Denseness estimation device and method

Abstract

To provide a congestion degree estimation device capable of accurately estimating the congestion degree of an individual of interest.SOLUTION: The difference calculating means 73 calculates a first difference which is a difference between the scattered power of the double reflection scattering when a herd of individuals occurs and the scattered power of the double reflection scattering when the herd of the individuals does not occur, and a second difference which is a difference between the scattering power of the volume scattering when the herd of the individuals occurs and the scattering power of the volume scattering when the herd of the individuals does not occur. Based on the first difference and the second difference, the estimation means 75 estimates the congestion degree which is the number of individuals per unit area.SELECTED DRAWING: Figure 15

Description

[Technical field]

[0001] The present invention relates to a density estimation device, a density estimation method, and a density estimation program for estimating the density of individuals. [Background technology]

[0002] A method for measuring the number of animals in a flock using a radar has been proposed (see, for example, Patent Document 1). Patent Document 1 describes a method for measuring the number of birds in a flock using the reflection intensity of radar radio waves.

[0003] Furthermore, Non-Patent Document 1 shows a model of volume scattering of radar radio waves in the case of a forest. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] JP 2016-90409 A [Non-patent literature]

[0005] [Non-Patent Document 1] Yoshio Yamaguchi, Toshifumi Moriyama, Motoi Ishido, Hiroyoshi Yamada, “Four-Component Scattering Model for Polarimetric SAR Image Decomposition”, IEEE, IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING, VOL. 43, NO. 8, AUGUST 2005 Summary of the Invention [Problem to be solved by the invention]

[0006] The technology described in Patent Document 1 measures the number of birds in a flock using only the reflection intensity of radio waves. Therefore, with the technology described in Patent Document 1, there is a possibility that objects other than the individual to be measured (birds in the case of Patent Document 1) may be recognized as the individual to be measured, and as a result, there is a possibility that the number of birds may not be measured accurately.

[0007] Here, the number of individuals per unit area is referred to as density.

[0008] An object of the present invention is to provide a density estimation device, a density estimation method, and a density estimation program that can accurately estimate the density of an individual of interest. [Means for solving the problem]

[0009] The density estimation device according to the present invention is capable of estimating the density of individuals when they are in a group. , calculated using polarimetric radar data The scattering power of double-reflection scattering and when no swarms of individuals are occurring , calculated using polarimetric radar data The first difference is the difference between the scattered power of the double reflected scattering and the scattered power of the single reflected scattering. , calculated using polarimetric radar data The scattering power of volume scattering and the scattering power when no swarms of individuals are occurring , calculated using polarimetric radar data The present invention is characterized by comprising a difference calculation means for calculating a second difference which is the difference with the scattering power of volume scattering, and an estimation means for estimating the density, which is the number of individuals per unit area, based on the first difference and the second difference.

[0010] In the method for estimating density according to the present invention, a computer calculates a density of a group of individuals when the group is generated. , calculated using polarimetric radar data The scattering power of double-reflection scattering and when no swarms of individuals are occurring , calculated using polarimetric radar data The first difference is the difference between the scattered power of the double reflected scattering and the scattered power of the single reflected scattering. , calculated using polarimetric radar data The scattering power of volume scattering and the scattering power when no swarms of individuals are occurring , calculated using polarimetric radar data A second difference, which is the difference from the scattering power of volume scattering, is calculated, and the density, which is the number of individuals per unit area, is estimated based on the first difference and the second difference.

[0011] The density estimation program according to the present invention is a program for estimating the density of a group of individuals by instructing a computer to estimate the density of a group of individuals when the group of individuals is generated. , calculated using polarimetric radar data The scattering power of double-reflection scattering and when no swarms of individuals are occurring , calculated using polarimetric radar data The first difference is the difference between the scattered power of the double reflected scattering and the scattered power of the single reflected scattering. , calculated using polarimetric radar data The scattering power of volume scattering and the scattering power when no swarms of individuals are occurring , calculated using polarimetric radar data The method is characterized by executing a difference calculation process that calculates a second difference, which is the difference with the scattering power of volume scattering, and an estimation process that estimates the density, which is the number of individuals per unit area, based on the first difference and the second difference. Effect of the Invention

[0012] According to the present invention, the density of an individual of interest can be estimated with high accuracy. [Brief description of the drawings]

[0013] [Figure 1] 1 is a block diagram showing an example of a congestion estimation device according to an embodiment of the present invention. [Diagram 2] 1 is a schematic diagram showing an area irradiated with radio waves when a polarized radar transmits and receives radio waves once. [Diagram 3] FIG. 1 is a schematic diagram showing a large number of pixels spread out on a plane. [Figure 4] FIG. 13 is a schematic diagram showing an example of pixel conversion. [Diagram 5] FIG. 2 is a schematic diagram showing a probability density function expressed by equation (8). [Figure 6] FIG. 2 is a schematic diagram showing a probability density function expressed by equation (15). [Figure 7] FIG. 1 is a schematic diagram showing which covariance matrix is ​​used depending on the value of 10log(<|SVV|2> / <|SHH|2>) when the individual is an animal. [Figure 8] FIG. 2 is a schematic diagram showing a probability density function expressed by equation (18). [Figure 9]FIG. 1 is a schematic diagram showing which covariance matrix is ​​used depending on the value of 10log(<|SVV|2> / <|SHH|2>) when the individual is a human. [Figure 10] 10 is a flowchart showing an example of a process flow when the density estimation device of the present embodiment learns a neural network. [Figure 11] 10 is a flowchart showing an example of a process flow when the density estimation device of the present embodiment estimates the density of individuals belonging to a flock. [Figure 12] 13 is a flowchart showing an example of the progress of a process for calculating the scattered power Pd of double reflective scattering and the scattered power Pv of volume scattering. [Figure 13] 13 is a flowchart showing an example of the progress of a process for calculating the scattered power Pd of double reflective scattering and the scattered power Pv of volume scattering. [Figure 14] 1 is a schematic block diagram showing an example of the configuration of a computer related to a congestion degree estimation device 10 according to an embodiment of the present invention. [Figure 15] 1 is a block diagram showing an overview of a congestion degree estimation device according to the present invention; DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0014] Hereinafter, an embodiment of the present invention will be described with reference to the drawings.

