JP2001357386A - Method and device for detecting road surface state - Google Patents

Abstract

PROBLEM TO BE SOLVED: To remove fixing and movement disturbance main factor, etc., and also to correctly detect a road surface state. SOLUTION: A standard road surface state such as assumed new snow, pressed snow and drying, which consists of a photograph, etc., is photographed by a camera 5, the road surface image is processed by a controller 13 and an amplitude characteristic is made into an image as a density and displayed on a monitor 19. Three feature values are obtained, that is, the amplitude feature value of the road surface image, a difference between the FFT amplitude feature values of the actual road surface image and the standard road surface state and a difference between the maximum and the minimum values from the R, G and B values of the road surface image. Then the normal distribution curve of each standard road surface state is calculated and a standard table is generated. The actual road surface image photographed by the camera is FFT-processed and the amplitude characteristic is made into an image as the density and displayed on the monitor. Then the actual feature values are obtained, i.e., the amplitude feature value of the actual road surface image, the difference between the FFT amplitude feature values of the actual road surface image and the standard road surface state and the difference between the maximum and minimum values from the R, G and B values of the road image. The actual feature values are fetched into the standard table to calculate each probability and to obtain an integrated probability. Then each road surface state is detected based on the integrated probability.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method and a device for detecting a road surface condition, and more particularly, to a method for detecting a snow compaction formed on a road surface which causes a decrease in a traveling speed and a traffic volume of a car in a snowy country in winter. The present invention relates to a method and a device for detecting a road surface condition for measurement.

[0002]

2. Description of the Related Art As a conventional road surface condition detecting device, there is known a device disclosed in Japanese Patent Publication No. 4-76638 and Japanese Patent Publication No. 5-61598. That is, as shown in FIG. 15, in the road surface condition detecting device 101, the measuring unit 107 rotates up and down via the mounting bracket 105 on the upper part of the column 103 erected on the ground GL outside the road. It is provided freely.

The measuring unit 107 is provided with a snow gauge, a snow gauge, and the like (not shown). The light wave transmitter projects light waves toward the snow surface, and the light reflected from the snow surface is received by the light wave receiver. Then, the depth (height) of the snow on the ground GL is measured, and the snowfall is detected. Then, the road surface condition such as the compacted snow is detected by judging the measurement results in combination.

[0004]

By the way, in the road surface condition detected by the above-mentioned conventional road surface condition detecting device, a fixed disturbance factor, a moving disturbance factor and the like cannot be removed, so that an accurate road surface condition can be detected. could not. Further, the conventional road surface condition detecting device includes devices such as a snow gauge and a snowfall meter, and the device itself has become large-scale.

SUMMARY OF THE INVENTION The present invention has been made to solve the above-mentioned problems, and an object of the present invention is to eliminate a fixed disturbance factor and a moving disturbance factor, and to detect an accurate road surface condition.
It is an object of the present invention to provide a road surface condition detection method and a device in which the device itself is made compact.

[0006]

According to a first aspect of the present invention, there is provided a method for detecting a road surface condition according to the present invention, which comprises the steps of assuming fresh snow, condensed snow, freezing, wetness and dryness comprising actual road surface conditions and photographs. The standard road surface condition is photographed by the monitoring camera, the photographed road surface image is processed by the FFT by the control device, the amplitude characteristic is converted into a density, and the amplitude characteristic value of the road surface image and the actual The difference between the road surface image and the FFT amplitude feature value of the standard road surface condition and the R of the road surface image,
Three characteristic values, which are the difference between the maximum value and the minimum value from the G and B values, are obtained, and a normal distribution curve of each standard road surface condition is calculated from each of the obtained characteristic values to prepare a standard table in advance. The surveillance camera shoots an actual road surface condition, the shot actual road surface image is processed by an FFT by an image processing device, and the amplitude characteristic is imaged as a density. Three actual feature values comprising the difference between the amplitude feature value, the difference between the actual road surface image and the FFT amplitude feature value of the standard road surface condition, and the difference between the maximum value and the minimum value from the R, G, and B values of the road surface image, This actual characteristic value is taken into the standard table to calculate each probability and obtain the total probability. Based on the obtained total probability, it is detected whether it is fresh snow, compact snow, freezing, wet, or dry. Is characterized by .

[0007] Accordingly, the surveillance camera captures the assumed standard road surface conditions of fresh snow, condensed snow, freezing, wetness and dryness, including actual road surface conditions and photographs. Then, the captured road surface image is processed by the FFT by the control device, and the amplitude characteristic is visualized as a density. Based on the density, the amplitude characteristic value of the road surface image, the actual road surface image and the FFT of the standard road surface condition are obtained. Three feature values, which are the difference between the amplitude feature values and the difference between the maximum value and the minimum value from the R, G, and B values of the road surface image, are obtained.
A normal distribution curve for each standard road surface condition is calculated from the obtained characteristic values, and a standard table is created in advance.

