Analysis device and program

Abstract

To provide an analysis device capable of calculating a volume as an occupation range of an aquatic plant from a point group measured for an underwater object.SOLUTION: The analysis device includes: a classification unit 7 for applying classification to a point group measured for an underwater object to obtain each partial point group corresponding to an individual as one object in which an aquatic plant occupies a continuous range; and a volume calculation unit 9 for calculating a volume of an underwater range occupied by the corresponding aquatic plant for each partial point group.SELECTED DRAWING: Figure 4

Description

[Technical Field]

[0001] The present invention relates to an analysis device and a program for calculating the volume of an area occupied by aquatic plants from a point cloud acquired about an underwater object. [Background technology]

[0002] In marine ecosystems where seaweeds such as eelgrass and kelp (hereinafter referred to as "seaweed") exist, carbon dioxide (CO2) is absorbed through photosynthesis and other processes, with some of the absorbed CO2 remaining in the soil and seawater. The Japan Blue Economy Technology Research Association (JBE) has established the J Blue Credit (registered trademark) scheme to target CO2 absorption by marine ecosystems. This scheme certifies and issues credits equivalent to the amount of CO2 absorbed, and aims to promote credit trading with companies or organizations seeking to reduce CO2 emissions.

[0003] A company or organization wishing to issue credits (in other words, a credit seller) submits an application for credits to JBE, for example, along with the results of seaweed bed observations. JBE then issues an appropriate amount of credits to the applicant based on the results of the seaweed bed observations. JBE has published an application guide (for example, Non-Patent Document 1), which states that the amount of CO2 absorbed by a seaweed bed (also known as the amount of blue carbon) is determined by the product of the area of the seaweed bed and the CO2 absorption coefficient. While it is permissible to use values found in literature or elsewhere for the absorption coefficient, actual measurement of the seaweed bed's area is required. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Japanese Patent Application Publication No. 2023-054691 [Non-patent literature]

[0005] [Non-Patent Document 1] Japan Blue Economy Technology Research Association (JBE), "J Blue Credit® Certification Application Guide - Climate Change Countermeasures Using Blue Carbon - Ver. 2.3," [online], August 2023, [Retrieved December 19, 2023], Internet <https: / / www.blueeconomy.jp / wp-content / uploads / jbc2023 / 20230816_J-BlueCredit_Guideline_v2.3.pdf> Summary of the Invention [Problem to be solved by the invention]

[0006] In order to implement climate change countermeasures, it is desirable to conduct credit trading appropriately. To do this, it is necessary to accurately estimate the amount of blue carbon, and as a prerequisite data, it is necessary to accurately determine the distribution of aquatic plants such as seaweed in seaweed beds.

[0007] However, the guide estimates the amount of CO2 absorbed as "seaweed bed area x absorption coefficient = CO2 absorption amount," and only uses the two-dimensional distribution of "seaweed bed area" for the distribution of aquatic plants. Aquatic plants are not only distributed in a two-dimensional plane, but also in the depth direction of the water, and are distributed in three dimensions as a volume occupied underwater, but conventional technology has not taken this three-dimensional distribution into account.

[0008] Meanwhile, point clouds are a conventional technology that can relatively easily obtain 3D distribution data based on video and other sources. For example, in Patent Document 1, an airspace in which a drone can fly is set by processing a point cloud obtained by 3D measuring the area around a steel tower. However, such conventional point cloud techniques are based on the premise of handling terrestrial data and are not designed to handle the distribution of aquatic plants that exist underwater, making it impossible to accurately determine the distribution.

[0009] In view of the above-mentioned problems with the conventional technology, the present invention aims to provide an analysis device and a program that can calculate the volume of the area occupied by aquatic plants from a point cloud measured for an underwater object. [Means for solving the problem]

[0010] To achieve the above object, the present invention provides an analysis device comprising: a classification unit that applies classification to a point cloud measured for an underwater object, and obtains partial point clouds corresponding to individual aquatic plants as one object occupying a continuous range; and a volume calculation unit that calculates the volume of the underwater area occupied by the corresponding aquatic plant for each of the partial point clouds. The present invention also provides a program that causes a computer to function as the analysis device. [Effects of the Invention]

[0011] According to the present invention, it is possible to classify the points measured for underwater objects into individuals that are single objects in which aquatic plants occupy a continuous range, and then calculate the volume of the area occupied by aquatic plants for each individual. [Brief explanation of the drawings]