[0015] 1 is a block diagram showing an example of a density estimation device according to an embodiment of the present invention. The density estimation device 10 of the present embodiment includes a radar data acquisition unit 11, a density acquisition unit 12, a scattering model power decomposition unit 13, a learning unit 14, an estimation unit 15, a population map generation unit 17, and an output unit 18.

[0016] The radar data acquisition unit 11 acquires polarized radar data from a polarized radar (not shown) that uses multiple types of polarized waves. In the following description, a case where the multiple types of polarized waves are horizontal and vertical polarization will be described as an example. An example of such a polarized radar is a synthetic aperture radar that uses horizontal and vertical polarization. The synthetic aperture radar is mounted on a moving object such as an aircraft, an artificial satellite, or a drone. The polarized radar that uses horizontal and vertical polarization may be a ground-mounted polarized radar that can change the irradiation direction or irradiation position of radio waves.

[0017] The radar data acquisition unit 11 is communicably connected to the polarized radar and acquires polarized radar data from the polarized radar through communication. In the following description, the polarized radar data includes the transmission electric field strengths of the horizontally polarized wave and the vertically polarized wave, and the reception electric field strengths of the horizontally polarized wave and the vertically polarized wave.

[0018] FIG. 2 is a schematic diagram showing an area irradiated with radio waves when the polarized radar transmits and receives radio waves once. Here, this area is referred to as "area A." When the polarized radar transmits and receives radio waves once, it transmits radio waves to band-shaped area A shown in FIG. 2 and receives the reflected waves in area A. In FIG. 2, the areas shown as squares are areas corresponding to individual parts of area A. The areas corresponding to individual parts are, for example, approximately squares with one side measuring about 1 m to 10 m, depending on the type of polarized radar, etc. The time at which the polarized radar receives the reflected waves differs for each part in area A, so the received electric field strength differs for each part in area A. The transmitted electric field strength is common to each part in area A.

[0019] The density estimation device 10 treats each portion in the area A as a pixel and estimates the density of individuals for each pixel. The number of pixels corresponding to each portion in the area A is, for example, several thousand. In addition, the polarized radar changes the irradiation position of the radio waves and provides the polarized radar data to the radar data acquisition unit 11 after each transmission and reception of the radio waves or after all transmissions and receptions of the radio waves have been completed, so that the density estimation device 10 can estimate the density of individuals for each of the many pixels spreading out in a plane as shown in FIG. 3.

[0020] Below, we will explain an example in which the radar data acquisition unit 11 acquires polarized radar data relating to an area corresponding to one pixel and its surrounding pixels, and the density estimation device 10 performs processing on the area corresponding to that one pixel (hereinafter sometimes referred to as the area of ​​interest).

[0021] In this embodiment, the individuals to be the targets of density estimation are "humans" or "animals other than humans."

[0022] Next, the processing of the elements of the density estimation device 10 (see FIG. 1) when learning a learning model for estimating the density of individuals will be described.

[0023] When it is known that a group of individuals has occurred in the region of interest, the density acquisition unit 12 acquires the density in the region of interest at that time from another system that calculates density (hereinafter, simply referred to as another system; not shown). This density is teacher data used when learning a learning model for estimating density, and can be said to be a known density.

[0024] The congestion degree acquisition unit 12 is, for example, communicably connected to another system, and acquires the congestion degree from the other system through communication.

[0025] The method of calculating the density by the other system is not particularly limited. For example, each individual belonging to a flock occurring in the area of ​​interest may carry a GPS (Global Positioning System) terminal, and the other system may identify the number of individuals in the area of ​​interest by counting the number of GPS terminals in the area of ​​interest, and calculate the density from the number of individuals and the known area of ​​the area of ​​interest. Also, for example, the other system may photograph the area of ​​interest, count the number of individuals in the area of ​​interest based on the resulting image, and calculate the density from the number of individuals and the known area of ​​the area of ​​interest. Since the density is the number of individuals per unit area, the density is calculated by dividing the number of individuals by the area of ​​the area in which the flock of individuals occurs.

[0026] Furthermore, the radar data acquisition unit 11 acquires polarized radar data of the region of interest and its surrounding region from the polarized radar at the time when the above-mentioned swarm occurs.

[0027] Furthermore, the radar data acquisition unit 11 acquires polarized radar data of the region of interest and its surrounding area from the polarized radar at a time when it is known that no flock of individuals has occurred in the region of interest.

[0028] The scattering model power decomposition unit 13 performs scattering model power decomposition on the polarized radar data acquired by the radar data acquisition unit 11 to calculate the scattering power of the double reflection scattering and the scattering power of the volume scattering. The scattering model power decomposition unit 13 calculates the scattering power of the double reflection scattering and the scattering power of the volume scattering when a flock occurs in the region of interest, and the scattering power of the double reflection scattering and the scattering power of the volume scattering when a flock does not occur in the region of interest. The calculation method of these scattering powers will be described later. Furthermore, the scattering model power decomposition unit 13 calculates the difference between the scattering power of the double reflection scattering when a flock occurs in the region of interest and the scattering power of the double reflection scattering when a flock does not occur in the region of interest. Similarly, the scattering model power decomposition unit 13 calculates the difference between the scattering power of the volume scattering when a flock occurs in the region of interest and the scattering power of the volume scattering when a flock does not occur in the region of interest. Hereinafter, the difference between the scattering power of the double reflection scattering when a flock occurs and the scattering power of the double reflection scattering when a flock does not occur in the region of interest is referred to as the first difference. Moreover, the difference between the scattering power of volume scattering when a school is occurring and the scattering power of volume scattering when a school is not occurring is referred to as a second difference.

[0029] The learning unit 14 learns a learning model for estimating the density by machine learning using a combination of the density acquired by the density acquisition unit 12 (known density) and the first difference and the second difference corresponding to the density. This learning model is a learning model for estimating the density when the first difference and the second difference are given.

[0030] In this embodiment, the learning model is described by taking as an example a neural network. The number of layers of the neural network and the number of neurons in each layer are predetermined. However, the weight between any neuron in any layer and any neuron in the next layer is not determined. A weight matrix whose elements are the weights between neurons in the kth layer and neurons in the next layer (the k+1th layer) is defined as [W k ], then [W k ] can be expressed as shown below in equation (1).

