Submersion depth determination program, submersion depth determination device, and submersion depth determination method

Abstract

Determining flood water depth from an image of a flooded situation as captured by a computer is difficult because no information exits for the conditions during normal periods. In the present invention, a flood water depth determination program causes a computer to execute a process of detecting a first type of object included in a first captured image and a first flood water position of the object, referencing depth information that associates the depth with a combination of the type and the flood water position, and outputting a first flood water depth that corresponds to the first type and the first flood water position. As a result, the computer is able to output the flood water depth from the image in which the flooded situation is captured even without information on the conditions during normal periods.

Description

Technical Field

[0001] The present invention relates to a technique for determining the depth of inundation.

Background Art

[0002] In flood prevention and disaster reduction activities during floods and inundations, information on the depth of inundation is important. More specifically, information on the depth of inundation distribution is important as a basis for determining which locations should be patrolled, which residents need to evacuate, and whether there are locations that require flood prevention activities. A knowledgeable expert visually determines the depth of inundation based on knowledge of the pre-inundation state of the location.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Patent Document 2

Summary of the Invention

Problems to be Solved by the Invention

[0004] However, it is difficult to determine the depth of inundation from an image captured of a flooded system because there is no information on the normal state.

[0005] On one aspect, it is an object to output the depth of inundation from a captured image even without information on the normal state.

Means for Solving the Problems

[0006] In one aspect, an inundation depth determination program causes a computer to execute a process of detecting a first type of object and a first inundation position of the object included in a first captured image, referring to depth information in which a depth is associated with a pair of the type and the inundation position, and outputting a first inundation depth corresponding to the first type and the first inundation position.

Advantages of the Invention

[0007] On one hand, even without information in normal state, the immersion depth can be output from the captured image.

Brief Description of the Drawings

[0008]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Modes for Carrying Out the Invention

[0009] Hereinafter, examples of the immersion depth determination program, the immersion depth determination device, and the immersion depth determination method according to the present embodiment will be described in detail with reference to the drawings. Note that the present embodiment is not limited by this example. Also, each example can be appropriately combined within a non - conflicting range.

[0010] [Functional Configuration of the Determination Device 10] Using FIG. 1, the functional configuration of the determination device 10 according to the present embodiment will be described. FIG. 1 is a block diagram showing a configuration example of the determination device 10. As shown in FIG. 1, the determination device 10 includes a communication unit 11, a storage unit 12, and a control unit 13.

[0011] The communication unit 11 is a processing unit that controls communication with other information processing devices.

[0012] The storage unit 12 is a storage device that stores various data and programs executed by the control unit 13. The storage unit 12 stores an image DB (database) 121, a detection model DB 122, a depth information table 123, and the like.

[0013] The image DB 121 stores images captured of the flooding situation. FIG. 2 is a diagram showing an example of a captured image according to the present embodiment. The captured image shown in FIG. 2 may be an image captured by a camera device or a smartphone equipped with a camera function. Further, the captured image may be an image captured by, for example, a resident in a flooded area and uploaded onto the Internet via an SNS (Social Networking Service) or the like.

[0014] The detection model DB 122 stores parameters for constructing a machine learning model generated by machine learning using the captured image as a feature amount and the type of object included in the captured image as a correct label, and training data for the model. Here, the type of object is, for example, a standing person, a squatting person, a sitting person, a car, a building, a utility pole, and the like.

[0015] Further, the detection model DB 122 stores parameters for constructing a machine learning model generated by machine learning using the object included in the captured image as a feature amount and the flooding position of the object as a correct label, and training data for the model. Here, the flooding position of the object is, for example, if the type of object is a standing person, any of below the knees, above the knees, up to the waist, and up to the shoulders; if the type of object is a squatting person or a sitting person, up to the waist or up to the shoulders. Further, the flooding position of the object is, for example, if the type of object is a car, any of up to the tires, up to the windows, and submersion; if the type of object is a building, any of under the floor, on the floor, up to the first floor, and up to the second floor. Further, the flooding position of the object is, for example, if the type of object is a utility pole, any of up to the street name and number display, up to the wrapped advertisement, and up to the hanging advertisement. Note that the model may be generated for each type of object.

