Water level estimation system

Abstract

To provide a water level estimation system that can determine a water level with higher accuracy than that of a prior art even for a water area having high transparency.SOLUTION: A water level estimation system of the present application, which is a system that estimates a water level on the basis of an on-site image including a water level marker partially inserted into water, includes scale detection means, scale luminance value calculation means, luminance change value calculation means and area determination means. The area determination means is means for determining an aerial area and an underwater area on the basis of a luminance change value extracted by the luminance change value calculation means. Then, the system determines as a water level a boundary between the aerial area and the underwater area, determined by the area determination means.SELECTED DRAWING: Figure 2

Description

Technical Field

[0001] The invention of the present application relates to a technology related to the water level in a water area such as a river. More specifically, it relates to a water level estimation system that can estimate the water level based on changes in the luminance value of a water gauge included in an image.

Background Art

[0002] It is beneficial to grasp the water level of a river. Especially when floods such as heavy rain are predicted, the water level of the river becomes extremely important information, such as issuing evacuation advisories according to the water level of the river. To observe the water level of a river, it has been done by monitoring contact-type relative pressure water level gauges or non-contact radio wave water level gauges, etc. so far, and sometimes the water gauge is also monitored as a reference value. This water gauge is a marker with a plurality of graduations displayed in its axial direction, and usually graduations of two colors (for example, white and black) are alternately attached at equal intervals. By visually observing the graduation at the position where the water surface overlaps with the water gauge attached to a river bank protection, bridge pier, dam, etc., the height attached to the graduation (for example, the height from the bottom of the water) can be determined as the water level.

[0003] Although investigators etc. may directly visually observe the water gauge, since it is necessary to continuously visually observe it for a certain period of time to grasp the change in the water level, a great burden is imposed on the investigator. In addition, these days, there are also many situations where it is advisable to avoid investigators approaching the water gauge, such as intense heavy rain occurring locally in a short period of time.

[0004] Therefore, a method of installing a camera (so-called fixed-point camera) so that the shooting angle does not change and using the fixed-point camera for water level monitoring has become common. That is, the fixed-point camera periodically takes pictures of the water gauge and transmits the images, and the water level is continuously monitored in a management building or the like far from the site. For example, in Patent Document 1, two types of cameras, a rotary space monitoring camera and a fixed camera (fixed-point camera), are used. The surrounding situation is confirmed with the rotary space monitoring camera, and a technique for analyzing the image of the water gauge taken by the fixed camera to determine the water level is proposed.

Prior Art Documents

Patent Documents

[0005]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0006] Normally, since the fixed-point camera takes pictures of the water gauge at short intervals, a large number of images are transmitted. That is, it is necessary to check the water level for a large number of images. If this checking work is done by workers, it will require a lot of people and time, and as a result, it will be costly. On the other hand, when estimating the water level by analyzing images as in Patent Document 1, labor costs can be suppressed and the water level can be grasped in a relatively short time. However, in conventional image analysis, incorrect water levels were sometimes output. In particular, in the case where the transparency of the water is high, it is difficult for even a person to detect the water surface from the image, and moreover, conventional image analysis sometimes shows significantly different water levels.

[0007] The problem of the present invention is to solve the conventional problems, that is, to provide a water level estimation system that can obtain the water level with higher accuracy than the prior art even in waters with high transparency.

Means for Solving the Problems

[0008] The invention of the present application focuses on the fact that among the water level markers contained in the image, the luminance change of the graduations in the air is generally constant, while the luminance change of the graduations in the water is unstable, and the water level is obtained by utilizing this fact. It is an invention made based on an idea that has never existed before.

[0009] The water level estimation system of the invention of the present application is a system for estimating the water level based on a field image containing a water level marker partially inserted into the water, and is provided with scale detection means, scale luminance value calculation means, luminance change value calculation means, and region determination means. Note that two or more types of graduations with different luminance values are displayed at equal intervals on the water level marker, and a plurality of graduations in the field image are arranged in a direction that is substantially vertical (including vertical). The water level marker constituting the water level estimation system is a means for detecting the "scale pixels" that constitute the graduations from the field image. Further, the scale luminance value calculation means sets a "scale pixel column" in which the scale pixels detected by the scale detection means are arranged in a row in order from below or above, and is a means for obtaining the luminance value for each scale pixel included in the scale pixel column. The luminance change value calculation means is a means for calculating a "luminance change value" indicating the degree of change in the luminance values of the scale pixels arranged vertically in the scale pixel column, and the region determination means is a means for determining the air region and the water region of the water level marker based on the luminance change value extracted by the luminance change value calculation means. Then, the boundary between the air region and the water region determined by the region determination means is estimated as the water level.

