Water depth measuring system

Abstract

To estimate a water depth of flooding efficiently.SOLUTION: Imaging means 12a mounted on a vehicle 12 captures images of a space outside the vehicle at a first position and a second position during flooding and normal conditions, to acquire a first captured image and a second captured image. Feature point detection means 16a detects a feature point on a boundary between a flooded surface and an object from the first captured image and the second captured image. Height calculation means 16b calculates a height from the feature point to the imaging means 12a. Water depth estimation means 16c estimates a depth of flooding by comparing the height calculated by the height calculation means 16b with a height from the ground to the imaging means 12a during the normal conditions.SELECTED DRAWING: Figure 1

Description

Technical Field

[0001] This specification discloses a water depth measurement system.

Background Art

[0002] Patent Document 1 discloses an inference program that can estimate the water depth at other locations if there are measurement points at some locations without measuring the water depth at all locations in the area where the immersion depth is to be grasped. The model of the inference program in Patent Document 1 is learned from the measured values of the water depth and fluid analysis.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] The purpose of the water depth measurement system disclosed in this specification is to efficiently estimate the water depth of flooding.

Means for Solving the Problems

[0005] The water depth measurement system disclosed in this specification includes imaging means mounted on a vehicle, which acquires a first captured image by imaging the external space of the vehicle at a first position during flooding, and acquires a second captured image by imaging the external space of the vehicle at a second position spaced apart from the first position; feature point detection means for detecting feature points on the boundary between the water surface and an object from the first captured image and the second captured image; height calculation means for calculating the height from the feature points to the imaging means; and water depth estimation means for estimating the water depth of the flooding by comparing the height calculated by the height calculation means with the height from the ground to the imaging means during normal times.

Effects of the Invention

[0006] According to the water depth measurement system disclosed in this specification, the water depth of flooding can be efficiently estimated.

Brief Description of the Drawings

[0007]

Figure 1

Figure 2

Figure 3

Modes for Carrying Out the Invention

[0008] FIG. 1 is a schematic configuration diagram of a water depth measurement system 10 according to this embodiment. The water depth measurement system 10 includes one or more vehicles 12, a data storage server 14, a computing server 16, and a user terminal 18.

[0009] The vehicle 12 is provided with imaging means 12a such as a camera. The external space of the vehicle 12 is imaged by the imaging means 12a. Specifically, when there is flooding, the imaging means 12a captures an imaging image including the water surface of the flooding (hereinafter referred to as the "flood water surface"), and when it is normal (when there is no flooding), the imaging means 12a captures an imaging image including the ground. Further, the imaging means 12a acquires a first imaging image by imaging at a first position, and after the vehicle 12 moves, acquires a second imaging image by imaging at a second position different from the first position. The imaging image captured by the imaging means 12a is transmitted to the data storage server 14 by wireless communication or the like and stored and accumulated in the data storage server 14. The arithmetic server 16 estimates the depth of the flood water (hereinafter simply referred to as the "water depth") based on the imaging images stored in the data storage server 14. The arithmetic server 16 includes feature point detection means 16a, height calculation means 16b, and water depth estimation means 16c. Details of the water depth estimation process by the arithmetic server 16 will be described later. The user terminal 18 is a computer used by the user of the water depth measurement system 10, and is, for example, a PC or a smartphone. The user transmits conditions such as the area and time for which the user wants to obtain the estimated result of the water depth as a request to the arithmetic server 16, and receives the estimated result by the arithmetic server 16 at the user terminal 18.

[0010] FIG. 2 is a flowchart showing the processing flow of the water depth measurement system 10 according to the present embodiment. First, with reference to FIG. 2(a), a general flow of the processing of the water depth measurement system 10 will be described.

[0011] In step S10, the arithmetic server 16 identifies the area to be analyzed, that is, the flooded road. It is identified by methods such as image classification and CAN data analysis. This is because it is unrealistic in terms of cost to continue the analysis 365 days a year, 24 hours a day.

[0012] In step S12, the arithmetic server 16 acquires the imaging images at the time of flooding in the identified range from the data storage server 14. In step S14, the arithmetic server 16 acquires the imaging images in the normal time in the identified range from the data storage server 14.