[0031]

number

[0032] On the right side of equation (1), w ij k represents the weight between the j-th neuron in the k-th layer and the i-th neuron in the k+1-th layer. In addition, in equation (1), if the number of neurons in the k-th layer is n k and the number of neurons in the k+1th layer is n k+1 In this case, the number of elements (weights) in the weight matrix is ​​n k ×n k+1 There are 10.

[0033] The learning unit 14 learns the neural network (i.e., defines the neural network) by learning a weight matrix for each pair of adjacent layers of the neural network through machine learning using a combination of the density and the first difference and the second difference corresponding to the density.

[0034] Next, the processing of the elements of the density estimation device 10 (see FIG. 1) when estimating the density of individuals will be described.

[0035] The radar data acquisition unit 11 acquires polarized radar data of a target area for which density is to be estimated at a time when it is known in advance that no flocks of individuals have occurred in that area. The scattering model power decomposition unit 13 then uses the polarized radar data to calculate the scattering power of double reflection scattering and the scattering power of volume scattering when no flocks have occurred in advance, and stores each of these scattering powers. The scattering power of double reflection scattering when no flocks have occurred, which has been calculated and stored in advance, is denoted as P d0 The scattering power of volume scattering when no flocks are occurring, which is calculated and stored in advance, is expressed as P v0 The congestion estimation device 10 performs the above process for each area corresponding to each pixel.

[0036] The radar data acquisition unit 11 also acquires polarized radar data at the time when a flock of individuals is generated and the density of the individuals is to be estimated. The scattering model power decomposition unit 13 then calculates the scattering power of double reflection scattering and the scattering power of volume scattering at that time. The scattering power of double reflection scattering is defined as P d1 The scattering power of this volume scattering is expressed as P v1 This is expressed as:

[0037] The scattering model power decomposition unit 13 calculates the first difference (the difference between the scattering power of double reflection scattering when a school is generated and the scattering power of double reflection scattering when a school is not generated) by the following formula (2). d represents the first difference.

[0038] ΔP d =P d1 -P d0 (2)

[0039] Moreover, the scattering model power decomposition unit 13 calculates the second difference (the difference between the scattering power of volume scattering when a school occurs and the scattering power of volume scattering when a school does not occur) by the following formula (3). v represents the second difference.

[0040] ΔP v =P v1 -P v0 (3)

[0041] The estimation unit 15 estimates the first difference ΔP d and the second difference ΔP v The density of individuals is estimated by applying

[0042] The congestion estimation device 10 performs the above process for each area corresponding to each pixel.

[0043] This results in an estimate of the density of individuals for each pixel.

[0044] The population map generating unit 17 generates a population map. The population map is information that represents the number of individuals for each region in a map form. It is assumed that the density is estimated for each pixel. The population map generating unit 17 generates the population map by converting the density estimated for each pixel into the number of individuals. The conversion from density to the number of individuals can be performed by multiplying the density by the area of ​​the region corresponding to one pixel. It is assumed that the area of ​​the region corresponding to one pixel is known.

[0045] Here, if the area corresponding to one pixel is an approximate square with one side of about 1 m to 10 m, one pixel in the population map may be divided too finely. In such a case, the population map generating unit 17 generates the population map by converting the pixel so as to combine multiple pixels into one pixel. In this case, the area of ​​the area corresponding to one pixel after conversion is larger than the area corresponding to one pixel before conversion. FIG. 4 is a schematic diagram showing an example of pixel conversion. The upper part of FIG. 4 is a map showing the density of each pixel before conversion, and shows a map for 25 pixels. Here, the density of each pixel is expressed in a manner according to the density (for example, a pattern, or color, etc.). It is assumed that the 25 pixels shown in the upper part of FIG. 4 are converted into one pixel. In this example, it is assumed that the 25 pixels shown in the upper part of FIG. 4 are converted into one pixel in the first row and second column of the lower part of FIG. 4. In this case, the population map generating unit 17 calculates the average value of the density in the 25 pixels before conversion (see the upper part of FIG. 4). The population map generating unit 17 then calculates the number of individuals for the pixel in the first row, second column in the lower part of Fig. 4 by multiplying the area of ​​the region corresponding to one pixel after the conversion (one pixel in the lower part of Fig. 4) by the average value. The population map generating unit 17 similarly calculates the number of individuals for each of the other pixels in the lower part of Fig. 4. The population map generating unit 17 then generates, as a population map, information that expresses the number of individuals for each pixel after the conversion (each pixel in the lower part of Fig. 4) in a manner corresponding to the number of individuals (for example, pattern, color, etc.).

[0046] The output unit 18 is an output device for outputting the population map. For example, the output unit 18 may be a display device. In this case, the population map generation unit 17 causes the display device (output unit 18) to display the generated population map (e.g., see the lower part of FIG. 4). Note that the display device is an example of the output unit 18, and the output unit 18 is not limited to a display device. For example, the output unit 18 may be a printer device, and the population map generation unit 17 may cause the printer device (output unit 18) to print the generated population map (e.g., see the lower part of FIG. 4).

[0047] Next, a process of calculating the scattering power of double reflection scattering and the scattering power of volume scattering by the scattering model power decomposition unit 13 will be described. Note that it is assumed that the radar data acquisition unit 11 acquires polarized radar data of the region of interest and its surrounding region from a polarized radar.

[0048] The received electric field strength of horizontally polarized waves and vertically polarized waves is E H s ,E V s Then, the received electric field strength of the polarimetric radar is [E H s ,E V s ] T where T means transposition. Similarly, the transmission electric field strength of the horizontally polarized wave and the vertically polarized wave can be expressed as E H t ,E V t Then, the transmission electric field strength of the polarimetric radar is [E H t ,E V t ] T The subscript H represents horizontal polarization, and the subscript V represents vertical polarization. This also applies to the other symbols shown below.

[0049] Here, the following equation (4) holds true.

[0050]

number

[0051] In equation (4), the first matrix on the right side is the scattering matrix. The scattering model power decomposition unit 13 calculates the scattering model power decomposition result by using equation (4) and the received electric field strength [E H s ,E V s ] T and the transmission field strength [E H t ,E V t ] T Further, the scattering model power decomposition unit 13 calculates a covariance matrix expressed as in the following equation (5) from the scattering matrix.