[0010] The water level estimation system of the invention of the present application can also be further provided with representative luminance value calculation means. When the luminance values of the scale pixels arranged vertically in the scale pixel column are substantially the same (including the same), this representative luminance value calculation means sets a "scale group" of a plurality of scale pixels assuming that these scale pixels are in the same scale, and calculates a "representative scale luminance value" by statistically processing the luminance values of the plurality of scale pixels constituting this scale group. In this case, the luminance change value calculation means calculates the degree of change in the representative scale luminance values adjacent vertically as the luminance change value.

[0011] The water level estimation system of the present invention can also be provided with average luminance value calculation means. This average luminance value calculation means is means for calculating a "moving average luminance value (average value of luminance values related to a predetermined number of scale pixels)" while moving a predetermined number of scale pixels in order from below or above with respect to the scale pixel column. In this case, the luminance change value calculation means calculates the degree of change in the moving average luminance values adjacent vertically in the scale pixel column as the luminance change value.

[0012] The water level estimation system of the present invention can also be configured to calculate the amount of change in the luminance value related to the scale pixel as the luminance change value. Here, the amount of change in the luminance value is the difference in the luminance values related to the scale pixels arranged vertically in the scale pixel column. In this case, the area determination means determines, as the underwater area, an area in the scale pixel column where the luminance change value is smaller than a predetermined change amount threshold.

[0013] The water level estimation system of the present invention can also be configured to calculate the change interval of the luminance value related to the scale pixel as the luminance change value. Here, the change interval of the luminance value is the distance between the scale pixel indicating the maximum value of luminance and the scale pixel indicating the minimum value of luminance in the scale pixel column. In this case, the area determination means determines, as the underwater area, an area in the scale pixel column where the luminance change value is smaller than a predetermined change interval threshold.

[0014] The water level estimation system of the present invention can also be configured to estimate the water level after setting a plurality of columns of scale pixel columns. In this case, the scale luminance value calculation means sets a plurality of columns of scale pixel columns according to the horizontal width dimension of the scale displayed on the water level indicator, and the area determination means determines the air area and the underwater area of the water level indicator for each scale pixel column set by the scale luminance value calculation means. Then, a plurality of water levels (so-called provisional water levels) are obtained based on the plurality of sets of air areas and underwater areas determined by the area determination means, and the water level (so-called determined water level) is estimated by statistically processing the plurality of water levels.

Advantages of the Invention

[0015] The water level estimation system of the present invention has the following advantages. (1) It can read the water level with high accuracy without requiring a lot of labor costs and working hours. (2) Even for a water level marker installed in water with high transparency, the water level line can be extracted more accurately compared to the prior art. (3) Even if so-called "poor-quality photos" are used, the water level line can be extracted more accurately compared to the prior art.

Brief Description of the Drawings

[0016]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Embodiments for Carrying Out the Invention

[0017] An example of an embodiment of the water level estimation system of the present invention will be described with reference to the drawings. The water level estimation system of the present invention can estimate the water level of a "water area" such as a river or a marsh. For convenience, an example in which this water area is a "river" will be described here.

[0018] The water level estimation system of the present invention estimates the water level of a river using an image that captures the situation of the river (hereinafter referred to as the "on-site image"). As a means of acquiring the on-site image, it is desirable to be able to automatically take pictures regularly. For example, a digital camera or a digital video camera can be used. Also, the on-site image may be acquired by a so-called fixed-point camera so that the angle of view does not change even when repeatedly photographed.

[0019] The on-site image includes a "water level marker" together with the river, and is an image that captures a situation where a part of the water level marker (especially the lower part) is inserted into the water as shown in Fig. 1(a). Here, the water level marker is one in which two or more scales are displayed at equal intervals, and the length from the end (lower end or upper end) of the water level marker can be read according to the position of the scale. As a representative example, a water gauge (water gauge board) can be cited. Of course, as long as the above conditions are satisfied, a ribbon rod, a steel tape, a leveling staff (rod), etc. can also be used, or the repetition of blocks and joints constructed as the local revetment can be used as the water level marker.

