value calculation unit 3 are embodiments of the distance information acquisition means and the reference value calculation means of the present disclosure, respectively. Moreover, the distance information acquisition unit 2 and the reference value calculation unit 3 correspond to the distance information acquisition unit 11 and the reference value calculation unit 13 in the first embodiment, respectively.

[0071] The distance information acquisition unit 2 acquires distance information calculated from an image including a floor surface captured by the imaging device, the distance information representing the distance from the imaging device to an imaging target corresponding to each pixel of the image. The reference value calculation unit 3 calculates a reference value of distance for distinguishing whether the imaging target is a floor surface or not, based on the longest distance among the distances having a maximum frequency in a frequency distribution of the distances related to each pixel included in the acquired distance information.

[0072] (Effects of the third embodiment) According to the third embodiment, the depth camera can distinguish between an obstacle and a floor surface with a simple configuration. The reason is that the distance information acquisition unit 2 of the image processing device 1 acquires distance information calculated from an image including a floor surface captured by an imaging device, the distance information indicating the distance from the imaging device to an imaging target corresponding to each pixel of the image, and the reference value calculation unit 3 calculates a reference value of distance for distinguishing whether the imaging target is the floor surface or not, based on the farthest distance among the distances having a maximum frequency in the frequency distribution of the distances related to each pixel included in the acquired distance information.

[0073] (Hardware configuration) In each of the above-described embodiments, each component of the image processing device 10 is represented by a functional block. A part or all of each component of each device may be realized by any combination of a computer 500 and a program.

[0074] Fig. 14 is a block diagram showing an example of a hardware configuration of a computer 500. Referring to Fig. 14, the computer 500 includes, for example, a CPU (Central Processing Unit) 501, a ROM (Read Only Memory) 502, a RAM (Random Access Memory) 503, a program 504, a storage device 505, a drive device 507, a communication interface 508, an input device 509, an output device 510, an input / output interface 511, and a bus 512.

[0075] The program 504 includes instructions for implementing each function of each device. The program 504 is stored in advance in the ROM 502, the RAM 503, or the storage device 505. The CPU 501 implements each function of each device by executing the instructions included in the program 504. For example, the CPU 501 implements the functions of the distance information acquisition unit 11, the reference value calculation unit 13, the floor surface determination unit 14, and the output control unit 15 by executing the instructions included in the program 504. Also, for example, the RAM 503 may store the data of the distance information storage unit 12. Also, for example, the storage device 505 may store the data of the distance information storage unit 12.

[0076] The drive device 507 reads and writes data from and to the recording medium 506. The communication interface 508 provides an interface with a communication network. The input device 509 is, for example, a mouse or a keyboard, and receives information input from an operator or the like. The output device 510 is, for example, a display, and outputs (displays) information to an operator or the like. The input / output interface 511 provides an interface with peripheral devices. The bus 512 connects these hardware components. The program 504 may be supplied to the CPU 501 via a communication network, or may be stored in advance in the recording medium 506, read out by the drive device 507, and supplied to the CPU 501.

[0077] It should be noted that the hardware configuration shown in FIG. 14 is merely an example, and other components may be added, or some components may not be included.

[0078] There are various variations in the method of realizing the image processing device 10. For example, the image processing device 10 may be realized by any combination of a different computer and a program for each component. Also, multiple components of each device may be realized by any combination of a single computer and a program.

[0079] In addition, some or all of the components of each device may be realized by general-purpose or dedicated circuits including a processor, etc., or a combination of these. These circuits may be configured by a single chip, or may be configured by multiple chips connected via a bus. Some or all of the components of each device may be realized by a combination of the above-mentioned circuits, etc., and a program.

[0080] Furthermore, when some or all of the components of each device are realized by a plurality of computers, circuits, etc., the plurality of computers, circuits, etc. may be arranged in a centralized manner or in a distributed manner.

[0081] Although the present disclosure has been described above with reference to the embodiments, the present disclosure is not limited to the above embodiments. Various modifications that can be understood by a person skilled in the art can be made to the configuration and details of the present disclosure within the scope of the present disclosure. Furthermore, the configurations in each embodiment can be combined with each other without departing from the scope of the present disclosure. [Explanation of symbols]

[0082] 1, 10 Image processing device 2, 11 Distance information acquisition section 3, 13 Reference value calculation section 12 Distance information storage section 14 Floor surface determination section 15 Output control section 500 Computers 501 CPU 502 ROM 503 RAM 504 Program 505 Storage device 506 Recording media 507 Drive Unit 508 Communication Interface 509 Input Device 510 Output Device 511 Input / Output Interface 512 Bus

Claims

1. an acquisition means for acquiring distance information calculated from an image including a floor surface captured by an imaging device, the distance information representing a distance from the imaging device to an imaging target corresponding to each pixel of the image; a calculation means for calculating a reference value of distance for distinguishing whether the imaging target is the floor surface or not, based on the longest distance among the distances at which the frequency of the frequency of the distances for each pixel included in the acquired distance information is a maximum value; The calculation means repeatedly sets a new interval including an average value based on an average value of distances within the interval in the frequency distribution, and when a change in the average value satisfies a predetermined condition, sets the average value to the reference value. Image processing device.

2. the calculation means sets a value obtained by subtracting a predetermined distance from the farthest distance as the reference value. The image processing device according to claim 1 .

3. the calculation means sets the new interval so that the width becomes narrower each time the repetition is performed.

3. The image processing device according to claim 1 or 2.

4. the calculation means sets an initial value of the average value for setting the section to a value greater than a possible value of the distance between the imaging device and the floor surface.

4. The image processing device according to claim 1.

5. the calculation means calculates the average value using a frequency distribution related to the distance information calculated from the images captured at different times for each repetition.

5. The image processing device according to claim 1.

6. Further, a determination means is provided for determining whether or not the imaging target corresponding to each pixel is the floor surface based on the distance related to each pixel included in the distance information and the reference value.

6. The image processing device according to claim 1,

7. acquiring distance information calculated from an image including a floor surface captured by an imaging device, the distance information representing a distance from the imaging device to an imaging target corresponding to each pixel of the image; calculating a reference distance value for determining whether the imaging target is the floor surface or not based on the longest distance among the distances at which the frequency of the frequency distribution of the distances related to each pixel included in the acquired distance information is a maximum value; In the step of performing the calculation, a new interval including the average value is repeatedly set based on the average value of the distance within the interval in the frequency distribution, and when a change in the average value satisfies a predetermined condition, the average value is set to the reference value. Image processing methods.

8. On the computer, acquiring distance information calculated from an image including a floor surface captured by an imaging device, the distance information representing a distance from the imaging device to an imaging target corresponding to each pixel of the image; calculating a reference distance value for determining whether the imaging target is the floor surface or not based on the longest distance among the distances at which the frequency of the frequency distribution of the distances related to each pixel included in the acquired distance information is a maximum value; In the process of performing the calculation, a new interval including the average value is repeatedly set based on the average value of the distance within the interval in the frequency distribution, and when a change in the average value satisfies a predetermined condition, the average value is set to the reference value. A program that executes a process.

Images