[0052]

number

[0053] Here, "<>" means the ensemble average of the pixel in question and its surrounding pixels, and "*" represents the complex conjugate.

[0054] Moreover, the covariance matrix representing the surface scattering model is given by the following formula (6): where β in formula (6) is an unknown quantity.

[0055]

number

[0056] Moreover, the covariance matrix representing the model of two-times reflection scattering is given by the following formula (7): where α in formula (7) is an unknown quantity.

[0057]

number

[0058] Next, the covariance matrix representing the model of volume scattering will be described. First, the covariance matrix representing the model of volume scattering when the individual is an "animal other than a human (hereinafter simply referred to as an animal)" will be described. When the individual is an animal, it can be considered similar to the case where a group of individuals is a forest. This is because when the individual is an animal, the orientation of the dipole is considered to be close to random. The model of volume scattering in the case of a forest is described in Non-Patent Document 1.

[0059] If we consider a flock of animals to be similar to a forest, and the dipoles point in random directions, the probability density function of the dipole orientation as seen by the polarimetric radar is expressed by equation (8).

[0060]

number

[0061] Fig. 5 is a schematic diagram showing the probability density function expressed by equation (8). Fig. 5 shows that the probability of the existence of a dipole is constant regardless of the orientation of the dipole.

[0062] Moreover, the scattering matrix rotated by θ is expressed as the following equation (9).

[0063]

number

[0064] The covariance matrix representing the volume scattering model is a 3-row, 3-column matrix with 9 elements. The 9 elements are expressed as shown in the following equation (10) using equation (9).

[0065]

number

[0066] In equation (10), Re() is a function that extracts the real part of a complex number. i(i:1,2,...,10) is expressed as the following equation (11).

[0067]

number

[0068] Using equations (8), (10) and (11), we can obtain the covariance matrix that represents the volume scattering model. Here, the scattering matrix of a horizontally oriented dipole is [S] h-dipole HV The scattering matrix of a vertically oriented dipole is written as [S] v-dipole HV [S] h-dipole HV and [S] v-dipole HV are expressed as the following equations (12) and (13), respectively.

[0069]

number

[0070] The right-hand side of the first equation in equation (11) is [S] shown in equation (12). h-dipole HV The covariance matrix when the first equation of equation (11) is substituted with [S] shown in equation (13) v-dipole HV Here, we will find the covariance matrix when [S] is substituted into the right-hand side of the first equation of (11). h-dipole HV Even if you substitute [S] into the right-hand side of the first equation of (11), v-dipole HV The same covariance matrix is ​​obtained even when the following equation (14) is substituted: The covariance matrix can be expressed as the following equation (14), which can be said to be a covariance matrix that represents a volume scattering model.

[0071]

number

[0072] As mentioned above, let us consider a group of animals as being similar to a forest. In actual forests, the structure in which dipoles are oriented vertically is dominant, and Non-Patent Document 1 shows a more accurate probability density function of a forest (probability density function of the dipole orientation). This probability density function is expressed as the following formula (15).

[0073]

number

[0074] Fig. 6 is a schematic diagram showing the probability density function expressed by equation (15). Fig. 6 shows that the probability of existence of a dipole pointing in the upward direction is higher than the probability of existence of a dipole pointing in any other direction.

[0075] The covariance matrix representing the volume scattering model when the probability density function expressed by equation (15) is used can also be obtained. In this case, the covariance matrix representing the volume scattering model is obtained using equations (15), (10), and (11). As in the previous case, the [S] shown in equation (12) is added to the right side of the first equation of equation (11). h-dipole HV The covariance matrix when the first equation of equation (11) is substituted with [S] shown in equation (13) v-dipole HV Substituting [S] into the right-hand side of the first equation of (11), we obtain the covariance matrix and h-dipole HV When this is substituted, the covariance matrix of the following equation (16) is obtained, and [S] is added to the right-hand side of the first equation of equation (11). v-dipole HV When this is substituted, the covariance matrix of equation (17) is obtained.

[0076]

number

[0077] The matrices shown in equations (14), (16), and (17) are all covariance matrices that represent a model of volume scattering when the individual is an animal. Here, 10log(<|SVV | 2 > / <|S HH | 2 The covariance matrix used is determined by the value of 10log(<|S >). VV | 2 > / <|S HH | 2 10log(<|S >) VV | 2 > / <|S HH | 2 If 10log(<|S >) is less than -2 dB, the covariance matrix shown in equation (17) is used. VV | 2 > / <|S HH | 2 When 10log(<|S >) is between -2 dB and 2 dB, the covariance matrix shown in equation (14) is used. VV | 2 > / <|S HH | 2 >) exceeds 2 dB, the covariance matrix shown in equation (16) is used.

[0078] Next, we will explain the covariance matrix that represents the model of volume scattering of a human group when the individuals are humans. In this case, the covariance matrix (covariance matrix shown in formula (14)) derived using the probability density function expressed by formula (8) is also used.

[0079] In addition, the arms of people in a group are often pointing vertically downward. If we consider an arm as a dipole, the probability density function of the dipole orientation is expressed by the following equation (18).

[0080]

number

[0081] Fig. 8 is a schematic diagram showing the probability density function expressed by equation (18). Fig. 8 shows that the probability of existence of a dipole pointing downward is particularly higher than the probability of existence of dipoles pointing in other directions.

[0082] The covariance matrix representing the volume scattering model when the probability density function expressed by equation (18) is used can also be obtained. In this case, the covariance matrix representing the volume scattering model is obtained using equations (18), (10), and (11). As in the previous case, the [S] shown in equation (12) is added to the right side of the first equation of equation (11). h-dipole HV The covariance matrix when the first equation of equation (11) is substituted with [S] shown in equation (13) v-dipole HV Substituting [S] into the right-hand side of the first equation of (11), we obtain the covariance matrix and h-dipole HV When this is substituted, the covariance matrix of the following equation (19) is obtained, and [S] is added to the right-hand side of the first equation of equation (11). v-dipole HV When this is substituted, the covariance matrix of equation (20) is obtained.