[0008] Then, an actual road surface condition is photographed by the surveillance camera, and the photographed actual road surface image is processed by an FFT by an image processing device, and an amplitude characteristic is visualized as a density. In addition, there are three actual values of the amplitude characteristic value of the actual road surface image, the difference between the FFT amplitude characteristic value of the actual road surface image and the FFT amplitude characteristic value of the standard road surface condition, and the difference between the maximum value and the minimum value from the R, G, and B values of the road surface image. Are obtained.

[0009] The actual characteristic values are fetched into the standard table to calculate respective probabilities and obtain the overall probabilities. Based on the obtained total probability, it is possible to accurately, easily and easily detect whether any one of fresh snow, condensed snow, freezing, wetness, and dryness.

According to a second aspect of the present invention, in the road surface condition detecting method according to the first aspect, when there is a fixed disturbance or a moving disturbance while the surveillance camera is capturing the actual road condition, It is characterized in that fixed disturbance removal processing and moving disturbance removal processing are performed.

Therefore, when there is a fixed disturbance and a moving disturbance while the monitoring camera is capturing the actual road surface condition, the fixed disturbance removing process and the moving disturbance removing process are performed, respectively.
By removing the fixed disturbance and the moving disturbance, it is possible to more accurately, easily and easily detect whether any one of fresh snow, condensed snow, freezing, wetness, and dryness.

According to a third aspect of the present invention, there is provided a road surface condition detecting device for photographing a road surface condition with a monitoring camera provided on an upper portion of the column while illuminating the illumination device provided on a column standing on the side of the road. After image processing the captured road surface image with an image processing device, the image processed video data is taken into a control device, and based on the actual road surface condition, fresh snow, condensed snow, freezing, wet, and dry A road surface condition detection device that detects any one of the following: the control device, the surveillance camera preliminarily actual road surface conditions and assumed assumed fresh snow, snow compaction, freezing, wet and dry standard road surface conditions. A road surface condition processing image monitor for taking an image with a surveillance camera, processing the captured road surface image by FFT, converting the amplitude characteristics into a density image, and displaying the image, and an image displayed on the road surface condition processing image monitor Three features consisting of the amplitude feature value of the road surface image based on the density, the difference between the amplitude feature value of the actual road surface image and the amplitude feature value of the standard road surface condition, and the difference between the maximum value and the minimum value from the R, G, and B values of the road surface image A memory for obtaining and storing values, a standard table for calculating and filing a normal distribution curve of each characteristic value based on the three characteristic values stored in the memory, Measured road surface condition The road surface image of the standard road surface condition is processed by FFT, the amplitude characteristics are visualized as density, and the amplitude characteristic value of the road surface image, the difference between the actual road surface image and the amplitude characteristic value of the standard road surface condition, and the road surface image The three actual feature values, which are the difference between the maximum value and the minimum value from the R, G, and B values, are taken into the normal distribution curve of each feature value calculated from the standard table, and each probability is calculated. Seeking fresh snow, pressure , Frozen, it is to wet, a discrimination unit for discriminating one of drying, characterized in that it comprises a.

[0013] Therefore, the same as in the first aspect, the assumed standard road surface conditions of fresh snow, condensed snow, freezing, wetness and dryness consisting of actual road surface conditions and photographs are photographed by the surveillance camera. Then, the captured road surface image is processed by the FFT by the control device, and the amplitude characteristic is visualized as a density, and the density is displayed on a road surface condition processing image monitor. The amplitude characteristic value of the road surface image and the actual road surface image are displayed. And standard road condition FFT
Three feature values, which are the difference between the amplitude feature values and the difference between the maximum value and the minimum value from the R, G, and B values of the road surface image, are obtained. A normal distribution curve for each standard road surface condition is calculated from the obtained characteristic values, and a standard table is created in advance.

An actual road surface condition is photographed by a monitoring camera provided above the pillar while illuminating with an illumination device provided on the pillar, and the photographed actual road surface image is processed by an image processing device. The FFT process is performed to image the amplitude characteristics as a density, and the density is displayed on a road surface condition processing image monitor. The amplitude characteristic value of the actual road surface image and the FFT amplitude characteristic value of the actual road surface image and the standard road surface condition are displayed. Three actual feature values, which are the difference and the difference between the maximum value and the minimum value from the R, G, B values of the road surface image, are obtained.