[0012] [Figure 1] FIG. 1 is a functional block diagram of an analysis system according to an embodiment. [Figure 2] 1 is a flowchart illustrating an operation of an analysis system according to an embodiment. [Figure 3] FIG. 1 is a diagram showing an overall overview of the processing content of the analysis system, divided into examples EX1 to EX3. [Figure 4] FIG. 2 is a functional block diagram of an analysis unit (analysis device) according to an embodiment. [Figure 5] FIG. 10 is a diagram illustrating a process performed by a polyline estimation unit. [Figure 6] 10A to 10C are diagrams showing schematic examples EX61 to EX63 for explaining the processing of the upper cross section estimation unit. [Figure 7]10A and 10B are diagrams illustrating an example of a process performed by a point cloud distance estimation unit; [Figure 8] 1A to 1C are diagrams showing schematic diagrams of embodiments of the bottom separator; [Figure 9] 1 is a diagram showing a schematic diagram of an embodiment of a bottom separator; FIG. [Figure 10] 10 is a diagram showing an example in which a volume calculation unit calculates a volume of a three-dimensional model based on line segment information. FIG. [Figure 11] FIG. 10 is a diagram illustrating an example in which a volume calculation unit calculates the volume of a partial point cloud using a three-dimensional model. [Figure 12] FIG. 1 is a diagram illustrating a hardware configuration of a typical computer. DETAILED DESCRIPTION OF THE INVENTION

[0013] FIG. 1 is a functional block diagram of an analysis system 100 according to one embodiment. The analysis system 100 comprises a measurement unit 1, an analysis unit 2, a recognition unit 3, an absorption amount DB (database) 4, and a display unit 5. FIG. 2 is a flowchart showing the operation of the analysis system 100 according to one embodiment, and FIG. 3 is a diagram showing an overall outline of the processing contents of the analysis system 100, which has the configuration of FIG. 1 and performs processing according to the flow of FIG. 2, divided into examples EX1 to EX3. Below, the processing contents of each functional unit in FIG. 1 will be explained while explaining each step of FIG. 2. In this case, the schematic example of FIG. 3 will be referred to as appropriate.

[0014] In step S1, the measurement unit 1 captures underwater images in a specified water area range for which the amount of blue carbon is to be estimated, obtains underwater image data, and outputs the underwater image data to the analysis unit 2 and the identification unit 3, before proceeding to step S2.

[0015] As shown in example EX1 of Figure 3, the underwater video captured by the measurement unit 1 in step S1 is schematically illustrated. The hardware that realizes the measurement unit 1 is configured as equipment for capturing underwater video. For example, the measurement unit 1 can be configured as a series of capture equipment, such as an underwater camera CM configured to capture underwater images from a surface drone (ship) D via an underwater cable CB. The surface drone D is equipped with position measurement equipment such as a GPS (Global Positioning System), and acquires information on the measured position of the surface drone D, information on the length of the underwater cable CB lowered into the water, which is managed by a device that controls the length of the cable. This information includes information on the water depth at the location of the underwater camera CM and information on the capture direction of the underwater camera CM (which may include information on the angle of view as the capture range), at each time t = 1, 2, ... during capture. This allows the underwater camera CM to capture underwater images Pic(t) of aquatic plants PL, such as seaweed, at each time.

[0016] In step S2, an orthoimage of the target water area is generated from the underwater image obtained in step S1, and the identification unit 3 identifies the types of aquatic plants present in each predetermined section of the orthoimage before proceeding to step S3. Here, a series of images taken with the underwater camera CM within a pre-determined range determined to be shallow and facing the bottom, capturing images of the aquatic plants PL from above, can be stitched (image stitched) using any existing method to generate an orthoimage from the underwater image. Alternatively, a portion of the underwater image captured in step S1 can be captured by moving the underwater camera CM while maintaining its underwater depth, facing vertically or diagonally toward the bottom, as dedicated images for orthoimage generation, and an existing method of stitching can be used to generate an orthoimage.

[0017] The identification unit 3 uses training data in advance to construct a model for identifying aquatic plants in orthoimages using any existing method such as deep learning, and by using this trained model, it is possible to obtain identification results for aquatic plants present in each specified area within the orthoimage (for example, a specified area divided into a grid pattern of 1m wide x 1m high).

