JP5770661B2 - Visibility information acquisition method and visibility information acquisition device - Google Patents

Description

  The present invention relates to a visibility information acquisition method and a visibility information acquisition device for determining a visibility state so as to analyze and evaluate the performance of, for example, a washer device and a wiper device provided in a vehicle.

  Conventionally, for example, the performance of a wiper device provided in a vehicle improves not only the performance of the wiping range, but also how good the visibility of the driver etc. is for rain (water droplets) on the windshield. It is done. That is, how well the wiper wipes water droplets (raindrops) on the windshield to improve visibility, specifically, for example, the ability to maintain visibility to easily recognize lanes and signs. is there.

However, these performances are generally evaluated only to the extent that they are visually compared with a sensory evaluation sample (standard sample).
Further, as a method for evaluating (detecting) the field of view through the glass, there is a method in which an image obtained by imaging the windshield is subjected to two-dimensional Fourier transform and evaluated based on the Fourier transform result (see, for example, Patent Document 1).

Japanese Patent No. 4226775

  However, the method disclosed in the above publication is for evaluating the visibility through the glass according to the climate at that time (detecting whether the glass is cloudy), However, the performance of a washer device or a wiper device could not be evaluated with high accuracy. Specifically, for example, evaluation for use in the optimum setting of the driving speed of the wiper and the optimum design of the shape of the wiper at the vehicle prototype stage or the like could not be performed.

  Also, of course, even in sensory evaluation, the state of view cannot be evaluated with high accuracy, for example, the evaluation becomes different depending on the evaluator, and the performance of the washer device or wiper device cannot be evaluated with high accuracy. .

  The present invention has been made to solve the above-mentioned problems, and its object is to provide a visual field information acquisition method and visual field information acquisition apparatus that can objectively evaluate the state of the visual field with high accuracy. Is to provide.

  In order to solve the above-mentioned problems, in the invention according to claim 1, a background member having a pattern with a constant interval including different brightness as a background of the glass to be evaluated is disposed, and the glass to be evaluated has the constant interval. An imaging step of capturing an image toward a pattern to obtain a captured image, a first quantification step of performing two-dimensional Fourier transform on the captured image and obtaining quantitative spatial frequency information based on the Fourier transform result, and the Fourier transform result And a second quantification step of calculating a contrast value for each frequency based on the inverse Fourier transform result and obtaining the contrast value for each frequency as quantitative contrast information, The gist is to evaluate the visibility state based on the spatial frequency information and the contrast information.

  According to the invention, in the imaging step, a background member having a pattern with a constant interval including different brightness as a background of the glass to be evaluated is arranged, and the glass to be evaluated is imaged toward the pattern with the constant interval. An image is obtained. In the first quantification step, the captured image is subjected to two-dimensional Fourier transform, and quantitative spatial frequency information is obtained based on the Fourier transform result. At this time, since a pattern with a constant interval is captured as the background of the glass to be evaluated, the identification in the Fourier transform result (power spectrum dispersion diagram in the spatial frequency domain) in a state where water droplets or the like are not attached to the glass to be evaluated The strength (power) of the frequency increases (concentrates). Further, in the state where water droplets or the like are attached to the glass to be evaluated, the strength (power) in the Fourier transform result is dispersed in a frequency region other than the specific frequency, and the influence of the water droplets appears strongly.

  In the second quantification step, the Fourier transform result is inverse Fourier transformed for each frequency, a contrast value for each frequency is calculated based on the inverse Fourier transform result, and the contrast value for each frequency is quantitative contrast. Obtained as information. Since the contrast information obtained in this way has a contrast value for each frequency, for example, more detailed information than that obtained by sampling a specific part of a captured image and calculating a single contrast value. Become.

  Since the visual field state is evaluated based on both the quantitative spatial frequency information and the contrast information, objective evaluation can be performed with high accuracy. In particular, information combining spatial frequency information and contrast information has been found to be greatly related to whether or not objects and characters, specifically lanes, signs, etc. can be recognized. Therefore, it is possible to objectively evaluate the significance of the visibility state with high accuracy and objectively. Therefore, for example, it is possible to objectively evaluate the performance of the wiper device and the washer device with high accuracy.

  According to a second aspect of the present invention, in the visual field information acquisition method according to the first aspect, the second quantifying step includes a minimum value and a maximum value of luminance from a minimum value and a maximum value of a waveform which are inverse Fourier transform results. And calculating the contrast value for each frequency from the minimum value and the maximum value of the luminance, and obtaining the contrast value for each frequency as quantitative contrast information.