[0020] As described above, the water level marker is one in which different types of scales are repeatedly displayed in the same pattern. In the on-site image in which the water level marker is included, these scales are arranged so as to be aligned in a substantially vertical (including vertical) direction, and moreover, different types of scales are displayed with different luminance values. For example, in the on-site image PS shown in Fig. 1(a), a water gauge WM is included as the water level marker, and this water gauge WM is provided with a "white scale SW" and a "black scale SB" as shown in Fig. 1(b). In this case, the white scale SW with a large luminance value and the black scale SB with a small luminance value are repeatedly displayed in the on-site image, and moreover, these scales are arranged in a substantially vertical direction. For the sake of convenience, an example using the water gauge WM as the water level marker will be described here. Also, as shown in Fig. 1(b), the longitudinal direction (vertical direction in the figure) of the water gauge WM is referred to as the "axial direction", and the direction perpendicular to the axial direction (left-right direction in the figure) is referred to as the "width direction".

[0021] The water level estimation system of the present invention cuts out a partial image of the water gauge WM (hereinafter referred to as "label image PM") from the on-site image PS shown in Fig. 1(a), so to speak, and estimates the water level of the river based on the pixels constituting the scale (hereinafter referred to as "scale pixels") in this label image PM. Specifically, as shown in Fig. 1(c), a line (hereinafter referred to as "reference line LS") passing through these scales (that is, scale pixels) at a predetermined position in the scale is set, and the water level is estimated by paying attention to the change in the luminance values of the group of scale pixels arranged on this reference line LS (hereinafter referred to as "scale pixel column").

[0022] Fig. 2 is a graph showing the luminance value of each scale pixel arranged in the scale pixel column (hereinafter referred to as "depth-luminance value graph"), where the horizontal axis indicates the ordinal number of the scale pixel counted from the upper end of the scale pixel column (for example, the 100th etc.), and the vertical axis indicates the luminance value related to the scale pixel. Since the water gauge WM is arranged in a posture with its axial direction being substantially vertical as shown in Fig. 1(a), the depth from the water surface can be grasped according to the order (hereinafter referred to as "arrangement rank") from the end (upper end or lower end) of the scale pixel in the scale pixel column. Therefore, in this figure, the horizontal axis is taken as the "depth in the axial direction".

[0023] The water level gauge WM is marked with white graduations SW and black graduations SB at regular intervals. Therefore, in the depth-luminance value graph, large luminance values (white graduations SW) and small luminance values (black graduations SB) should be plotted periodically and repeatedly. However, although large and small luminance values are repeated periodically in the air, the inventors have found that this tendency breaks down in water. For example, in most of the left part of FIG. 2, large and small luminance values are plotted periodically, but in a part of the right side, large and small luminance values are repeated irregularly. That is, the range where large and small luminance values are plotted repeatedly with a certain tendency (the left part in FIG. 2) is the part of the water level gauge WM placed in the air (hereinafter referred to as the "air region"), and the range where large and small luminance values are plotted randomly (a part of the right side in FIG. 2) can be understood as the part of the water level gauge WM placed in water (hereinafter referred to as the "water region"). The inventors have also found that the interval between large and small luminance values is smaller in water than in the air, and the difference is also smaller. And it can be estimated that the boundary between the air region and the water region of the water level gauge WM is the water level of the river.

[0024] FIG. 3 is a block diagram showing the main configuration of the water level estimation system 100 of the present invention. As shown in this figure, the water level estimation system 100 of the present invention includes a graduation detection means 101, a graduation luminance value calculation means 102, a luminance change value calculation means 103, and a region determination means 104, and can further include a representative luminance value calculation means 105, an average luminance value calculation means 106, a mark detection means 107, a water level output means 108, a local image storage means 109, a mark information storage means 110, etc.

[0025] The scale detection means 101, scale luminance value calculation means 102, luminance change value calculation means 103, region determination means 104, representative luminance value calculation means 105, average luminance value calculation means 106, sign detection means 107, and water level output means 108 that make up the water level estimation system 100 of the present invention can be manufactured as dedicated ones, or general-purpose computer devices can also be used. That is, by causing a computer device to execute arithmetic processing according to a predetermined program, the processing of various means is carried out. This computer device includes a processor such as a CPU (Central Processing Unit) or GPU (Graphics Processing Unit), and a memory such as a ROM or RAM. Some also include input means such as a mouse and keyboard, and a display, and can be configured by, for example, a personal computer (PC) or a server.