[0083]

number

[0084] The matrices shown in equations (14), (19), and (20) are all covariance matrices that represent a model of volume scattering when the individual is a human. VV | 2 > / <|S HH | 2 The covariance matrix used is determined by the value of 10log(<|S >). VV | 2 > / <|S HH | 2 10log(<|S >) VV | 2 > / <|S HH |2 If 10log(<|S >) is less than -3.5 dB, the covariance matrix shown in equation (20) is used. VV | 2 > / <|S HH | 2 When 10log(<|S >) is between -3.5 dB and 3.5 dB, the covariance matrix shown in equation (14) is used. VV | 2 > / <|S HH | 2 >) exceeds 3.5 dB, the covariance matrix shown in equation (19) is used.

[0085] As mentioned above, whether the individual is an animal or a human, the covariance matrix that represents the volume scattering model is 10log(<|S VV | 2 > / <|S HH | 2 >).

[0086] Next, a method for calculating the scattering power of double reflective scattering and the scattering power of volume scattering will be described.

[0087] Covariance matrix <[C]> that represents the surface scattering model surface hv , the covariance matrix <[C]> that represents the model of double-reflection scattering double hv , the covariance matrix <[C]> that represents the volume scattering model vol hv Using the covariance matrix <[C]> HV is expressed as the following equation (21).

[0088]

number

[0089] In formula (21), f s ,f d ,f v are expansion coefficients and unknowns. As mentioned above, α and β are also unknowns. Here, fs ,f d ,f v , α, and β are used to obtain the following equation (22).

[0090]

number

[0091] f in each equation shown in (22) v Coefficient of m ij is the element in the i-th row and j-th column of the covariance matrix that represents the volume scattering model. As mentioned above, the covariance matrix that represents the volume scattering model depends on whether the individual is an animal or a human, and 10log(<|S VV | 2 > / <|S HH | 2 It depends on the value of m > shown in equation (22). 11 ,m 22 ,m 33 ,m 13 The following is a concrete example.

[0092] m when the individuals in the herd are animals 11 ,m 22 ,m 33 ,m 13 The value of is shown in the following equation (23).

[0093]

number

[0094] m when the individuals in the group are humans 11 ,m 22 ,m 33 ,m 13 The value of is shown in the following equation (24).

[0095]

number

[0096] The scattering power of the surface scattering is P sThe scattering power of the double reflection scattering is P d The scattering power of the volume scattering is P v These scattering powers P s ,P d ,P v can be expressed as the following equation (25).

[0097]

number

[0098] In addition, the scattering power P s ,P d ,P v The sum of these is expressed as shown in equation (26) below.

[0099]

number

[0100] The scattering model power decomposition unit 13 obtains a scattering matrix (the first matrix on the right side of the equation (4)) from the polarized radar data (specifically, the received electric field strength and the transmitted electric field strength) acquired by the radar data acquisition unit 11 and the equation (4). From the scattering matrix, the scattering model power decomposition unit 13 obtains a covariance matrix <[C]> HV As a result, <|S HH | 2 >,<|S HV | 2 >,<|S VV | 2 >,<|S HH S VV * The scattering model power decomposition unit 13 substitutes these values ​​into equation (22) and solves the simultaneous equations using the eight equations shown in equations (22), (25), and (26), thereby obtaining the unknown P s ,P d ,P v ,f s ,f d ,f v , α, β are calculated. As a result, the scattering power P dand the scattering power of volume scattering P v is obtained.

[0101] In addition, it is assumed that whether the individual to be the subject of density estimation is an animal or a human is specified in advance. If the individual is an animal, 10log(<|S VV | 2 > / <|S HH | 2 >) depending on the value of m 11 ,m 22 ,m 33 ,m 13 The value of is determined based on equation (23) and substituted into equation (22). In addition, if the individual is a human, 10log(<|S VV | 2 > / <|S HH | 2 >) depending on the value of m 11 ,m 22 ,m 33 ,m 13 The value of is determined based on equation (24) and substituted into equation (22).

[0102] If a group of individuals occurs in the target area for estimating density, the P calculated as above can be used to estimate the density of individuals using the polarimetric radar data at the time of estimating the density of individuals. d ,P v are the aforementioned P d1 ,P v1 is equivalent to.

[0103] In addition, the P calculated as above was calculated using polarimetric radar data at a time when it was known that no flocks of individuals were occurring in the target area. d ,P v are the aforementioned P d0 ,P v0 Here, using polarimetric radar data from a time when it is known that no flocks of individuals have occurred in the target area, P d ,P vWhen calculating, it is possible to specify whether the individual belonging to the group in the event of a group occurrence is an animal or a human. In that case, as in the above case, if the specified individual is an animal, 10log(<|S VV | 2 > / <|S HH | 2 >) depending on the value of m 11 ,m 22 ,m 33 ,m 13 The value of is determined based on equation (23) and substituted into equation (22). In addition, if the specified individual is a human, 10log(<|S VV | 2 > / <|S HH | 2 >) depending on the value of m 11 ,m 22 ,m 33 ,m 13 The value of is determined based on equation (24) and substituted into equation (22). In this case, P d0 ,P v0 and P when the specified individual is a human. d0 ,P v0 There may be cases where the above differs.

[0104] In addition, using polarimetric radar data from a time when it is known that no flocks of individuals have occurred in the target area, d ,P v When calculating P, regardless of whether the designated individual is an animal or a human, d ,P v If no population of individuals occurs in the target area, 10log(<|S VV | 2 > / <|S HH | 2 >) is considered to be near 0 dB. In this case, whether the individual is a human or an animal, (m 11 ,m 22 ,m 33 ,m 13 ) is common to (3 / 8, 1 / 4, 3 / 8, 1 / 8). Therefore, m 11 ,m22 ,m 33 ,m 13 Substituting 3 / 8, 1 / 4, 3 / 8, and 1 / 8 into equation (22), we get P d ,P v In this case, P may be calculated depending on whether the individual is an animal or a human. d0 ,P v0 are unchanged.

[0105] As described above, the scattering model power decomposition unit 13 calculates the first difference ΔP d (the difference between the scattered power of the double reflection scattering when a school occurs and the scattered power of the double reflection scattering when a school does not occur) is calculated by the calculation of the formula (2). In addition, the scattering model power decomposition unit 13 calculates the second difference ΔP v (The difference between the scattering power of volume scattering when a swarm occurs and the scattering power of volume scattering when a swarm does not occur) is calculated using equation (3).