  According to the invention, in the second quantification step, the minimum value and the maximum value of the luminance are calculated from the minimum value and the maximum value of the waveform as the inverse Fourier transform result, and each frequency is calculated from the minimum value and the maximum value of the luminance. And the contrast value for each frequency is obtained as quantitative contrast information. Therefore, the effect of the invention of claim 1 can be specifically obtained.

  According to a third aspect of the present invention, in the visibility information acquisition method according to the first or second aspect, a threshold value for pass / fail judgment based on the spatial frequency information and the contrast information is set in advance, The gist of the invention is that it includes a pass / fail evaluation step for determining pass / fail status and evaluating whether or not the total information obtained by combining the spatial frequency information and the contrast information exceeds the threshold value.

  According to the invention, in the pass / fail evaluation step, a threshold value for pass / fail determination based on the spatial frequency information and the contrast information is set in advance, and the spatial frequency information and the contrast information are combined. Since the visibility state is determined as pass / fail depending on whether the information exceeds the threshold value, it is possible to evaluate pass / fail with high accuracy and objective evaluation. For example, by setting the threshold value to correspond to the pass / fail of the wiper device, it can be easily used for optimal setting of the wiper drive speed, optimal design of the wiper shape, etc. at the vehicle prototype stage.

  According to a fourth aspect of the present invention, the background member, the imaging means for performing the imaging step, and the first quantification step for performing the first quantification step in the visibility information acquisition method according to the first or second aspect. The gist is provided with a quantification means and a second quantification means for performing the second quantification step.

  According to the configuration, the visibility information acquisition method according to claim 1 or 2 can be easily executed, and the visibility information acquisition device capable of obtaining the effect of the invention according to claim 1 or 2 is provided. be able to.

  According to a fifth aspect of the present invention, in the field-of-view information acquisition device according to the fourth aspect, storage means for storing a threshold value for pass / fail judgment based on the spatial frequency information and the contrast information; The gist of the present invention is to provide a pass / fail judgment means for judging pass / fail status by evaluating whether or not the total information obtained by combining the spatial frequency information and the contrast information exceeds the threshold value.

  According to this configuration, it is possible to easily execute the visual field information acquisition method according to the third aspect, and it is possible to provide a visual field information acquisition apparatus that can obtain the effects of the invention according to the third aspect.

  ADVANTAGE OF THE INVENTION According to this invention, the visual field information acquisition method and visual field information acquisition apparatus which enable objective evaluation of the visual field state with high precision can be provided.

The schematic diagram of the visual field information acquisition apparatus in this Embodiment. The partial front view of the background member in this Embodiment. Captured image. (A) The power spectrum dispersion | distribution figure in case there are few adhesions of a water droplet. (B) Power spectrum dispersion diagram when water droplets are frequently attached. (A) The characteristic view for demonstrating a threshold value. (B) An image for visualizing the relationship between the spatial frequency and the contrast value (contrast sensitivity) in (a).

Hereinafter, an embodiment embodying the present invention will be described with reference to FIGS.
As shown in FIG. 1, the visual field information acquisition apparatus of this embodiment includes a background member 11, a video camera 12 as an imaging means, a first and second quantifying means, a storage means, and a personal computer as a pass / fail judgment means. And the like.

  The background member 11 is a fixed background in which the state does not change, has a pattern 11a (see FIG. 2) with a constant interval including different brightness, and is used as a background of the windshield 21 as the glass to be evaluated. Arranged forward. The pattern 11a (see FIG. 2) with a constant interval is a pattern in which the luminance of the captured image changes greatly within the predetermined interval and repeats the same pattern at every predetermined interval. It is a vertical stripe pattern in which a bright portion (white) and a dark portion (black) are arranged in parallel with a half width (4 cm in the present embodiment) of a half width (4 cm in the present embodiment). In addition, the pattern 11a (see FIG. 2) at regular intervals is a pattern in which the strength (power) of a specific frequency is increased (concentrated) when two-dimensional Fourier transform is performed. Is a vertical stripe pattern (black and white alternate vertical stripe pattern) in which the strength (power) of a specific frequency in the horizontal direction is increased.