[0018] That is, by dividing the orthoimage into, for example, a grid, the identification result can be obtained for each area Dist(i,j) with abscissa i (i is an integer) and ordinate j (j is an integer) as follows: The aquatic plant in area D(1,1) is "Zostera marina". The aquatic plant in area D(100,100) is "Togemoku". The aquatic plants in area D (500,500) are "not applicable" (non-existent).

[0019] The classification result obtained by the classification unit 3 as in the above example may be output to the display unit 5 and also to the analysis unit 2. (Note that in an embodiment in which it is not necessary to use the classification result by the classification unit 3 in the analysis unit 2, outputting the classification result by the classification unit 3 to the analysis unit 2 may be omitted.)

[0020] In step S3, the analysis unit 2 generates a point cloud from the underwater image obtained in step S1, and then analyzes this point cloud to calculate the volume of the aquatic plants, outputting it to the display unit 5, and then proceeding to step S4. As shown in example EX2 of Figure 3, which schematically shows the volume calculation process by the analysis unit 2, the analysis unit 2 separates aquatic plants such as seaweed from the point cloud into individual pieces such as single pieces or clusters (individual pieces as a single object occupying a continuous range, which does not necessarily coincide with an individual piece of aquatic plant that is a living organism), and then calculates the volume of each individual piece.

[0021] The analysis unit 2 constitutes the analysis device 2 according to this embodiment, but will be described below as the analysis unit 2. The processing of the analysis unit 2 will be described in detail later.

[0022] In step S4, the CO2 absorption amount of the area Dist(i,j) is calculated using the identification results of the type of aquatic plant for each area Dist(i,j) obtained by the identification unit 3 in step S2, the volume information obtained by the analysis unit 3 in step S3, and the information on the CO2 absorption amount for each type of aquatic plant and occupied volume recorded in the absorption amount DB4, which is a pre-constructed database.The display unit 4 then displays and outputs the CO2 absorption amount of this area Dist(i,j), and the flow of Figure 2 is terminated.

[0023] As shown in example EX3 of Figure 3, the calculation and display process on the display unit 4 is schematically shown. If the type of aquatic plant in Dist(i,j) is K(i,j), its volume is V(i,j), and the CO2 absorption per volume of the aquatic plant of type K(i,j) recorded in the absorption amount DB4 is v(i,j), the CO2 absorption amount of the area Dist(i,j) can be calculated as v(i,j) * V(i,j). Note that if the type K(i,j) is "not applicable," it is assumed that there are no aquatic plants in the area Dist(i,j), and the CO2 absorption amount is calculated as zero.

[0024] The volume calculation by the analysis unit 2 (the analysis device 2 according to this embodiment) in step S3 will be described in detail below. Fig. 4 is a functional block diagram of the analysis unit 2 according to one embodiment for explaining the details. The analysis unit 2 includes a point cloud generation unit 6, a classification unit 7, a water bottom separation unit 8, and a volume calculation unit 9.

[0025] The point cloud generation unit 6 generates a point cloud of objects present in the water (transparent or translucent water itself is excluded from the objects) from a group of images {Pic(t)|t=1,2,...,N} at each time t=1,2,...,N (N is the total number of frame images of the underwater image) that constitute the underwater image obtained by the measurement unit 1, and outputs this point cloud to the classification unit 7 and the water bottom separation unit 8. Any existing method such as SfM (Structure from Motion; (3D) structure recovery from motion) may be used to generate a point cloud from the image. As is well known, SfM extracts a large number of feature points {(x1j ,y1 j )∈Pic(t1)|j=1,2,…},{(x2 k ,y2 k )∈Pic(t2)|k=1,2,…}, and the correspondence of the same feature points between different images is determined by determining whether the image features of the feature points match. j ,y1 j )←→(x2 k ,y2 k )”, that is, the j-th feature point (x1 j ,y1 j ) and the kth feature point (x2 k ,y2 k ) correspond to the same point, and then, using the principle of stereo vision, these identical feature points "(x1 j ,y1 j )←→(x2 k ,y2 k By obtaining the three-dimensional world coordinates (X, Y, Z) corresponding to each of the points "Pic(t)|t=1, 2, ..., N}", a large number of three-dimensional point clouds (X, Y, Z) can be generated. When generating the point cloud, information on the spatial position and shooting direction of the underwater camera CM when the original image group {Pic(t)|t=1, 2, ..., N} was acquired by the measurement unit 1 at the corresponding time t may also be used.