The actual characteristic values are fetched into the standard table to calculate respective probabilities and obtain a total probability. Based on the obtained total probability, it is possible to accurately, easily and easily detect whether any one of fresh snow, condensed snow, freezing, wetness, and dryness.

According to a fourth aspect of the present invention, in the road surface condition detecting device according to the third aspect, the control device includes a fixed disturbance removing process determining device and a moving disturbance removing process determining device. It is characterized by the following.

Therefore, as in the second aspect, when there is a fixed disturbance or a moving disturbance while the actual surveillance camera is photographing the actual road surface condition, the fixed disturbance removing process and the moving disturbance removing process are broken, respectively. By removing the fixed disturbance and the moving disturbance, it is possible to more accurately, easily and easily detect whether any one of fresh snow, condensed snow, freezing, wetness, and dryness.

[0018]

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a block diagram showing an embodiment of a road surface condition detecting method and apparatus according to the present invention.

Referring to FIG. 1, a road surface condition detecting device 1
Has a column 3 standing upright on the ground GL beside the road surface R, and a camera housing 7 having an example of an ITV camera 5 as a surveillance camera is provided above the column 3 in a range of, for example, 45 degrees. It is installed so that it can swing. An illumination device 9 is attached to the column 3 below the camera housing 7.

According to the above configuration, the lighting device 9 illuminates the road surface condition detection area RS of the road surface R with light. Then, in this state, the road surface condition of the road surface condition detection area RS is photographed by the ITV camera 5. That is,
The road surface condition in the road surface condition detection area RS is received by the ITV camera 5 with the reflected light amount, transferred to the image processing device 11 and subjected to image processing, and then sent to the control device 13 where the road surface condition is detected.

The control device 13 has a CPU 15 as shown in FIG. 2. The CPU 15 is connected to an input means 17 for inputting various data with a keyboard, a touch sensor or the like. A display means 19 as an image processing monitor such as a CRT is connected.

The CPU 15 has the image processing device 11
, The difference m_FFT_D between the actual road surface image and the FFT amplitude characteristic value of the standard road surface condition, R (red), G (green), and B (blue) 3) Three characteristic value / memory 21 for storing a difference RGB_dif between the maximum value and the minimum value from the value are connected. In addition, the CPU
A fixed disturbance removal processing device 23 and a moving disturbance removal processing device 25 are connected to 15. Further, a standard table file 27 and a discrimination device 29 are connected to the CPU 15.

Next, the road surface condition detecting operation will be described with reference to the flowcharts shown in FIGS. 3 and 4. First, in FIG. 3, after the initialization processing is performed in step S1, the determination table is determined in step S2. A determination is made whether there is. If there is a judgment table, the process proceeds to step S3, and the standard table is loaded from the standard table file 27. If there is no judgment table, a standard table is created in step S4. This standard table is created from a photograph or the like which has been taken in advance or an actual road surface condition. That is, for example,

M_FFT_Z table Fresh snow: m_FFT_SN_tbl [250] Snow: m_FFT_YA_tbl [250] Freezing: m_FFT_TO_tbl [250] Drying: m_FFT_KN_tbl [250] Wet: m_FFT_SI_tbl [250]

[Expression 2] RGB_dif table Fresh snow: m_COL_SN_tbl [250] Snow: m_COL_YA_tbl [250] Freezing: m_COL_TO_tbl [250] Drying: m_COL_KN_tbl [250] Wet: m_COL_SI_SI_tbl [250]

Equation 1 shows a table of amplitude characteristic values m_FFT_Z.

In the equation (2), a table of a difference RGB_dif between the maximum value and the minimum value from the R (red), G (green) and B (blue) values is filed in the standard table file 27.

[0024]