[0026] Also, the on-site image storage means 109 and the sign information storage means 110 can use the storage device of a general-purpose computer (for example, a personal computer), or can be constructed in a database server. When constructing in a database server, it can be placed on a local area network (LAN: Local Area Network), or can be a cloud server stored via the Internet.

[0027] Hereinafter, each main element that makes up the water level estimation system 100 will be described in detail.

[0028] (On-site Image Storage Means and Sign Information Storage Means) The on-site image storage means 109 is a means for storing the on-site image PS. When the fixed-point camera installed on-site periodically acquires the on-site image PS, it is advisable to adopt a specification in which the on-site image storage means 109 sequentially stores these on-site images PS. Also, it is possible to adopt a specification in which the on-site image PS acquired by the fixed-point camera is transmitted to a remote location and the on-site image storage means 109 sequentially stores the transmitted on-site image PS. The marking information storage means 110 is a means for storing the attribute information regarding the water gauge WM. Examples of the attribute information of the water gauge WM include the overall length of the water gauge WM, the dimension in the axial direction of the scale (i.e., the vertical width), the dimension in the width direction of the scale (hereinafter referred to as the "lateral dimension"), the standard luminance value for each scale, and a table showing the correspondence between the position of the scale and the distance from the end of the water gauge WM.

[0029] (Marking detection means) The marking detection means 107 is a means for detecting the marking image PM (Fig. 1(b)) from the on-site image PS (Fig. 1(a)). When detecting the marking image PM, various conventionally used image analysis techniques can be adopted. For example, the marking image PM can be detected by template matching. At this time, template matching can also be executed while rotating, enlarging, or reducing the on-site image PS.

[0030] (Scale detection means) The scale detection means 101 is a means for detecting "scale pixels (pixels constituting the scale)" from the marking image PM. When detecting the scale pixels, similar to the detection of the marking image PM, various conventionally used image analysis techniques can be adopted. For example, the marking image PM can be detected by template matching. Of course, template matching can also be executed while rotating, enlarging, or reducing the marking image PM. In the water level estimation system 100, it is also possible to adopt a specification in which the scale detection means 101 directly detects the scale pixels from the marking image PM. In this case, the marking detection means 107 can also be omitted.

[0031] (Scale luminance value calculation means) The scale brightness value calculation means 102 is means for setting a scale pixel column and obtaining a brightness value related to the scale pixels. Hereinafter, the process executed by the scale brightness value calculation means 102 will be described in detail. First, the scale brightness value calculation means 102 sets a reference line LS passing through these scales (that is, scale pixels) at a predetermined position. At this time, depending on the horizontal width dimension of the scale, a plurality of reference lines LS can also be set. For example, in FIG. 4, three reference lines LS are set at three locations of the scale.

[0032] After setting the reference line LS, the scale pixels arranged on this reference line LS are arranged from the upper end (or lower end) to generate a scale pixel column. Of course, when a plurality of reference lines LS are set, a scale pixel column may be generated for each reference line LS. After generating the scale pixel column, the brightness value related to the scale pixels constituting the scale pixel column is obtained. In this way, when the scale pixel column is constituted by the scale pixels having brightness values, so to speak, a depth-brightness value graph (FIG. 2) can also be created. At this time, when a plurality of reference lines LS are set, a depth-brightness value graph may be created for each reference line LS. For example, in the case of FIG. 5, ten reference lines LS are set, and therefore ten types of depth-brightness value graphs are created.

[0033] (Representative brightness value calculation means) The representative brightness value calculation means 105 is means for calculating a representative brightness value (hereinafter referred to as "representative scale brightness value") for each scale. Hereinafter, the procedure until the representative brightness value calculation means 105 calculates the representative scale brightness value will be described in detail. First, the representative brightness value calculation means 105 compares the brightness values related to the scale pixels adjacent vertically to each other. At this time, the scale pixels to be compared are those constituting the scale pixel column, and of course, they are arranged according to the arrangement order. Therefore, it is preferable to compare the adjacent scale pixels in order from the upper end (or lower end) of the scale pixel column.