[0013] In step S16, the arithmetic server 16 calculates the height from the water surface during flooding to the imaging means 12a based on the captured image during flooding obtained in step S12. Also, in step S16, the arithmetic server 16 calculates the height from the ground to the imaging means 12a based on the captured image during normal times obtained in step S14. The process of step S16 may be executed for each vehicle included in the water depth measurement system 10. Details of the process of step S16 will be described later with reference to the flowchart of FIG. 2(b).

[0014] In step S18, the arithmetic server 16 estimates the water depth based on the height from the water surface during flooding to the imaging means 12a and the height from the ground to the imaging means 12a calculated in step S16. Details of the process of step S18 will be described later with reference to the flowchart of FIG. 2(c).

[0015] Next, with reference to FIG. 2(b), details of the process of step S16 will be described.

[0016] To calculate the height from the water surface during flooding to the imaging means 12a, two captured images (the first captured image and the second captured image described above) captured at different positions (the first position and the second position described above) are required. Therefore, in step S30, the arithmetic server 16 determines the first captured image and the second captured image from among a plurality of captured images captured by the imaging means 12a of one vehicle 12. In order to improve the calculation accuracy, the moving distance (the distance between the first position and the second position) should not be too short or too long, and it is necessary to compare the images at an optimal moving distance. Therefore, the arithmetic server 16 calculates the moving distance using data such as the vehicle speed and determines an optimal comparison frame for a certain frame.

[0017] Also, to calculate the height from the water surface during flooding to the imaging means 12a, the focal length of the captured image and the mounting height of the imaging means 12a are required. Also, in order to correct the distortion of the captured image, other imaging parameters are required. Therefore, in step S32, the arithmetic server 16 reads these parameters.

[0018] The processes from step S34 to S44 are respectively executed for the first captured image and the second captured image. The processes from step S34 to S42 are the processes of detecting feature points by the feature point detection means 16a, and the process of step S44 is the process of calculating the height from the feature points to the imaging means 12a by the height calculation means 16b.

[0019] In step S34, the feature point detection means 16a determines the boundary in order to identify the feature points on the boundary between the crown water surface and houses or other objects. Here, the region division method is used to identify the boundary. In addition to the region division method, the boundary may be identified using an edge detection method or the like after object detection.

[0020] In step S36, the feature point detection means 16a determines the boundary line by performing edge detection in the vicinity of the boundary line obtained in step S34. This is because the boundary line obtained by the region division method is often ambiguous.

[0021] In step S38, the feature point detection means 16a detects feature points by applying the prior art. Also, in step S40, the feature point detection means 16a performs feature point matching by applying the prior art. Further, in step S42, the feature point detection means 16a selects only the points on the boundary specified in step S36 from among those for which feature point matching was possible in step S40. The feature points thus selected are the crown water surface feature points on the boundary between the crown water surface and the object.

[0022] In step S44, the height calculation means 16b calculates the height from the feature points acquired in step S42 to the imaging means 12a (that is, the height from the crown water surface to the imaging means 12a).

[0023] Fig. 3(a) shows parameters for calculating the height. In Fig. 3(a), reference numeral 12-1 represents the position of the vehicle 12 when the imaging means 12a acquires the first captured image (the position of the imaging means 12a when the vehicle 12 is at this position is the first position), and reference numeral 12-2 represents the position of the vehicle 12 when the imaging means 12a acquires the second captured image (the position of the imaging means 12a when the vehicle 12 is at this position is the second position). Le-1 and Im-1 are the positions of the lens and the image sensor of the imaging means 12a when the imaging means 12a is at the first position, and Le-2 and Im-2 are the positions of the lens and the image sensor of the imaging means 12a when the imaging means 12a is at the second position. d is the distance between the first position and the second position, f is the focal length of the imaging means 12a, and FP is a feature point. Also, Y 1 is the Y coordinate of the feature point FP in the first captured image, and Y 2 is the Y coordinate of the feature point FP in the second captured image. Further, X is the horizontal distance from the position of the imaging means 12a at the first position to the feature point FP, and H cf is the height from the feature point FP to the imaging means 12a, that is, the value to be calculated in step S44. Among the above parameters, values other than H cf and X are known values.

[0024] In Fig. 3(a), first, the following equations (1) and (2) hold.