[0016] The depth information table 123 stores information associating the depth with the combination of the type of target and the flooding position. FIG. 3 is a diagram showing an example of the depth information table 123 according to the present embodiment. As shown in FIG. 3, in the depth information table 123, for example, the type of target, the flooding position, and the flooding depth are set in association with each other. The depth information table 123 shown in FIG. 3 indicates that, for example, when the type of target included in the captured image is an upright person and the flooding reaches below the knees, the flooding depth is 30 cm. The determination device 10 can refer to the depth information table 123 using, for example, the type of target and the flooding position detected from the captured image as search keys to search for the flooding depth. Note that the depth information table 123 shown in FIG. 3 is merely an example, and each set value is not limited to the example in FIG. 3. Also, the depth information table 123 may be generated by collecting captured images of past flooding situations, detecting the type of target and the flooding position from the captured images using a machine learning model, and associating the actual flooding depth with the combination of the detected type of target and the flooding position.

[0017] Note that the above information stored in the storage unit 12 is merely an example, and the storage unit 12 can store various other information in addition to the above information.

[0018] The control unit 13 is a processing unit that controls the entire determination device 10, and includes an acquisition unit 131, a detection unit 132, and an output unit 133.

[0019] The acquisition unit 131 acquires an image capturing the flooding situation. For example, the acquisition unit 131 collects captured images uploaded onto the Internet via SNS or the like and stores them in the image DB 121. Note that the captured images may be collected for each area or time zone where flooding has occurred.

[0020] The detection unit 132 detects the type of the object and the flooding position included in the captured image acquired by the acquisition unit 131. For example, the detection unit 132 takes the captured image as a feature amount, and inputs the captured image into a machine learning model generated by machine learning with the type of the object included in the captured image as the correct label, thereby detecting the type of the object. Further, for example, the detection unit 132 takes the object included in the captured image as a feature amount, and inputs the object included in the captured image into a machine learning model generated by machine learning with the flooding position of the object as the correct label, thereby detecting the flooding position of the object.

[0021] The output unit 133 refers to the depth information table 123 using the type of the object and the flooding position detected by the detection unit 132 as search keys, searches for the flooding depth, and outputs it. The output of the flooding depth by the output unit 133 may be performed via an output device such as a display device connected to the determination device 10, for example.

[0022] FIG. 4 is a diagram showing an example of the flooding depth output according to the present embodiment. As shown in FIG. 4, the output unit 133 can, for example, surround the detected object in the captured image with a bounding box and output it in association with the type of the object, the flooding position, and the flooding depth.

[0023] [Flow of processing] Next, with reference to FIG. 5, the flow of the flooding depth determination process by the determination device 10 will be described. FIG. 5 is a flowchart showing an example of the flow of the flooding depth determination process according to the present embodiment.

[0024] First, as shown in FIG. 5, the determination device 10 acquires an image that has captured the flooding situation (step S101). The acquisition of the captured image in step S101 may be to acquire in real time a captured image uploaded onto the Internet via an SNS or the like, or to acquire a captured image stored in the image DB 121 collected in advance.

[0025] Next, the determination device 10 inputs the captured image acquired in step S101 into the machine learning model, and detects the type of the object included in the captured image (step S102). The machine learning model used in step S102 is, for example, a machine learning model generated by machine learning using the captured image as a feature amount and the type of the object included in the captured image as a correct label.

[0026] Next, the determination device 10 inputs the object included in the captured image acquired in step S101 into the machine learning model, and detects the flooding position of the object included in the captured image (step S103). The machine learning model used in step S103 is, for example, a machine learning model generated by machine learning using the object included in the captured image as a feature amount and the flooding position of the object as a correct label.