  The video camera 12 is arranged in the vehicle driver's seat 24 (assuming the driver's eye point EP (see FIG. 3)) so as to face the pattern 11a at a predetermined interval through the windshield 21. The video camera 12 is capable of obtaining moving image data obtained by capturing moving images of the windshield 21 toward the pattern 11a at a predetermined interval. Note that the video camera 12 according to the present embodiment can obtain moving image data including captured images (see FIG. 3) that are 30 frames (30) still images per second. The video camera 12 of the present embodiment has hundreds of thousands to millions of effective pixels, and in the present embodiment, it is possible to obtain the captured image of about 2 million pixels. .

The arithmetic device 13 performs various processes described later.
Further, a display 14 capable of displaying various information such as the calculation result of the calculation device 13 is connected to the calculation device 13.

Here, various processes performed by the arithmetic device 13 will be described in detail while explaining the visibility information acquisition method.
The visibility information acquisition method includes an “imaging process”, a “first quantification process”, a “second quantification process”, and a “pass / fail evaluation process”.

  First, in the “imaging process”, as shown in FIG. 1, a background member 11 having a pattern 11a (see FIG. 2) with a constant interval is arranged as a background of the windshield 21, and a wiper 23 that is driven while water droplets are applied thereto is applied. The windshield 21 to be wiped away is taken as a moving image toward the pattern 11a at a constant interval to obtain moving image data (captured images that are a large number of still images).

  Next, in the “first quantification step”, the arithmetic device 13 performs two-dimensional Fourier transform on the captured image, which is a still image of the moving image data, and obtains quantitative spatial frequency information based on the Fourier transform result. The captured image used (processed) in the present embodiment is an image of a preset setting range (so-called eye point EP) located in front of the driver's seat, as shown in FIG.

  Specifically, in the “first quantification step” of the present embodiment, first, captured images that are still images at fixed time intervals (for example, every 0.05 seconds) in the moving image data are each two-dimensionally Fourier transformed, and the Fourier is obtained. As a result of conversion, each power spectrum dispersion diagram (which is a spatial frequency region, for example, see FIGS. 4A and 4B) is obtained. FIG. 4A is a power spectrum dispersion diagram P1 when there is little water droplet adhesion (immediately after wiping), and FIG. 4B is a power spectrum when there is much water droplet adhesion (immediately before wiping). It is a spectral dispersion diagram P2.

  In the “first quantification step”, the spatial frequency information is obtained by calculating where the frequency (power) in the Fourier transform result is distributed in the frequency range from where to where. Specifically, in this embodiment, it is determined whether or not the strength (power) in the Fourier transform result exceeds a preset reference value, and the frequency range (dispersion range) exceeding the spatial frequency is determined. Information.

  Next, in the “second quantification step”, the arithmetic unit 13 performs inverse Fourier transform on the Fourier transform result for each frequency (for each preset frequency range), and based on the inverse Fourier transform result, the contrast value for each frequency. And the contrast value for each frequency is obtained as quantitative contrast information. Specifically, the arithmetic unit 13 calculates the minimum value and the maximum value of the luminance from the minimum value and the maximum value of the waveform as the inverse Fourier transform result, and the contrast for each frequency from the minimum value and the maximum value of the luminance. The value is calculated, and the contrast value for each frequency is obtained as quantitative contrast information. In the present embodiment, the contrast value is calculated by calculating the minimum value and the maximum value of the brightness from the minimum value and the maximum value of the waveform without displaying the waveform itself as the inverse Fourier transform result on the display 14 in particular. The contrast value of the present embodiment uses Michelson contrast. Specifically, the maximum luminance value is Lmax, and the minimum luminance value is Lmin. Is calculated.

  In the “pass / fail evaluation step”, a threshold value Z for pass / fail judgment based on the spatial frequency information and the contrast information (see FIG. 5A) is set (stored) in advance in the arithmetic unit 13 (storage means). The pass / fail state is determined based on whether or not the total information S1 to S4 obtained by combining the spatial frequency information and the contrast information obtained in the first and second quantification steps exceeds the threshold value Z. And evaluate.

  Specifically, as shown in FIG. 5 (a), the threshold value Z is a value set for information combining the contrast sensitivity and the spatial frequency. This is a value set in advance based on the questionnaire result, which allows a large number of people to conduct a questionnaire on whether or not characters can be identified.