[0034] When the difference in luminance values between the upper and lower scale pixels is less than a predetermined threshold (hereinafter referred to as the "difference threshold"), that is, when the luminance values of both are substantially the same (including the same), the representative luminance value calculation means 105 determines that these scale pixels are the same (the same type and at the same position) scale. On the other hand, when the difference in luminance values between the upper and lower scale pixels is equal to or greater than the difference threshold, that is, when the luminance values of both are different, the representative luminance value calculation means 105 determines that these scale pixels are different scales. Then, by sequentially repeating this determination process from the upper end to the lower end (or from the lower end to the upper end) of the scale pixel column, a set (hereinafter referred to as a "scale group") composed of a plurality of scale pixels in the same scale is set. When a plurality (usually the number of scales attached to the water gauge WM) of scale groups are set, a representative scale luminance value representing the scale is calculated based on the luminance values of the plurality of scale pixels constituting the scale group. When calculating the representative scale luminance value, various statistical processes such as obtaining an average value, a median value, or a mode value may be performed.

[0035] (Average luminance value calculation method) The average luminance value calculation means 106 is a means for calculating a moving average of luminance values (hereinafter referred to as the "moving average luminance value"). Hereinafter, the procedure until the average luminance value calculation means 106 calculates the moving average luminance value will be described in detail. First, the average luminance value calculation means 106 extracts a predetermined number (a previously set number) of scale pixels and obtains a moving average luminance value, which is the average value of the luminance values of these extracted scale pixels. However, the extracted scale pixels are those constituting the scale pixel column and are a predetermined number of scale pixels that are consecutive in the array order. Then, while sequentially moving the predetermined number from the upper end to the lower end (or from the lower end to the upper end) of the scale pixel column, the moving average luminance value is repeatedly obtained.

[0036] (Luminance change value calculation means) The luminance change value calculation means 103 is a means for calculating a "luminance change value" based on the scale pixels constituting the scale pixel column. Here, the luminance change value is a value indicating the degree of change in the luminance values of the scale pixels arranged vertically, and it can be the "amount of change in luminance value" as the luminance change value, or the "interval of change in luminance value" as the luminance change value, or alternatively the "gradient of change in luminance value" as the luminance change value.

[0037] When obtaining the luminance change value as the amount of change in luminance value, focus on the scales adjacent vertically, extract the luminance value related to any one of the scale pixels constituting the upper scale, and extract the luminance value related to any one of the scale pixels constituting the lower scale. Then, the difference between these luminance values can be used as the luminance change value. Alternatively, using the representative scale luminance value obtained by the representative luminance value calculation means 105, the difference between the representative scale luminance values adjacent vertically can be used as the luminance change value, or using the moving average luminance value obtained by the moving average luminance value calculation means 106, the difference between the moving average luminance values adjacent vertically can be used as the luminance change value.

[0038] When obtaining the luminance change value as the interval of change in luminance value, the distance between the scale pixel indicating the maximum value of luminance and the scale pixel indicating the minimum value of luminance can be used as the luminance change value. Here, the extreme value is the luminance value at the point where the gradient reverses in the depth-luminance value graph (Figure 2). Among the extreme values, the maximum value is the luminance value at the point where the ascending gradient reverses to the descending gradient, and the minimum value among the extreme values is the luminance value at the point where the descending gradient reverses to the ascending gradient. Then, compare the scale pixels of adjacent maximum and minimum values, and obtain the distance between the two as the luminance change value. The depth-luminance value graph for obtaining the interval of change in luminance value may be one in which the luminance value is plotted for each scale pixel, or in addition to the one in which the representative scale luminance value obtained by the representative luminance value calculation means 105 is plotted, it may also be one in which the moving average luminance value obtained by the moving average luminance value calculation means 106 is plotted.

[0039] When obtaining the luminance change value as the change gradient of the luminance value, the value obtained by dividing the "amount of change in luminance value" by the "interval of change in luminance value" can be used as the luminance change value. In this case as well, a depth-luminance value graph will be created. However, it is also possible to use a depth-luminance value graph in which the luminance value is plotted for each scale pixel, or a depth-luminance value graph in which the representative scale luminance value obtained by the representative luminance value calculation means 105 is plotted, or a depth-luminance value graph in which the moving average luminance value obtained by the moving average luminance value calculation means 106 is plotted.