## EQU3 ## m_FFT_Table (1) When the fresh snow situation is based on the FFT difference: fresh snow → new snow: m_D_SN_tbl. m_sn [25
0] fresh snow-> compact snow: m_D_SN_tbl. m_ga [25
0] fresh snow → freeze: m_D_SN_tbl. m_to [25
0] fresh snow-> dry: m_D_SN_tbl. m_kn [25
0] fresh snow-> wet: m_D_SN_tbl. m_si [25
0] (2) When the snow compaction condition is based on the FFT difference: compaction snow → new snow: m_D_YA_tbl. m_sn [25
0] compacted snow-> compacted snow: m_D_YA_tbl. m_ya [25
0] compacted snow-> freeze: m_D_YA_tbl. m_to [25
0] pressed snow-> dry: m_D_YA_tbl. m_kn [25
0] snow-> wet: m_D_YA_tbl. m_si [25
0] (3) When the freezing condition is based on the FFT difference: frozen-> fresh snow: m_D_TO_tbl. m_sn [25
0] Freeze-> Pressed snow: m_D_TO_tbl. m_ya [25
0] Freeze-> Freeze: m_D_TO_tbl. m_to [25
0] Freeze-> Dry: m_D_TO_tbl. m_kn [25
0] Freeze-> wet: m_D_TO_tbl. m_si [25
0] (4) When the wet condition is based on the FFT difference: wet-> fresh snow: m_D_SI_tbl. m_sn [25
0] wet-> pressure snow: m_D_SI_tbl. m_ya [25
0] wet-> freeze: m_D_SI_tbl. m_to [25
0] wet-> dry: m_D_SI_tbl. m_kn [25
0] wet-> wet: m_D_SI_tbl. m_si [25
0] (5) When the drying condition is based on the FFT difference: drying-> fresh snow: m_D_KN_tbl. m_sn [25
0] Drying-> Pressed snow: m_D_KN_tbl. m_ya [25
0] Dry-> Freeze: m_D_KN_tbl. m_to [25
0] Dry-> Dry: m_D_KN_tbl. m_kn [25
0] dry-> wet: m_D_KN_tbl. m_si [25
0] above

Equation 3 shows the difference m_FFT_D between the actual road surface image and the FFT amplitude characteristic value between the standard road surface condition and the data when the fresh snow condition, the snow compaction condition, the freezing condition, the wet condition, and the dry condition are each based on the FFT difference. Have been.

[0025]

[Equation 4] Each road surface condition establishment calculation table (A) (1) From m_FFT_Z Fresh snow: PJ_SN_FFT = m_FFT_SN_tbl [m_FFT_Z (i)] Snowfall: PJ_YA_FFT = m_FFT_YA_tbl [m_FFT_Z (i)] Freezing: PJ_F_F_F_Z ] Dry: PJ_KN_FFT = m_FFT_KN_tbl [m_FFT_Z (i)] Wet: PJ_SI_FFT = m_FFT_SI_tbl [m_FFT_Z (i)] (2) From RGB_dif Fresh snow: PJ_SN_COL_m_i__ P _______ = m_COL_J ___ = Pulse Freezing: PJ_TO_COL = m_COL_TO_tbl [RGB_dif (i)] Drying: PJ_KN_COL = m_COL_KN_tbl [RGB_dif (i)] Wet: PJ_SI_COL = m_COL_SI_tbl [RGB_dif (i)] (from snow_F_F_D_F_F_D_F_F_D_F_D_F_F_D_F_D_F_D_F_D_F_D_F_F_D_F_D_F_F_D_F_F_D_F_D_F_D_F_D_F_F_F_D_F_F_D_F_F_D_F_F_D_F_F_D_F_F_F_D_F_F_D_F_F_D_F_F_D_F_F_F_D_FFT_F_D_FFT_F_F_F_D_FFT : PJ_YA_D = m_FFT_YA_tbl [m_FFT_D (i)] Freezing: PJ_TO_D = m_FFT_TO_tbl [m_FFT_D (i)] Drying: PJ_KN_D = m_FFT_KN_tbl [m_FFT_D (_) _ F_SI__________________________________ Of the above

Equation 4 shows m_F as each road surface condition probability calculation table.
Fresh snow, compact snow, frozen, wet, and dry data from FT_Z, RGB_dif, and m_FFT_D are shown.

[0026]

[Equation 5] Total probability: Fresh snow: PJ_SN (i) = PJ_SN_FFT * 0.5 + PJ_SN_D * 0.3 + PJ_SN
_COL * 0.2 Pressed snow: PJ_YA (i) = PJ_YA_FFT * 0.5 + PJ_YA_D * 0.3 + PJ_YA
_COL * 0.2 Freeze: PJ_TO (i) = PJ_TO_FFT * 0.5 + PJ_TO_D * 0.3 + PJ_TO
_COL * 0.2 Wet: PJ_SI (i) = PJ_SI_FFT * 0.5 + PJ_SI_D * 0.3 + PJ_SI
_COL * 0.2 Drying: PJ_KN (i) = PJ_KN_FFT * 0.5 + PJ_KN_D * 0.3 + PJ_KN
_COL * 0.2 PJ_MAX = max (PJ_SN (i), PJ_YA (i), PJ_TO (i), PJ_SI
(i), PJ_KN (i)) IF PJ_MAX = PJ_SN (i) Road surface condition ==> fresh snow IF PJ_MAX = PJ_YA (i) Road surface condition ==> consolidation IF PJ_MAX = PJ_TO (i) Road surface condition ==> frozen IF PJ_MAX = PJ_SI (i) Road surface condition ==> wet IF PJ_MAX = PJ_KN (i) Road surface condition ==> dry

[Equation 5] shows the total probability calculation table as fresh snow, compact snow,
Freezing, wetting, and drying data are shown along with the conditions.