  In the “pass / fail evaluation step”, the visibility information is determined as pass / fail according to whether or not at least a part of the total information S1 to S4 in which the spatial frequency information and the contrast information are combined exceeds the threshold value Z. To do. In the present embodiment, assuming the rain of several tens of millimeters, the wiper 23 is driven at high speed while applying the corresponding water droplets in the “imaging process”, so that the minimum visibility is always maintained (even immediately before wiping). It is decided to make a pass / fail judgment. Here, the general information S1 is information immediately after wiping and is in a state corresponding to FIG. The general information S2 is information immediately before wiping and is in a state corresponding to FIG. Since some of the total information S1 and S2 does not exceed the threshold value Z, the arithmetic unit 13 (pass / fail judgment means) makes a pass judgment. The comprehensive information S3 is information immediately before wiping in another wiper device (for example, the shape of the wiper 23 and the driving device (driving speed) are different), and the contrast value (information) is higher than the comprehensive information S2. Since the spatial frequency (information) is wider than the general information S2 and partly exceeds the threshold value Z, in this case, the arithmetic device 13 (pass / fail judgment means) makes a failure judgment. Further, the general information S4 is information immediately before wiping in another wiper device (for example, the shape of the wiper 23 or the driving device (driving speed) is different), and the entire information exceeds the threshold value Z. In this case, the arithmetic device 13 (pass / fail judgment means) makes a failure judgment. FIG. 5B is an image for visually visualizing the relationship between the spatial frequency and the contrast value (contrast sensitivity) in FIG.

Next, characteristic effects of the above-described embodiment (method and apparatus) will be described below.
(1) In the “imaging process”, the background member 11 having the pattern 11a with a constant interval including different brightness as the background of the windshield 21 is arranged, and the windshield 21 is imaged and imaged toward the pattern 11a with a constant interval. An image is obtained. In the “first quantification step”, the captured image is subjected to two-dimensional Fourier transform, and quantitative spatial frequency information is obtained based on the Fourier transform result. At this time, since the pattern 11a with a constant interval is captured as the background of the windshield 21, in a state where water droplets or the like are not attached to the windshield 21, in the Fourier transform result (power spectrum dispersion diagram, spatial frequency region) The intensity (power) of a specific frequency becomes stronger (concentrated). Further, in the state where water droplets or the like are attached to the windshield 21, the strength (power) in the Fourier transform result is dispersed in a frequency region other than the specific frequency, and the influence of the water droplets appears strongly.

  In the “second quantification step”, the Fourier transform result is inverse Fourier transformed for each frequency, a contrast value for each frequency is calculated based on the inverse Fourier transform result, and the contrast value for each frequency is quantitative. Can be obtained as accurate contrast information. Specifically, in the present embodiment, the minimum value and the maximum value of the luminance are calculated from the minimum value and the maximum value of the waveform that is the result of the inverse Fourier transform, and each frequency is calculated from the minimum value and the maximum value of the luminance. The contrast value is calculated, and the contrast value for each frequency is obtained as quantitative contrast information. Since the contrast information obtained in this way has a contrast value for each frequency, for example, more detailed information than that obtained by sampling a specific part of a captured image and calculating a single contrast value. Become.

  Since the visual field state is evaluated based on both the quantitative spatial frequency information and the contrast information, objective evaluation can be performed with high accuracy. In particular, the information (general information S1 to S4) in which the spatial frequency information and the contrast information are combined is information that is greatly related to whether or not an object or a character, specifically, a lane or a sign can be recognized. In the present invention, it is possible to objectively evaluate the significance of the visibility state with high accuracy. Therefore, for example, it is possible to objectively evaluate the performance of the wiper device and the washer device with high accuracy.

  (2) In the “pass / fail evaluation step”, a threshold value Z for pass / fail determination based on the spatial frequency information and the contrast information is set in advance, and the spatial frequency information and the contrast information are combined. Since the visibility state is determined as pass / fail depending on whether or not the information S1 to S4 exceeds the threshold value Z, it is possible to evaluate the pass / fail objective with high accuracy. For example, as in this embodiment, by setting the threshold value Z to correspond to the pass / fail of the wiper device, it is easy to use for setting the wiper drive speed, designing the wiper shape, etc. be able to.