[0027] Next, the determination device 10 refers to the depth information table 123 using the type and the flooding position of the object acquired in each of steps S103 and S104 as search keys, searches for and outputs the flooding depth (step S104). After the execution of step S104, the flooding depth determination process shown in FIG. 5 ends.

[0028] [Effect] As described above, the determination device 10 detects the first type of the object included in the first captured image and the first flooding position of the object, refers to the depth information associating the depth with the combination of the type and the flooding position, and outputs the first flooding depth corresponding to the first type and the first flooding position.

[0029] In this way, the determination device 10 detects the type and the flooding position of the object included in the image capturing the flooding situation, and refers to the depth information associated with the combination of the type and the flooding position to output the flooding depth. Thereby, the determination device 10 can output the flooding depth from the captured image even without information on the normal state.

[0030] In addition, the process of detecting the first type executed by the determination device 10 includes a process of detecting the first type by inputting the first captured image into a first machine learning model generated by machine learning using the captured image as a feature amount and the type as a correct label.

[0031] Thereby, the determination device 10 can more accurately detect the type of the object included in the image capturing the flooding situation.

[0032] In addition, the process of detecting the first flooding position executed by the determination device 10 includes a process of detecting the first flooding position by inputting the object included in the first captured image into a second machine learning model generated by machine learning using the object included in the captured image as a feature amount and the flooding position as a correct label.

[0033] Thereby, the determination device 10 can more accurately detect the flooding position of the object included in the image capturing the flooding situation.

[0034] In addition, the process of detecting the first type executed by the determination device 10 includes a process of detecting at least one of a standing person, a squatting person, a sitting person, a vehicle, a building, and a utility pole as the first type.

[0035] Thereby, the determination device 10 can detect the types of various objects included in the image capturing the flooding situation and output the flood depth based on the detected types.

[0036] In addition, the process of detecting the first flooding position executed by the determination device 10 includes detecting at least one of below the knees, above the knees, up to the waist, up to the shoulders if the first type is a standing person, up to the waist or up to the shoulders if the first type is a squatting person or a sitting person, up to the tires, up to the windows, and inundation if the first type is a vehicle, under the floor, on the floor, up to the first floor, up to the second floor if the first type is a building, up to the geographical name and number display, up to the rolled advertisement, and up to the hanging advertisement if the first type is a utility pole as the first flooding position.

[0037] Accordingly, the determination device 10 can detect the flooding positions of various objects included in the image capturing the flooding situation, and output the flooding depth based on the detected flooding positions.

[0038] [System] The processing procedures, control procedures, specific names, information including various data and parameters shown in the above documents and drawings may be arbitrarily changed unless otherwise specified. Also, the specific examples, distributions, numerical values, etc. described in the embodiments are merely examples and may be arbitrarily changed.

[0039] Also, the specific forms of the dispersion and integration of the components of the determination device 10 are not limited to those shown. For example, the detection unit 132 of the determination device 10 may be dispersed into a plurality of processing units, or the detection unit 132 and the output unit 133 of the determination device 10 may be integrated into one processing unit. That is, all or part of its components may be functionally or physically dispersed and integrated in any unit according to various loads, usage situations, etc. Furthermore, each processing function of each device may be realized in whole or in any part by a CPU and a program analyzed and executed by the CPU, or may be realized as hardware by wired logic.

[0040] FIG. 6 is a diagram showing a hardware configuration example of the determination device 10 according to the present embodiment. As shown in FIG. 6, the determination device 10 includes a communication interface 10a, a HDD (Hard Disk Drive) 10b, a memory 10c, and a processor 10d. Also, each part shown in FIG. 6 is mutually connected by a bus or the like.

[0041] The communication interface 10a is a network interface card or the like and performs communication with other information processing devices. The HDD 10b stores programs and data for operating the functions shown in FIG. 1 and the like.

[0042] The processor 10d is a CPU (Central Processing Unit), MPU (Micro Processing Unit), GPU (Graphics Processing Unit), etc. Further, the processor 10d may be realized by an integrated circuit such as an ASIC (Application Specific Integrated Circuit) or FPGA (Field Programmable Gate Array). The processor 10d is a hardware circuit that executes a process of realizing each function described in FIG. 1, etc., by reading a program that executes the same processing as each processing unit shown in FIG. 1, etc., from the HDD 10b, etc., and expanding it in the memory 10c.