[0040] (Area determination means) The area determination means 104 is a means for determining the air region and the underwater region among the water level marks WM. Specifically, the range in which the luminance change value obtained by the luminance change value calculation means 103 is below a predetermined threshold value (hereinafter referred to as the "area threshold value") is determined as the underwater region, and the range in which it is equal to or above the area threshold value is determined as the air region. For example, when the amount of change in luminance value is used as the luminance change value, after setting a predetermined amount of change (hereinafter referred to as the "amount of change threshold value") as the area threshold value and comparing it with the luminance change value, the range in which the luminance change value is below this amount of change threshold value is determined as the underwater region, and the range in which it is equal to or above the amount of change threshold value is determined as the air region. Similarly, when the interval of change in luminance value is used as the luminance change value, after setting a predetermined interval of change (hereinafter referred to as the "interval of change threshold value") as the area threshold value and comparing it with the luminance change value, the range in which the luminance change value is below this interval of change threshold value is determined as the underwater region, and the range in which it is equal to or above the interval of change threshold value is determined as the air region. When the change gradient of the luminance value is used as the luminance change value, after setting a predetermined change gradient (hereinafter referred to as the "change gradient threshold value") as the area threshold value and comparing it with the luminance change value, the range in which the luminance change value is below this change gradient threshold value is determined as the underwater region, and the range in which it is equal to or above the change gradient is determined as the air region. When a plurality of reference lines LS are set, it is advisable to determine the air region and the underwater region for each reference line LS.

[0041] (Water level output means) The water level output means 108 is a means for using the boundary between the air region and the water region determined by the region determination means 104 as the water level of the river and outputting this to a display or a printer. In actuality, scale pixels are extracted as the boundary between the air region and the water region. Therefore, for example, the height from the river bottom may be obtained based on the position (array rank) of the extracted scale pixels and the attribute information of the water gauge WM (particularly, the relationship between the position of the scale and the distance from the end of the water gauge WM), and this height may be output as the water level of the river.

[0042] By the way, when a plurality of reference lines LS are set, the air region and the water region may be determined for each reference line LS, that is, a plurality of different water levels may be estimated. In this case, it is advisable to estimate the final water level (so-called, determined water level) based on the plurality of estimated water levels (so-called, provisional water levels). When estimating the final water level, various statistical processes such as obtaining an average value, obtaining a median value, or obtaining a mode value may be performed.

[0043] (Flow of processing) Hereinafter, the main processing of the water level estimation system 100 will be described in detail with reference to FIG. 6. FIG. 6 is a flowchart showing an example of the flow of the main processing of the water level estimation system 100, showing the processing to be executed in the central column, the things necessary for that processing in the left column, and the things resulting from that processing in the right column.

[0044] In order to estimate the water level of the river by the water level estimation system 100, first, as shown in FIG. 6, a "label image PM" is detected from the on-site image PS using the label detection means 107 (Step 201 in FIG. 6), and further, "scale pixels" are detected from the label image PM using the scale detection means 101 (Step 202 in FIG. 6).

[0045] When scale pixels are detected from the identification image PM, after setting the reference line LS (Step 203 in FIG. 6), a "scale pixel column" is set using the scale luminance value calculation means 102 (Step 204 in FIG. 6), and the luminance value related to the scale pixels constituting the scale pixel column is obtained (Step 205 in FIG. 6). Also, a "representative scale luminance value" for each scale is calculated using the representative luminance value calculation means 105, or a "moving average luminance value" is calculated using the average luminance value calculation means 106.

[0046] When the luminance value related to the scale pixels is obtained and the representative scale luminance value and the moving average luminance value are calculated, a "luminance change value" is calculated using the luminance change value calculation means 103 (Step 206 in FIG. 6). At this time, the "change amount of the luminance value" can be used as the luminance change value, or the "change interval of the luminance value" or the "change gradient of the luminance value" can be used as the luminance change value. Also, as described above, the luminance change value can be calculated using the luminance value related to the scale pixels selected for each scale, or the representative scale luminance value or the moving average luminance value can be used to calculate the luminance change value.