[0106] The estimation unit 15 estimates the first difference ΔP d and the second difference ΔP v The density of individuals is estimated by applying the above to a neural network.

[0107] The radar data acquisition unit 11 and the density acquisition unit 12 are realized, for example, by a central processing unit (CPU) of a computer that operates according to a density estimation program and a communication interface of the computer. For example, the CPU may read the density estimation program from a program recording medium such as a program storage device of the computer, and operate as the radar data acquisition unit 11 and the density acquisition unit 12 using a communication interface according to the density estimation program. In addition, the scattering model power decomposition unit 13, the learning unit 14, the estimation unit 15, and the population map generation unit 17 are realized, for example, by a CPU of a computer that operates according to a density estimation program. For example, the CPU may read the density estimation program from a program recording medium as described above, and operate as the scattering model power decomposition unit 13, the learning unit 14, the estimation unit 15, and the population map generation unit 17 according to the density estimation program.

[0108] Next, the process will be described.

[0109] 10 is a flowchart showing an example of the process flow when the density estimation device 10 of this embodiment learns a neural network. Descriptions of matters that have already been explained will be omitted as appropriate.

[0110] It is assumed that whether the individuals belonging to the group are humans or animals is specified in advance. In addition, it is assumed that a group of individuals occurs in an area where another system calculates the density. Furthermore, when a group of individuals does not occur, the scattering power P d (P d0 ), and the scattering power of the volume scattering P v (P v0 ) is assumed to be calculated in advance.

[0111] The radar data acquisition unit 11 acquires polarized radar data from a polarized radar (step S1). This polarized radar data is data acquired by the polarized radar irradiating a target area for which another system calculates the density by irradiating the target area with radio waves.

[0112] Next, the scattering model power decomposition unit 13 uses the polarimetric radar data obtained in step S1 to calculate the scattering power P d (P d1 ), and the scattering power of the volume scattering P v (P v1 ) is calculated (step S2).

[0113] Next, the scattering model power decomposition unit 13 calculates the first difference ΔP d (The scattering power P of double reflection scattering when a swarm occurs d1 and the scattering power P of double reflection scattering when no flocking occurs. d0 ) and the second difference ΔP v (The scattering power P v1 and the scattering power P of volume scattering when no swarms are occurring. v0 The difference between the two is calculated (step S3).

[0114] Furthermore, the congestion degree acquisition unit 12 acquires, from the other system, the congestion degree calculated by the other system (step S4).

[0115] Next, the learning unit 14 calculates the density obtained in step S4 and the first difference ΔP d and the second difference ΔP v The neural network is trained by learning a weight matrix for each pair of adjacent layers of the neural network using (step S5). As already described, the number of layers of the neural network and the number of neurons in each layer are predetermined.

[0116] The crowding estimation device 10 performs steps S1 to S5 for each of the cases where an animal is specified as an individual and where a human is specified as an individual, and generates a neural network corresponding to the animal and a neural network corresponding to the human. In the process of step S2, branching processing is performed between the case where the individual is a human and the case where the individual is an animal. The process of step S2 will be described later.

[0117] 11 is a flowchart showing an example of the process flow when the density estimation device 10 of this embodiment estimates the density of individuals belonging to a flock. Descriptions of matters that have already been explained will be omitted as appropriate.

[0118] It is assumed that whether the individuals belonging to the group are humans or animals is specified in advance. In addition, the scattering power P d (P d0 ), and the scattering power of the volume scattering P v (P v0 ) is assumed to be calculated in advance.

[0119] The radar data acquisition unit 11 acquires polarized radar data from the polarized radar (step S11). This polarized radar data is data acquired by the polarized radar irradiating an area in which the density of individuals is to be estimated with radio waves.

[0120] Next, the scattering model power decomposition unit 13 calculates the scattering power P d (P d1 ), and the scattering power of the volume scattering P v (P v1 ) is calculated (step S12).

[0121] Next, the scattering model power decomposition unit 13 calculates the first difference ΔP d , and the second difference ΔP v is calculated (step S13).

[0122] Steps S12 and S13 are similar to the above-mentioned steps S2 and S3 (see FIG. 10).

[0123] Next, the estimation unit 15 calculates the first difference ΔP d , and the second difference ΔP v to the neural network to estimate the density of the individuals (step S14). When an animal is specified as an individual, the estimation unit 15 may use a neural network corresponding to the animal in step S14. When a human is specified as an individual, the estimation unit 15 may use a neural network corresponding to the human in step S14.

[0124] The population density estimation device 10 performs the processes of steps S11 to S14 for each region corresponding to each pixel. As a result, the population density of individuals is obtained for each pixel. After the population density of individuals is obtained for each pixel, the population map generation unit 17 generates a population map (see, for example, the lower part of FIG. 4) based on the density of each pixel. The method by which the population map generation unit 17 generates the population map has already been described, so the description will be omitted here. The population map generation unit 17 displays the population map on, for example, a display device (output unit 18).

[0125] As described above, step S2 (see FIG. 10) and step S12 (see FIG. 11) are similar processes, and in both cases, the scattering power P d and the scattering power of volume scattering P v This is the process of calculating the scattering power P d and the scattering power of volume scattering P v 12 and 13 show the process of calculating the scattering power P d and the scattering power of volume scattering P v 13 is a flowchart showing an example of the progress of a process for calculating the .times. ...

[0126] In the following explanation, 10log(<|SVV | 2 > / <|S HH | 2 >) is represented by the symbol L.

[0127] The scattering model power decomposition unit 13 calculates a scattering matrix (the first matrix on the right side of equation (4)) using equation (4) and the reception electric field strength and transmission electric field strength included in the polarized radar data (step S21).

[0128] Next, the scattering model power decomposition unit 13 calculates the covariance matrix <[C]> HV is calculated (step S22).

[0129] Next, scattering model power decomposition unit 13 judges whether the pre-specified individual is a human being or an animal (step S23).

[0130] If the pre-specified individual is a human being, the process proceeds to step S24.

[0131] In step S24, the scattering model power decomposition unit 13 calculates the value of L (i.e., 10log(<|S VV | 2 > / <|S HH | 2 >) to determine the compartment to which it belongs.