Referring again to FIG. 3, the inspection is started in step S5, and the processed image capture di is started in step S6.
f_flag = 0, and the buffer read_in step S7
dif (). Then, in step S8, Dif_f
It is determined whether lag is 0 or 1, and if it is 0, step S9
In step S10, it is set as dif_flag = 1 as a difference image capture. In the case of 1, the process proceeds to step S12 as a single capture in step S11.

In step S12, a time difference process is performed.
You. In step S13, vehicles and pedestrians move in the measurement area.
When there is a moving object such as an object, dif_V (Δt) = [fr
am-dif_ (t_{i + 1}) -Frame-init
(T_i)] + 0 × 80, Δt = t_{i + 1}-T _i(About 50m
s), when Dif> 135, a fixed disturbance
Then, the process returns to step S6. And
If Dif> 135, it is determined that there is no fixed disturbance.
Proceed to step S14.

In step S14, noise of the image is removed by a nodian filter, and contour extraction is performed in step S15. That is, when there is a fixed object such as a parked vehicle in the measurement area, the contour of the object is extracted. Step S
At 16 the binarization is performed. That is, the image whose contour has been extracted is binarized using the brightness of the video as a threshold.

Next, in step S17 in FIG. 4, the upper binarized image is label-linked, the number of label links is calculated, and the area of the label-linked image is calculated. Step S
At 18, the maximum value of the area is obtained, and the maximum value of the area is compared with a previously determined fixed disturbance determination value S_max.
If the area> S_max, it is determined that there is a movement disturbance, and the process returns to the step before step S6. And area> S_ma
If it is not x, it is determined that there is no movement disturbance and step S1
Go to 9.

In step S19, an FFT process is performed.
That is, the road surface image is formed into an image with the amplitude characteristic as the density by the FFT processing, and is displayed on the display means 19 as a road surface condition processing image monitor as shown in FIG. The determination characteristic value of the road surface condition is obtained by the FFT processing. The main parameters are m_FFT_Z and m_FFT_D.

That is, m_FFT_Z is an amplitude characteristic value of a road surface image and is obtained by the following equation.

M_FFT_Z Standard road surface condition = A × √X ² + Y ² However, the amplitude characteristic value is the density value of point A imaged.

As shown in FIG. 6, m_FFT_D is the difference between the actual road surface image and the amplitude characteristic value of the FFT difference reference road surface condition, and is obtained by the following equation.

M_FFT_D = | ΣP (i, j) -ΣQ (i, j) | In step S20, color processing is performed. That is, it is one of the parameters for determining the fixed disturbance using the color information of the road surface image.

RGB_max = max (R, G, B) RGB_min = min (R, G, B) RGB_dif = RGB_max-RGB_min If Col_Dif> Rd, it is determined that there is a fixed disturbance, and the process returns to before step S6. It is. And C
If ol_Dif> Rd is not satisfied, it is determined that there is no fixed disturbance, and the process proceeds to step S21. In step S21, the road surface condition is determined.

That is, when the road surface conditions are fresh snow, compact snow, freezing,
Whether it is wet or dry is determined from the numerical feature value of the image processing. The determination method compares the characteristic value of the measured road surface image with the characteristic value of the standard road surface, and calculates and determines the probability of what road surface condition the measured road surface is. As described above, for the standard road surface condition, a standard data table is created from photographs and actual road surfaces, and the standard data table assumes a normal distribution of the selected feature values.

First, a normal distribution curve of three characteristic values m_FFT_Z, m_FFT_D, and RGB_dif obtained from the standard road surface condition is calculated, and a standard data table is created in advance. Then, the characteristic value of the actual road surface condition is taken into the standard data table, and the probability of the road surface condition is obtained.
Further, the overall probability is obtained from the probabilities of the three feature values. Then, in step S22, the road surface condition is determined for the maximum probability among the three characteristic values. That is, when the values of the maximum probabilities are calculated for fresh snow, condensed snow, freezing, wet, and dry, respectively, it is determined to be fresh snow, condensed snow, freezing, wet, and dry. Then, as an on-timer loop, step S5
And if the road surface table is updated in step S23, the process returns to step S4 and restarts.

Next, when the result output is executed in step S24, graph drawing, Logfile saving, and Bit output are performed in steps S25, S26, and S27, respectively.

Next, from step S21 to step S27
A specific determination method in the flow up to will be described. For example, there are photos of three standard road conditions: assumed fresh snow, compacted snow, and dry. Image processing is performed on these three photographs, and the three characteristic values are as follows.