[0047] When the luminance change value is obtained, the air region and the underwater region of the water level gauge WM are determined using the region determination means 104 (Step 207 in FIG. 6). Then, the water level of the river is output using the water level output means 108 (Step 208 in FIG. 6). At this time, it is preferable to output the "height from the river bottom" obtained based on the position of the extracted scale pixels and the attribute information of the water level gauge WM as the water level of the river.

Industrial Applicability

[0048] The water level estimation system of the present invention and the river flow rate estimation system of the present invention are particularly useful for river administrators such as countries and local governments. Considering that the present invention can quickly predict flood disasters and other water disasters associated with rivers, and as a result, can protect many residents from disasters, it can be said that the present invention is not only industrially applicable but also expected to make a great social contribution.

Explanation of Reference Numerals

[0049] 100 Water level estimation system of the present invention 101 Scale detection means (of the water level estimation system) 102 Scale luminance value calculation means (of the water level estimation system) 103 Luminance change value calculation means (of the water level estimation system) 104 Region determination means (of the water level estimation system) 105 Representative luminance value calculation means (of the water level estimation system) 106 Average luminance value calculation means (of the water level estimation system) 107 Mark detection means (of the water level estimation system) 108 Water level output means (of the water level estimation system) 109 Local image storage means (of the water level estimation system) 110 Mark information storage means (of the water level estimation system) LS Reference line PS Local image PM Mark image SB Black scale SW White scale WM Water gauge

Claims

1. A system for estimating a water level based on a on-site image containing a water level marker with a part inserted into water, comprising: On the water level marker, two or more scales with different luminance values are displayed at equal intervals, and a plurality of such scales in the on-site image are arranged in a vertical or substantially vertical direction; Scale detection means for detecting scale pixels constituting the scales from the on-site image; Scale luminance value calculation means for setting a scale pixel column in which the scale pixels detected by the scale detection means are arranged in a single column in order from below or above, and obtaining a luminance value for each scale pixel included in the scale pixel column; Luminance change value calculation means for calculating a luminance change value indicating the degree of change in the luminance values of the scale pixels arranged vertically in the scale pixel column; Region determination means for determining an air region and a water region of the water level marker based on the luminance change value extracted by the luminance change value calculation means; and Estimating the boundary between the air region and the water region determined by the region determination means as the water level. A water level estimation system characterized by the above.

2. When the luminance values of the scale pixels arranged vertically in the scale pixel column are the same or substantially the same, a scale pixel group is set with the scale pixels being in the same scale, and representative luminance value calculation means for calculating a representative scale luminance value by statistically processing the luminance values of the plurality of scale pixels constituting the scale pixel group is further provided; The luminance change value calculation means calculates the degree of change in the representative scale luminance values adjacent vertically as the luminance change value. The water level estimation system according to Claim 1, characterized by the above.

3. Average luminance value calculation means for calculating a moving average luminance value, which is the average value of the luminance values of a predetermined number of the scale pixels, while moving the predetermined number of the scale pixels in order from below or above with respect to the scale pixel column is further provided; The luminance change value calculation means calculates the degree of change in the moving average luminance values adjacent vertically in the scale pixel column as the luminance change value. The water level estimation system according to Claim 1, characterized by the above.

4. The luminance change value calculation means calculates the amount of change in the luminance value of the scale pixel as the luminance change value, and the amount of change is the difference in the luminance values of the scale pixels arranged vertically in the scale pixel column. The area determination means determines, as the underwater area, an area in the scale pixel column where the luminance change value is smaller than a preset change amount threshold value. The water level estimation system according to claim 1, characterized in that.

5. The luminance change value calculation means calculates, as the luminance change value, the change interval of the luminance value related to the scale pixel. The change interval is the distance between the scale pixel indicating the maximum luminance value and the scale pixel indicating the minimum luminance value in the scale pixel column. The area determination means determines, as the underwater area, an area in the scale pixel column where the luminance change value is smaller than a preset change interval threshold value. The water level estimation system according to claim 1, characterized in that.

6. The scale luminance value calculation means sets a plurality of columns of the scale pixel columns according to the horizontal width dimension of the scale displayed on the water level indicator. The area determination means determines the air area and the underwater area of the water level indicator for each of the scale pixel columns set by the scale luminance value calculation means. Based on a plurality of sets of the air area and the underwater area determined by the area determination means, a plurality of water levels are obtained, and the water level is estimated by statistically processing the plurality of water levels. The water level estimation system according to claim 1, characterized in that.

Images