[0043] In addition, the determination device 10 can also realize the same functions as those in the above embodiments by reading the above program from the recording medium by the medium reading device and executing the read program. Note that the program in this other embodiment is not limited to being executed by the determination device 10. For example, the above embodiments may be similarly applied when another information processing device executes the program, or when these cooperate to execute the program.

[0044] This program may be distributed via a network such as the Internet. Further, this program may be recorded on a recording medium readable by an information processing device such as a hard disk, flexible disk (FD), CD-ROM, MO (Magneto-Optical disk), DVD (Digital Versatile Disc), etc., and executed by being read from the recording medium by the information processing device.

Explanation of Reference Numerals

[0045] 10 Determination device 10a Communication interface 10b HDD 10c Memory 10d Processor 11 Communication unit 12 Memory unit 13 Control unit 121 Image DB 122 Detection model DB 123 Depth information table 131 Acquisition unit 132 Detection unit 133 Output unit

Claims

1. Detect the first type of the object included in the first captured image and the first flooding position of the object, Refer to the depth information associating the depth with the pair of the type and the flooding position, and output the first flooding depth corresponding to the first type and the first flooding position, Cause a computer to execute the process, The process of detecting the first type includes inputting the first captured image into a first machine learning model generated by machine learning using the captured image as a feature amount and the type of the object included in the captured image as a correct label, thereby detecting the first type, The process of detecting the first flooding position includes inputting the object of the first type detected using the first machine learning model into a second machine learning model generated by machine learning using the object included in the captured image as a feature amount and different flooding positions for each type as correct labels, thereby detecting the first flooding position, Flooding depth determination program.

2. The process of detecting the first type includes detecting at least one of a standing person, a squatting person, a sitting person, a vehicle, a building, and a utility pole as the first type, The flooding depth determination program according to claim 1, characterized in that.

3. The process of detecting the first flooding position includes, if the first type is a standing person, any one of below the knees, above the knees, up to the waist, and up to the shoulders; if the first type is a squatting person or a sitting person, up to the waist or up to the shoulders; if the first type is a vehicle, any one of up to the tires, up to the windows, and submersion; if the first type is a building, any one of under the floor, on the floor, up to the first floor, and up to the second floor; if the first type is a utility pole, any one of up to the street name and number display, up to the rolled advertisement, and up to the hanging advertisement, including the process of detecting at least one of them as the first flooding position, The flooding depth determination program according to claim 2, characterized in that.

4. Detect the first type of the object included in the first captured image and the first flooding position of the object, Refer to the depth information associating the depth with the pair of the type and the flooding position, and output the first flooding depth corresponding to the first type and the first flooding position, Comprising a control unit that executes the process, The front control unit, Using the captured image as a feature amount, input the first captured image into a first machine learning model generated by machine learning using the type of the object included in the captured image as a correct label, thereby detecting the first type, A water depth determination device that inputs the object of the first type detected using the first machine learning model into a second machine learning model generated by machine learning, with the object included in the captured image as a feature amount and different water immersion positions for each type as correct labels, to detect the first water immersion position.

5. Detect the first type of the object included in the first captured image and the first water immersion position of the object, Refer to depth information associating a depth with a pair of a type and a water immersion position, and output a first water depth corresponding to the first type and the first water immersion position, The computer executes the process, The process of detecting the first type includes the process of inputting the first captured image into a first machine learning model generated by machine learning, with the captured image as a feature amount and the type of the object included in the captured image as a correct label, to detect the first type. The process of detecting the first water immersion position includes the process of inputting the object of the first type detected using the first machine learning model into a second machine learning model generated by machine learning, with the object included in the captured image as a feature amount and different water immersion positions for each type as correct labels, to detect the first water immersion position. A water depth determination method.

Images