M_FFT_Z: Fresh snow: m_FFT_Z. SN = 10 pressed snow: m_FFT_Z. YA = 40 Dry: m_FFT_Z. KN = 120 m_FFT_D: (FFT based on assumed fresh snow road conditions)
The difference between the values of is m_FFT_D. ) Fresh snow → fresh snow: m_FFT_D. SN_SN = 0 Fresh snow → consolidation snow: m_FFT_D. SN_YA = 20 Fresh snow → Drying: m_FFT_Z. SN_KN = 40 RGB_dif: Fresh snow: RGB_dif. SN = 0 pressed snow: RGB_dif. YA = 1 Dry: RGB_dif. KN = 5 Then, a normal distribution curve of the above characteristic values is calculated, and a standard data table is created in advance.

[0042]

(Equation 6)

(Equation 7)

(Equation 8)

(Equation 9)

(Equation 10)

[Equation 11]

(Equation 12)

(Equation 13)

[Equation 14] A graph of the normal distribution curve calculated above is as shown in FIG. 7, for example. In FIG. 7, D
Indicates the width of the normal distribution curve. Moreover, the normal distribution curve of m_FFT_Z is, for example, as shown in FIG.

M_FFT_Z = 20, D =
The 30 road surface state probabilities are calculated as follows:

[0044]

(Equation 15)

(Equation 16)

[Expression 17] Drying: m FFT D. KN-D = 120-30 = 90>0; m FFT D. KN + D = 120 + 30 = 150; 90 <= X <= 150; Z K (20) = 0 Then, when determining the probability from M_FFT_D, becomes as shown in FIG.

M_FFT_D = 15, D =
The 30 road surface state probabilities are calculated as follows:

[0046]

(Equation 18)

[Equation 19] New snow → dry: D _{S — K} (15) = 0 The normal distribution curve of RGB_dif is as shown in FIG. 10, for example.

Assumed actual measurement road surface RGB_dif = 3, D =
The 30 road surface state probabilities are calculated as follows:

[0048]

(Equation 21)

(Equation 22)

(Equation 23) The overall probability is calculated by multiplying the probability obtained from the above three characteristic values by a weighting factor. The weighting factors are as follows, for example.

M_FFT_Z: 0.5 m_FFT_D: 0.3 RGB_dif: 0.2 Overall probability Fresh snow: PJ_SN = 0.5 × Z _S (20) + 0.3 × D _{S — S} (15) + 0.2 × R _S (3) = 0.5 × 0.75 + 0.3 × 0.5 + 0.2 × 0.98 = 0.72 Snow: PJ_YA = 0.5 × Z _Y (20) + 0.3 × D _{S — Y} (15 ) + 0.2 × R _Y (3) = 0.5 × 0.25 + 0.3 × 0.93 + 0.2 × 0.99 = 0.60 Drying: PJ_KN = 0.5 × Z _K (20) +0.3 × D _{S — K} (15) + 0.2 × R _K (3) = 0.5 × 0 + 0.3 × 0 + 0.2 × 0.99 = 0.2 When the maximum values of the above three probabilities are obtained, Therefore, since the value of PJ_MAX is the probability of fresh snow, the road surface condition is determined to be fresh snow.

Moreover, the on-timer loop is executed every 10 seconds to determine the road surface condition, and the final result is a one-hour determination result. This result is a result obtained by collecting the determination results of 10 seconds in the on-timer loop and obtaining the maximum value.

For example, as shown in FIG. 11, the data is output as a graph or a log file (Log).
file) or bit.

For example, a log file (Logfile)
One log file is created for the road surface determination result every hour from 9:00 in the morning to 8:00 in the morning on the following day.

Log file location: IDS_ROMDATA_PATH: C: \ WINN \ Profile \ Administr
ator \ Desktop \ ROM data Log file name: LogYYYYMMDDHH. For example, it is output as txt bits as follows.

[0054] Therefore, it is possible to accurately, easily and easily detect whether it is fresh snow, condensed snow, freezing, wet or dry. In addition, when the ITV camera 5 captures an actual road surface condition, when there is a fixed disturbance and a moving disturbance, the fixed disturbance removing process and the moving disturbance removing process are performed, respectively.
By removing the fixed disturbance and the moving disturbance, it is possible to more accurately, easily and easily detect whether any one of fresh snow, condensed snow, freezing, wetness, and dryness.

The display means 19 of the road surface condition detecting device 1 reflects a standard table screen, a measurement screen, and a display screen of snow compaction as shown in FIGS. 12, 13 and 14, for example. This is a great help for those who measure. Further, various other data and an operation screen at the time of measurement can be displayed on the display unit 19.