The above embodiment may be modified as follows.
In the above-described embodiment, the visibility state (and consequently, whether or not the total information S1 to S4 in which the spatial frequency information and the contrast information are combined in the “pass / fail evaluation step” by the arithmetic unit 13 exceeds the threshold value Z) The wiper device) is evaluated based on the pass / fail determination, but the visibility state is not limited to the pass / fail determination by the arithmetic device 13, and the visibility state may be evaluated from a different viewpoint. That is, if the visual field state is evaluated based on the spatial frequency information and the contrast information, it is not particularly necessary to perform the pass / fail determination by the arithmetic unit 13, and evaluation of how much the visual field state is achieved at each timing. May be performed. In addition, if each information is acquired about the captured image which is a still image for every fixed time (for example, every 0.05 second) in moving image data like the said embodiment, the visual field state in each timing can be evaluated, respectively. For example, it is possible to evaluate the performance of the washer device itself (for example, the visibility state due to the landing of the washer liquid) or the combination of the washer device and the wiper device (for example, the wiped state with the wiper after the washer liquid is jetted). Further, when performing different evaluations in this way, the threshold value Z may be set based on different criteria, or a plurality of stepwise threshold values may be set to perform not only pass / fail determination but also level determination. Also good.

  In the above embodiment, the pattern 11a (see FIG. 2) with a constant interval is replaced with a portion (white) with a lightness at a half width (2 cm in this embodiment) of the fixed interval (4 cm in this embodiment). Although a vertical stripe pattern in which dark portions (black) are arranged side by side is not limited to this, for example, a horizontal stripe pattern may be used.

  In the “second quantification step” of the above embodiment, the minimum value and the maximum value of the luminance are calculated from the minimum value and the maximum value of the waveform as the inverse Fourier transform result, and the minimum value and the maximum value of the luminance are calculated. Although the contrast value for each frequency is calculated, it may be changed to another method as long as the contrast value for each frequency can be calculated based on the inverse Fourier transform result. For example, the contrast value may be directly calculated from the difference between the minimum value and the maximum value of the waveform that is the inverse Fourier transform result, that is, the amplitude of the waveform that is the inverse Fourier transform result.

  In the above embodiment, the Michelson contrast “(Lmax−Lmin) / (Lmax + Lmin)” is used. However, a contrast value calculated by another formula may be used. For example, the contrast value may be calculated by “Lmax−Lmin” or may be calculated by “Lmax−T (T is a preset value)”.

  In the above embodiment, the video camera 12 is used as the imaging unit, but the present invention is not limited to this, and the video camera 12 may be changed to a camera that can capture only still images.

  DESCRIPTION OF SYMBOLS 11 ... Background member, 11a ... Pattern of fixed interval, 12 ... Video camera (imaging means), 13 ... Arithmetic unit (first and second quantification means, storage means, and pass / fail judgment means), 21 ... Windshield (covered) Evaluation glass), S1 to S4 ... general information, Z ... threshold value.

Claims (5)

An imaging step of arranging a background member having a pattern with a constant interval including different brightness as the background of the glass to be evaluated, imaging the glass to be evaluated toward the pattern with the constant interval, and obtaining a captured image;
A first quantification step of performing two-dimensional Fourier transform on the captured image and obtaining quantitative spatial frequency information based on the Fourier transform result;
A second quantification step of performing inverse Fourier transform on the Fourier transform result for each frequency, calculating a contrast value for each frequency based on the inverse Fourier transform result, and obtaining the contrast value for each frequency as quantitative contrast information; And a visual field state evaluation method based on the spatial frequency information and the contrast information. The visibility information acquisition method according to claim 1,
The second quantification step calculates the minimum value and maximum value of the luminance from the minimum value and maximum value of the waveform as the inverse Fourier transform result, and calculates the contrast value for each frequency from the minimum value and maximum value of the luminance. And obtaining a contrast value for each frequency as quantitative contrast information. In the visibility information acquisition method according to claim 1 or 2,
A threshold value for pass / fail judgment based on the spatial frequency information and the contrast information is set in advance, and whether or not the total information obtained by combining the spatial frequency information and the contrast information exceeds the threshold value. A visibility information acquisition method comprising: a pass / fail evaluation step for determining pass / fail status and evaluating it. The background member in the visibility information acquisition method according to claim 1 or 2,
Imaging means for performing the imaging step;
First quantification means for performing the first quantification step;
A field-of-view information acquisition apparatus comprising: second quantification means for performing the second quantification step. In the visibility information acquisition apparatus according to claim 4,
Storage means for storing a threshold value for pass / fail judgment based on the spatial frequency information and the contrast information;
Visibility information comprising: pass / fail judgment means for judging pass / fail status by evaluating whether or not total information obtained by combining the spatial frequency information and the contrast information exceeds the threshold value Acquisition device.