[0132] When the value of L is less than −3.5 dB, the scattering model power decomposition unit 13 calculates m 11 ,m 22 ,m 33 ,m 13 Then, 5 / 8, 1 / 4, 1 / 8, and 1 / 8 are respectively substituted for (step S25). After step S25, the process proceeds to step S32.

[0133] When the value of L is −3.5 dB or more and 3.5 dB or less, the scattering model power decomposition unit 13 calculates m 11 ,m 22 ,m 33 ,m 13Then, 3 / 8, 1 / 4, 3 / 8, and 1 / 8 are respectively substituted for (step S26). After step S26, the process proceeds to step S32.

[0134] When the value of L exceeds 3.5 dB, the scattering model power decomposition unit 13 calculates m 11 ,m 22 ,m 33 ,m 13 Then, 1 / 8, 1 / 4, 5 / 8, and 1 / 8 are substituted into the following (step S27). After step S27, the process proceeds to step S32.

[0135] If the pre-specified individual is an animal, the process proceeds to step S28 (see FIG. 13).

[0136] In step S28, the scattering model power decomposition unit 13 calculates the value of L (i.e., 10log(<|S VV | 2 > / <|S HH | 2 >) to determine the compartment to which it belongs.

[0137] When the value of L is less than −2 dB, the scattering model power decomposition unit 13 calculates m 11 ,m 22 ,m 33 ,m 13 After step S29, the process proceeds to step S32 (see FIG. 12).

[0138] When the value of L is −2 dB or more and 2 dB or less, the scattering model power decomposition unit 13 calculates m 11 ,m 22 ,m 33 ,m 13 Then, 3 / 8, 1 / 4, 3 / 8, and 1 / 8 are respectively substituted for (step S30). After step S30, the process proceeds to step S32.

[0139] When the value of L exceeds 2 dB, the scattering model power decomposition unit 13 calculates m 11 ,m22 ,m 33 ,m 13 Then, 1 / 5, 4 / 15, 8 / 15, and 2 / 15 are respectively substituted into (step S31). After step S31, the process proceeds to step S32.

[0140] When the process proceeds to step S32 (see FIG. 12), m 11 ,m 22 ,m 33 ,m 13 In step S32, the scattering model power decomposition unit 13 solves the simultaneous equations using the eight equations shown in equations (22), (25), and (26) to obtain the scattering power P d and the scattering power of volume scattering P v Calculate.

[0141] Steps S21 to S32 shown in Figs. 12 and 13 are an example of the progress of processing by scattering model power decomposition section 13 in step S2 (see Fig. 10) or step S12 (see Fig. 11).

[0142] According to this embodiment, the scattering model power decomposition unit 13 calculates the difference (first difference) between the scattering power of double reflection scattering when a swarm occurs and the scattering power of double reflection scattering when a swarm does not occur, and the difference (second difference) between the scattering power of volume scattering when a swarm occurs and the scattering power of volume scattering when a swarm does not occur. Then, the estimation unit 15 applies the first difference and the second difference to a neural network to estimate the density of individuals. Therefore, the density estimation device 10 can estimate the density of individuals while eliminating the influence of stationary objects.

[0143] Furthermore, the estimation unit 15 estimates the density of individuals without using data on the scattering power of surface scattering. Therefore, the density estimation device 10 can estimate the density of individuals while eliminating the influence of objects that cause surface scattering. Note that humans and animals do not fall under the category of objects that cause surface scattering.

[0144] Therefore, the density of individuals can be estimated with high accuracy.

[0145] When the individual is an animal, the scattering model power decomposition unit 13 converts a value obtained from a covariance matrix (see FIG. 7) that represents a model of the volume scattering of an animal into m 11 ,m 22 ,m 33 ,m 13 When the individual is a human, the scattering model power decomposition unit 13 substitutes the value obtained from the covariance matrix (see FIG. 9) representing the model of volume scattering of a human into m 11 ,m 22 ,m 33 ,m 13 Therefore, whether the density of animals or the density of people is to be estimated, the density can be estimated with high accuracy.

[0146] In the above embodiment, the learning model is a neural network. The learning model may be a learning model obtained by other machine learning techniques. For example, the learning model may be a learning model obtained based on linear regression or log-linear regression.

[0147] In the above embodiment, the estimation unit 15 calculates the first difference ΔP d and the second difference ΔP v A case has been described in which the density of individuals is estimated by applying the above to a learning model. The estimation unit 15 may estimate the density of individuals without using a learning model. In this case, the learning unit 14 does not need to be provided. Hereinafter, a case will be described in which the estimation unit 15 estimates the density of individuals without using a learning model.

[0148] Double scattering occurs in human and animal crowds when the density of individuals is low to moderate, and as the density increases, the scattering power of double scattering increases.

[0149] Furthermore, when the density of individuals is high, such as in a crowd of people or animals, the scattering power of double scattering becomes saturated and volume scattering occurs, and as the density increases, the scattering power of volume scattering increases.

[0150] Therefore, when the congestion is low to medium, the first difference ΔP d Hereinafter, this function will be referred to as the first function. When the density is high, the second difference ΔP v The density can be expressed by the function below. This function is called the second function.

[0151] The first function and the second function may be determined in advance. Then, the estimation unit 15 determines whether to estimate the density using the first function or the second function based on ΔP d / ΔP v That is, the estimation unit 15 determines the value of ΔP d / ΔP v Calculate the value of ΔP d / ΔP v If the value of is equal to or greater than a predetermined threshold (represented as Th), the estimation unit 15 estimates the first function and the first difference ΔP d The density can be estimated from the value of ΔP d / ΔP v If the value of is less than the threshold Th, the estimation unit 15 estimates the second function and the first difference ΔP v The density can be estimated from the value of

[0152] 14 is a schematic block diagram showing an example of the configuration of a computer relating to the congestion estimation device 10 of the embodiment of the present invention. The computer 1000 includes, for example, a CPU 1001, a main memory device 1002, an auxiliary memory device 1003, an interface 1004, a display device 1005, and a communication interface 1006.

[0153] A congestion estimation device 10 according to an embodiment of the present invention is realized by a computer 1000. The operation of the congestion estimation device 10 is stored in the form of a program (congestion estimation program) in an auxiliary storage device 1003. A CPU 1001 reads the program from the auxiliary storage device 1003, loads the program in a main storage device 1002, and executes the processing described in the above embodiment according to the program.