The present invention is not limited to the above-described embodiment, but can be embodied in other modes by making appropriate changes.

[0057]

As will be understood from the above description of the embodiments of the present invention, according to the first aspect of the present invention, it is assumed that fresh snow, compact snow, freezing,
Standard wet and dry road conditions are captured by surveillance cameras. Then, the captured road surface image is FFed by the control device.
T, image the amplitude characteristic as a density, and based on this density, the amplitude characteristic value of the road surface image, the difference between the FFT amplitude characteristic value of the actual road surface image and the FFT amplitude characteristic value of the standard road surface condition, and the R, G, Three feature values consisting of the difference between the maximum value and the minimum value from the B value are obtained. A normal distribution curve for each standard road surface condition is calculated from the obtained characteristic values, and a standard table is created in advance.

Then, an actual road surface condition is photographed by the surveillance camera, and the photographed actual road surface image is processed by an FFT by an image processing apparatus, and the amplitude characteristic is visualized as a density. In addition, there are three actual values of the amplitude characteristic value of the actual road surface image, the difference between the FFT amplitude characteristic value of the actual road surface image and the FFT amplitude characteristic value of the standard road surface condition, and the difference between the maximum value and the minimum value from the R, G, and B values of the road surface image. Are obtained.

The actual characteristic values are fetched into the standard table, each probability is calculated, and the total probability is obtained. Based on the obtained total probability, it is possible to accurately, easily and easily detect whether any of fresh snow, condensed snow, freezing, wetness, and drying.

According to the second aspect of the present invention, when there is a fixed disturbance or a moving disturbance while the actual surveillance camera is photographing the actual road surface condition, the fixed disturbance removing process and the moving disturbance removing process are respectively broken and the fixed By removing the disturbance and the movement disturbance, it is possible to more accurately, easily and easily detect whether any of fresh snow, condensed snow, freezing, wetness, and dryness.

According to the third aspect of the present invention, it is the same as the first aspect, and the surveillance camera shoots the actual road surface conditions and the assumed standard road surface conditions of fresh snow, condensed snow, freezing, wetness and dryness which are composed of photographs and the like. Is done. Then, the captured road surface image is processed by the FFT by the control device, and the amplitude characteristic is visualized as a density, and the density is displayed on a road surface condition processing image monitor. The amplitude characteristic value of the road surface image and the actual road surface image are displayed. Of the FFT amplitude characteristic value of the road surface condition and the standard road surface condition, and R of the road surface image,
Three feature values, which are the difference between the maximum value and the minimum value from the G and B values, are obtained. A normal distribution curve for each standard road surface condition is calculated from the obtained characteristic values, and a standard table is created in advance.

Then, while being illuminated by the illumination device provided on the column, an actual road surface condition is photographed by a monitoring camera provided on the column, and the photographed actual road surface image is converted into an image processing device. The FFT process is performed to image the amplitude characteristics as a density, and the density is displayed on a road surface condition processing image monitor. The amplitude characteristic value of the actual road surface image and the FFT amplitude characteristic value of the actual road surface image and the standard road surface condition are displayed. Three actual feature values, which are the difference and the difference between the maximum value and the minimum value from the R, G, B values of the road surface image, are obtained.

The actual characteristic values are fetched into the standard table to calculate respective probabilities and obtain the overall probabilities. Based on the obtained total probability, it is possible to accurately, easily and easily detect whether any of fresh snow, condensed snow, freezing, wetness, and dryness.

According to a fourth aspect of the present invention, in the same manner as the second aspect, while the actual surveillance camera captures an actual road surface condition,
When there is a fixed disturbance or a moving disturbance, the fixed disturbance removing process and the moving disturbance removing process are performed, respectively, and the fixed disturbance and the moving disturbance are removed. It can be detected accurately, easily and easily.

[Brief description of the drawings]

FIG. 1 is a perspective view of a road surface condition detecting device according to the present invention.

FIG. 2 is a configuration block diagram of a control device.

FIG. 3 is a flowchart of road surface condition detection.

FIG. 4 is a flowchart of road surface condition detection.

FIG. 5 is an explanatory diagram for obtaining m_FFT_Z.

FIG. 6 is an explanatory diagram for obtaining m_FFT_D.

FIG. 7 is a diagram illustrating an example of a normal distribution curve.

FIG. 8 is a diagram illustrating an example of a normal distribution curve of m_FFT_Z.

FIG. 9 is a diagram illustrating an example of a normal distribution curve of m_FFT_D.

FIG. 10 is a diagram illustrating an example of a normal distribution curve of RGB_dif.