Equation

Equation

[0025] In the above description, the arithmetic server 16 has obtained the height from the crown water surface feature point on the boundary between the crown water surface and the object to the imaging means 12a. However, the arithmetic server 16 also obtains the height from the ground feature point on the boundary between the ground and the object to the imaging means 12a by the same process.

[0026] Next, with reference to FIG. 2(c), the details of the process of step S18 will be described.

[0027] The flow in FIG. 2(c) executes the process for each point where the height from the ground calculated using the normal-time imaging image to the imaging means 12a could be calculated.

[0028] In step S60, the water depth estimation means 16c searches for the height from the crown water surface to the imaging means 12a that could be estimated at the same location as the location where the height from the ground to the imaging means 12a could be estimated. If not available, it searches for the estimation result at the nearest location.

[0029] In step S62, the water depth estimation means 16c estimates the water depth by comparing the height from the crown water surface to the imaging means 12a read in step S60 with the height from the ground to the imaging means 12a in normal times.

[0030] In FIG. 3(b), the parameters for calculating the height are shown. The left figure in FIG. 3(b) is the figure in normal times, and the right figure is the figure at the time of crown water. H cn is the height from the ground feature point FPg to the imaging means 12a, and H cf is the height from the crown water surface feature point FP fr to the imaging means 12a. Also, H cr1 is the height from the road in normal times to the imaging means 12a, and H cr2 is the height from the road at the time of crown water to the imaging means 12a. Further, H g is the height from the road surface to the ground feature point FP g to FP fr is the height from the road surface to the crown water surface feature point FP fr to FP fg is the water depth to be obtained. Among the above parameters, H cr1 and H cr2 are known values.

[0031] H fg is represented by the following formula (4).

Mathematics

Mathematics

Mathematics

[0032] As described above, the embodiments of the water depth measurement system according to the present disclosure have been described. However, the water depth measurement system according to the present disclosure is not limited to the above embodiments, and various changes can be made without departing from the spirit thereof.

[0033] For example, the arithmetic server 16 may detect an object with a known height or shape reflected in the imaging means 12a, narrow down the water depth, and guarantee the reliability of the estimated result of the water depth. For example, in the captured image taken during flooding, if a guardrail with a known height is visible, the water depth can be narrowed down to be less than or equal to the height of the guardrail. Also, in the captured image taken during flooding, if an object with a known height cannot be seen, the water depth can be narrowed down to be greater than or equal to the height of the object. Further, the ratio of an object with a known height and shape (e.g., a tire) immersed in water may be estimated. For example, when 30% of the tire is submerged, the outer shape of the tire × 0.3 is the water depth. Note that by specifying the vehicle type through object detection, the outer shape of the tire can be narrowed down and the error can be reduced. Furthermore, an object with a known dimension may be used as a measuring stick to calculate the distance from an arbitrary point to the ground. For example, by counting the number of pixels in which a license plate with a known dimension appears, the spatial resolution in the depth direction can be known. The water depth may be estimated by calculating the distance from an arbitrary point to the ground during normal times and the distance from the same point to the water surface during flooding and taking the difference.

[0034] In addition, although the present embodiment has described the estimation of water depth, the present invention can also be applied to the estimation of the height or depth of objects, not limited to the water depth, such as the depth of snow accumulation or the height of vegetation on the road.

Description of Reference Numerals

[0035] 10 Water depth measurement system, 12 Vehicle, 12a Imaging means, 14 Data storage server, 16 Computation server, 16a Feature point detection means, 16b Height computation means, 16c Water depth estimation means, 18 User terminal.

Claims

【Claim 1】 Imaging means mounted on a vehicle, which acquires a first imaging image by imaging the external space of the vehicle at a first position during flooding, and acquires a second imaging image by imaging the external space of the vehicle at a second position spaced apart from the first position; Feature point detection means for detecting a feature point on the boundary between the flood surface and an object from the first imaging image and the second imaging image; Height calculation means for calculating the height from the feature point to the imaging means; Water depth estimation means for estimating the water depth of the flooding by comparing the height calculated by the height calculation means with the height from the ground to the imaging means during normal times; A water depth measurement system, characterized by comprising the above.

Images