[0154] The auxiliary storage device 1003 is an example of a non-transient tangible medium. Other examples of non-transient tangible media include a magnetic disk, a magneto-optical disk, a CD-ROM (Compact Disk Read Only Memory), a DVD-ROM (Digital Versatile Disk Read Only Memory), and a semiconductor memory connected via the interface 1004. In addition, when a program is distributed to the computer 1000 via a communication line, the computer 1000 that receives the program may load the program into the main storage device 1002 and execute the processing described in the above embodiment according to the program.

[0155] In addition, some or all of the components may be realized by general-purpose or dedicated circuits, processors, etc., or a combination of these. These may be configured by a single chip, or may be configured by multiple chips connected via a bus. Some or all of the components may be realized by a combination of the above-mentioned circuits, etc., and programs.

[0156] When some or all of the components are realized by a plurality of information processing devices, circuits, etc., the plurality of information processing devices, circuits, etc. may be centrally or distributed. For example, the information processing devices, circuits, etc. may be realized as a client-server system, a cloud computing system, etc., in a form in which each is connected via a communication network.

[0157] Next, an overview of the present invention will be described. Fig. 15 is a block diagram showing an overview of the congestion degree estimation device of the present invention. The congestion degree estimation device of the present invention comprises a difference calculation means 73 and an estimation means 75.

[0158] The difference calculation means 73 (e.g., the scattering model power decomposition unit 13) calculates a first difference which is the difference between the scattering power of double reflection scattering when a flock of individuals has occurred and the scattering power of double reflection scattering when a flock of individuals has not occurred, and a second difference which is the difference between the scattering power of volume scattering when a flock of individuals has occurred and the scattering power of volume scattering when a flock of individuals has not occurred.

[0159] The estimation means 75 (eg, the estimation unit 15) estimates the density, which is the number of individuals per unit area, based on the first difference and the second difference.

[0160] With such a configuration, the density of the individual of interest can be estimated with high accuracy.

[0161] In addition, the difference calculation means 73 may be configured to calculate a first difference and a second difference when the individual is a human using a model of double reflection scattering and a model of volume scattering of a human flock, and the estimation means 75 may be configured to estimate the density when the individual is a human based on the first difference and the second difference.

[0162] The configuration may also include a population map generating means (for example, population map generating unit 17) that generates a population map, which is information that represents the number of individuals in each area in a map form based on the density of each area.

[0163] The estimation means 75 may be configured to estimate the density by applying the first difference and the second difference to a learning model that has been learned in advance.

[0164] The configuration may also include a learning means (e.g., the learning unit 14) that learns a learning model by machine learning using a known density and the first difference and the second difference corresponding to the known density.

[0165] The learning model may also be a neural network. [Industrial Applicability]

[0166] The present invention is preferably applied to a density estimation device that estimates the density of individuals. [Explanation of symbols]

[0167] 10 Density estimation device 11 Radar data acquisition section 12 Density acquisition part 13 Scattering model power decomposition part 14 Learning Department 15 Estimation part 17 Population map generator 18 Output section

Claims

1. a difference calculation means for calculating a first difference which is the difference between the scattering power of double reflection scattering calculated using the polarized radar data when a swarm of individuals has occurred and the scattering power of double reflection scattering calculated using the polarized radar data when a swarm of individuals has not occurred, and a second difference which is the difference between the scattering power of volume scattering calculated using the polarized radar data when a swarm of individuals has occurred and the scattering power of volume scattering calculated using the polarized radar data when a swarm of individuals has not occurred; and an estimation means for estimating a density, which is the number of individuals per unit area, based on the first difference and the second difference. A congestion estimation device comprising:

2. the difference calculation means calculates a first difference and a second difference when the individual is a human, using a double-reflection scattering model calculated using the polarized radar data and a volume scattering model of a human group calculated using the polarized radar data; The estimation means estimates a density when the individual is a human being based on the first difference and the second difference. The congestion estimation device according to claim 1 .

3. The system includes a population map generating means for generating a population map, which is information that represents the population of each area in a map form based on the density of each area. The congestion estimation device according to claim 1 or 2.

4. The estimation means estimates the density by applying the first difference and the second difference to a pre-trained learning model. The congestion estimation device according to any one of claims 1 to 3.

5. A learning means is provided for learning a learning model by machine learning using a known density and a first difference and a second difference corresponding to the known density. The congestion estimation device according to claim 4 .

6. The learning model is a neural network. The congestion estimation device according to claim 4 or 5.

7. The computer calculating a first difference which is the difference between the scattering power of double reflection scattering calculated using the polarized radar data when a flock of individuals has occurred and the scattering power of double reflection scattering calculated using the polarized radar data when a flock of individuals has not occurred, and a second difference which is the difference between the scattering power of volume scattering calculated using the polarized radar data when a flock of individuals has occurred and the scattering power of volume scattering calculated using the polarized radar data when a flock of individuals has not occurred; A density, which is the number of individuals per unit area, is estimated based on the first difference and the second difference. A method for estimating congestion, comprising:

8. The computer when calculating the first difference and the second difference, a model of double reflection scattering calculated using the polarized radar data and a model of volume scattering of a human group calculated using the polarized radar data are used to calculate the first difference and the second difference when the individual is a human; Estimating a density when the individual is a human based on the first difference and the second difference. The method of claim 7 .

9. On the computer, a difference calculation process for calculating a first difference which is the difference between the scattering power of double reflection scattering calculated using the polarized radar data when a flock of individuals has occurred and the scattering power of double reflection scattering calculated using the polarized radar data when a flock of individuals has not occurred, and a second difference which is the difference between the scattering power of volume scattering calculated using the polarized radar data when a flock of individuals has occurred and the scattering power of volume scattering calculated using the polarized radar data when a flock of individuals has not occurred; and An estimation process for estimating a density, which is the number of individuals per unit area, based on the first difference and the second difference. A density estimation program for executing the above.

10. On the computer, in the difference calculation process, a first difference and a second difference are calculated when the individual is a human being, using a model of double reflection scattering calculated using the polarized radar data and a model of volume scattering of a human group calculated using the polarized radar data; In the estimation process, a density in the case where the individual is a human being is estimated based on the first difference and the second difference. The congestion estimation program according to claim 9 .

Images