[0026] The processing of the point cloud generation unit 6 may be performed in advance outside the analysis unit 2, so that the analysis unit 2 does not receive the underwater image as input, but instead reads a point cloud of underwater objects corresponding to the underwater image into the classification unit 7 and the bottom separation unit 8. For example, the measurement unit 1 may immediately perform processing equivalent to that of the point cloud generation unit 6 on the image obtained by the underwater camera CM using dedicated hardware, etc., so that a point cloud is obtained immediately at the time of underwater photography. The hardware constituting the measurement unit 1 may be underwater LiDAR (light detection and ranging), so that a point cloud of underwater objects is obtained immediately.

[0027] The classification unit 7 classifies the point cloud (referred to as PG) obtained by the point cloud generation unit 6 into multiple point clouds pg1, pg2, ... (or, in some embodiments, the point clouds pg1, pg2, ... obtained as a polyline or a single line segment) that are estimated to correspond to individual aquatic plants as a single line or group, and outputs the result to the volume calculation unit 9. Note that, since the point cloud PG may generally contain noise, some points may be excluded as a result of classification by the classification unit 7 as they do not correspond to individual aquatic plants.

[0028] The water bottom separation unit 7 separates points determined to correspond to the water bottom from the point cloud PG obtained by the point cloud generation unit 6 and generates a separation result (PG [水底] The separation result PG obtained by the bottom separation unit 7 is output to the volume calculation unit 9. The details will be described later. [水底] By outputting the above to the classification unit 7, the classification unit 7 does not perform classification processing on the entire point cloud PG, but instead performs classification processing on only those points in the point cloud PG that are determined to correspond to the bottom of the water. [水底] "PG\PG" [水底] (where the backslash "\" represents the set difference) may be used for classification processing.

[0029] 4, the classification unit 7 includes a polyline estimation unit 71, an upper cross section estimation unit 72, a point cloud distance estimation unit 73, and a cluster estimation unit 74, and each unit performs its own estimation process to obtain classification results pg1, pg2, ... of the point cloud PG based on the estimation results. That is, each of the units 71 to 74 is responsible for a separate embodiment of the classification process by the classification unit 7. The process of only one of the units 71 to 74 may be applied to the entire point cloud PG, or which of the units 71 to 74 is applied may vary for each partial data of the point cloud PG.

[0030] The polyline estimation unit 71 obtains the classification result through the following steps 71a, 71b, and 71c. Fig. 5 is a diagram schematically showing the processing contents of the polyline estimation unit 71, and examples EX51, EX52, and EX53 show examples of steps 71a, 71b, and 71c, respectively.

[0031] <Step 71a> The point cloud PG is clustered (possibly using the clustering technique of the cluster estimation unit 74 described later) to obtain multiple clusters CL1, CL2, ..., and then polylines (connecting multiple straight lines) are generated from each cluster CL1, CL2, ... to obtain polylines PL1, PL2, .... The polyline generation process may involve manual input from a data analyst (user) for each cluster CL1, CL2, ..., so that the individual line segments that make up the polyline are specified. Alternatively, polylines may be generated automatically without human intervention, by generating a single line segment (as a special case of multiple line segments) connecting the points with the largest and smallest Z-axis (vertical) coordinates among the points in each cluster CL1, CL2, .... In either the manual or automatic process, the Z-axis (vertical) coordinate of the point closest to the bottom of the polyline may be manually or automatically corrected using information about the depth of the bottom of the water obtained by the water bottom separation unit 8 described later, so that it matches the value of the bottom of the water.

[0032] <Step 71b> From among these multiple polylines, those determined to be noise that does not correspond to aquatic plants such as seaweed are excluded. Polylines that are excluded can be those whose length (which can be the sum of the lengths of the constituent line segments or the distance between the two end points) is determined to be short based on a threshold ("Reason 3" in example EX52 of Figure 5), those whose orientation (which can be the orientation of the line segments connecting the two ends of the polyline) is determined to be nearly horizontal based on a threshold ("Reason 2" in the same example), or those whose position overlaps an area lower than the water bottom ("Reason 1" in the same example). The method for acquiring water bottom information will be explained in the section on the water bottom separation unit 8, which will be described later.

[0033] <Step 71c> For each of the polylines that have not been excluded and remain as aquatic plants, a predetermined three-dimensional volume (information on the occupied range, for example, a cylinder or a rectangular prism) is assigned according to the type K(i,j) of aquatic plants obtained by the identification unit 3 in the corresponding section Dist(i,j).