FIG. 11 is a diagram illustrating an example of a result of detecting a road surface condition;

FIG. 12 is a diagram illustrating an example in which a standard table screen is displayed.

FIG. 13 is an example of a display screen of a measurement screen.

FIG. 14 is a diagram illustrating an example in which a display screen of compacted snow is displayed.

FIG. 15 is a side view of a conventional road surface condition detecting device.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Road surface condition detection device 3 Support 5 ITV camera (surveillance camera) 9 Illumination device 11 Image processing device 13 Control device 19 Display means 21 Three feature values / memory 23 Fixed disturbance elimination device 25 Moving disturbance elimination device 27 Standard table file 29 Identification device

 ──────────────────────────────────────────────────続 き Continued on the front page F term (reference) 2G059 AA05 BB08 EE13 FF01 GG00 KK04 MM01 MM02 MM05 MM09 MM10 PP04 5B057 AA16 BA02 CA01 CA08 CA12 CA16 CE02 CE12 DA12 DB02 DB06 DB09 DC04 DC16 DC23 DC30 DC32 5H180 ACOA EE BA04 CA04 DA02 EA05 EA43 FA06 FA23 FA37 FA59 GA08 GA30 GA40 JA11 JA22

Claims (4)

[Claims] 1. A standard camera condition of fresh snow, condensed snow, freezing, wetness and dryness consisting of actual road surface conditions and photographs is photographed by a monitoring camera, and the photographed road surface image is processed by an FFT by a control device. Then, the amplitude characteristic is visualized as a density, and based on this density, the amplitude characteristic value of the road surface image, the difference between the FFT amplitude characteristic value of the actual road surface image and the FFT amplitude characteristic value of the standard road surface condition, and the R, G, B values of the road surface image are obtained. The three characteristic values consisting of the difference between the maximum value and the minimum value are calculated, the normal distribution curve of each standard road surface condition is calculated from the obtained characteristic values, and a standard table is created in advance. The road condition of the road surface is photographed, the photographed actual road surface image is processed by the FFT by the image processing device, the amplitude characteristic is imaged as a density, and based on the density, the amplitude characteristic value of the actual road surface image, Actual road image and standard road The three actual characteristic values, which are the difference between the FFT amplitude characteristic value of the surface condition and the difference between the maximum value and the minimum value from the R, G, and B values of the road surface image, are obtained, and these actual characteristic values are taken into the standard table. A road surface condition detection method characterized by calculating each probability and calculating a total probability, and detecting any one of fresh snow, consolidation snow, freezing, wetness, and dryness based on the calculated total probability. . 2. The fixed disturbance removing process and the moving disturbance removing process are respectively performed when there is a fixed disturbance and a moving disturbance while the monitoring camera is capturing an actual road surface condition. Road condition detection method. 3. A road surface condition is photographed by a surveillance camera provided on an upper part of the column while illuminating with a lighting device provided on a column standing on the side of the road, and the photographed road surface image is imaged. After the image processing by the processing device, the image data obtained by the image processing is taken into the control device, and based on the actual road surface condition, the road surface condition detection for detecting whether it is fresh snow, condensed snow, freezing, wet, or dry. A device, wherein the control device captures, with a surveillance camera, a standard road surface condition of a presumed fresh snow, snow compaction, freezing, wetness and dryness which is made up of a real road surface condition and a photograph in advance with a surveillance camera. A road surface condition processing video monitor that processes a road surface image by FFT and visualizes and displays the amplitude characteristics as density, and an amplitude feature value of the road surface image based on the density of the image displayed on the road surface condition processing image monitor. of A memory for obtaining and storing three characteristic values including a difference between amplitude feature values of a road surface image and a standard road surface condition and a difference between a maximum value and a minimum value of R, G, and B values of a road surface image, and a memory for storing the three characteristic values. A standard table for calculating and filing a normal distribution curve of each characteristic value based on the stored three characteristic values, and a road surface image of the standard road surface condition actually measured by the surveillance camera by FFT. Process, image the amplitude characteristics as density, and convert the amplitude characteristic value of the road surface image, the difference between the amplitude characteristic value of the actual road surface image and the amplitude characteristic value of the standard road surface condition, and the maximum value and the minimum value from the R, G, and B values of the road surface image. The three actual feature values consisting of the difference are taken into the normal distribution curve of each feature value calculated from the standard table to calculate each probability and obtain the overall probability to obtain any one of fresh snow, compact snow, freezing, wet, and dry. Discriminator for discriminating If, road surface condition detecting apparatus characterized by including the. 4. The road surface condition detecting device according to claim 3, wherein the control device includes a fixed disturbance removing process determining device and a moving disturbance removing process determining device.