[0034] The three-dimensional volume information can be used in an embodiment of the volume calculation unit 9, which will be described later, and the details thereof will be described later when the volume calculation unit 9 is explained.

[0035] As described above, the polyline estimation unit 71 clusters the point cloud PG and divides it into individual clusters, and then converts each cluster into a polyline, thereby detecting aquatic plants that have a shape that extends roughly vertically from the bottom of the water toward the water surface as the polyline model, and also uses a predetermined shape model corresponding to the corresponding type of aquatic plant K(i,j) when calculating the volume.

[0036] On the other hand, the upper cross section estimation unit 72 is an estimation process that is suitable for cases where polylining is not suitable, such as when individuals or clusters can be identified one by one at the top near the water surface, but the roots at the bottom at the bottom cannot be seen because they blend in with other individuals, and estimates the length of each individual from the cross section shape of the upper part of aquatic plants such as seaweed.

[0037] 6 is a diagram showing schematic examples EX61 to EX63 for explaining the processing of the upper cross section estimation unit 72. For the sake of explanation below, with respect to the Cartesian coordinates XYZ of each three-dimensional point (X, Y, Z) belonging to the point group PG, it is assumed that the Z-axis direction represents the vertical direction, i.e., the water depth direction, and the X-axis and Y-axis represent the horizontal direction.

[0038] The upper cross section estimation unit 72 fixes the horizontal position of a rectangular sliding window with a width of X1, a length of Y1, and a height of Z1 as shown in example EX63 of Figure 6 within the range in which the point cloud PG exists, and then slides it in the depth Z direction at each horizontal position to detect areas where multiple individuals are separated on the water surface side but have merged into one entity on the bottom side.

[0039] That is, as shown in Examples EX61 and EX62, by sliding the sliding window of the rectangular parallelepiped in the depth Z direction, for example, three individual regions R1, R2, and R3 exist in a cross section CR(d1) at depth d1, but only one region R4 exists in a cross section CR(d2) at depth d2 on the bottom side of the water below depth d1. When the sliding window is moved in a direction sinking from the water surface side to the bottom side, the number of individual regions decreases (there are multiple regions on the shallow side of the water, and the number decreases from multiple to one on the deep side), the upper cross section estimation unit 72 can determine that a situation has occurred in which multiple individuals are integrated on the base side (referred to as a "multiple individual integrated situation").

[0040] In the upper cross section estimation unit 72, for the location where a multiple-individual integration situation occurs, based on the change in the number of areas at each depth d when the sliding window is moved in the depth direction, the number of areas with the largest number of areas on the shallower side of depth d is obtained as the number of aquatic plants present in the multiple-individual integration situation, and the depth (length of a vertical straight line) from the position of each individual closest to the water surface to the bottom of the water is calculated as the length of each individual.

[0041] In this way, the upper cross section estimation unit 72 identifies the multiple individuals in a location where multiple individuals are integrated, and then detects a single line segment as a model of each individual instead of the polyline. As with the polyline estimation unit 71, the volume of the line segment can be assigned using a predetermined shape model corresponding to the type K(i,j) of the aquatic plant at the location where the line segment exists.

[0042] The number of regions present within the sliding window can be determined by projecting the points present within a rectangular parallelepiped sliding window having a width of X1, a length of Y1, and a height of Z1 onto the XY plane while ignoring the Z coordinates of each point, as shown in Example EX63, and identifying each of the regions R surrounding the points distributed on the XY plane. To identify each of the regions R, clustering similar to that performed by the cluster estimation unit 74, which will be described later, may be used on the XY plane.

[0043] The size and horizontal position of the sliding window, that is, the width X1 and height Y1, can be set so that the entire individual cluster obtained by clustering the point cloud PG is contained within the range of movement when the sliding window is moved in the depth Z direction.

[0044] The polyline estimation unit 71 and upper cross section estimation unit 72 described above identify individual aquatic plants such as seaweed one by one, assuming that the aquatic plants have an elongated shape. The point cloud distance estimation unit 73 and cluster estimation unit 74, which will be described next, estimate individual aquatic plants as clumps like seaweed, regardless of their shape, not limited to when the aquatic plant is composed of a single elongated individual, such as when the aquatic plant is present in a spherical clump.

[0045] 7 is a diagram showing a schematic example of the processing of the point cloud distance estimation unit 73, and is suitable for aquatic plants such as seaweed that cannot be identified individually as a group of seaweed, such as Sargassum, as shown in Example EX71. In this case, as shown in Example EX72, for example, four mutually adjacent point clouds pga, pgb, pgc, and pgd are obtained, and the point cloud distance estimation unit 73 groups these mutually adjacent point clouds into one as shown in Example EX73, and then identifies the interior of the group connecting the outer point clouds as a single individual.

[0046] That is, the point cloud distance estimation unit 73 performs clustering on the entire point cloud PG in the same manner as the cluster estimation unit 74 described later to obtain cluster point groups pg1, pg2, pg3, ..., and groups together those whose distance between their centers of gravity is determined to be below a threshold, and then identifies the grouped point groups as a single individual, as in example EX73.

[0047] The cluster estimation unit 74 applies DBSCAN (Density-Based Spatial Clustering of Applications with Noise), which is a density-based clustering method as an existing method, to the point cloud PG, and obtains each cluster as a classification result for each individual.

[0048] The DBSCAN procedure is as follows: after setting parameters in the pre-configuration, steps 1 to 3 are repeated for all points P belonging to the point cloud PG using those parameters, and the clustering results are obtained by determining which cluster each point P belongs to, or does not belong to any cluster. As is well known, density-based clustering is performed by defining a sphere of radius eps in step 2, and determining whether the number of points inside this sphere of a specified volume is dense or not in step 3 based on whether the number of points inside is equal to or greater than a specified minPts. <Preliminary settings> Determine the values of eps and minPts as clustering parameters. <Step 1> Calculate the distance from a point P to all other points. <Step 2> Points that exist within a distance eps are considered to be adjacent points to P. <Step 3> If there are more than minPts points, the point is in the same cluster as point P.

[0049] The above-described units 71, 72, and 73 are equivalent to units that perform further processing on the results of DBSCAN in the cluster estimation unit 74. DBSCAN itself is a general-purpose clustering method that takes density into consideration, but the units 71, 72, and 73 also simplify modeling of point clouds, which generally have complex distributions, making it possible to simply calculate the volume of aquatic plants such as seaweed. (Note that when the upper cross-section estimation unit 72 moves the sliding window in the depth direction, DBSCAN is applied to the XY plane at each depth to identify each of the regions R1, R2, R3, R4, etc.)

[0050] The water bottom separation unit 7 separates points determined to correspond to the water bottom from the point cloud PG obtained by the point cloud generation unit 6 according to any one of the following embodiments (a) to (c). [水底] ) can be obtained as

[0051] (a) Among the clusters resulting from applying DBSCAN described in the cluster estimation unit 74 to the point cloud PG, if the range in the horizontal direction (XY plane direction) (the range may be evaluated by the area of the circumscribing rectangle of the cluster on the XY plane, or may be evaluated by the area of the polygon connecting the outermost periphery of the cluster) is determined to be wide by threshold judgment, one or more clusters that are determined to span a wide planar range are determined to be clusters on a surface along the bottom of the water, thereby PG [水底] You may try to obtain the following.

[0052] (b) A ruler is placed on the bottom of the water in advance, and can be used to obtain length information in the generated point cloud PG. The depth of the position where the ruler is detected by object detection or the like in the underwater image obtained in step S1 is set as the bottom of the water (Z=0), and the point cloud obtained on the deeper side (Z<0) is set as PG. [水底] The range of Z<0 may be below the water bottom, where aquatic plants such as seaweed cannot exist. Note that this embodiment b is based on the premise (prior knowledge) that the depth of the water bottom is constant within the range of the underwater image obtained in step S1.

[0053] FIG. 8 is a diagram showing the above-described embodiments a and b of the bottom separating section 7 as an example EX81.

[0054] (c) As an additional process when obtaining underwater images in step S1, the water depth Z = Z(X,Y) at each horizontal position (X,Y) is also measured using a depth meter installed on the water drone D, and information on the bottom topography is obtained as a given. The point cloud obtained at a position deeper than the water depth Z = Z(X,Y) is used as the PG [水底] 9 is a diagram showing the embodiment c of the bottom separating unit 7 as an example EX82.

[0055] As also shown in Figure 9, when the polyline estimation unit 71 and the upper cross section estimation unit 72 obtain information on the length of the elongated distribution of aquatic plants such as seaweed as a polyline or a single line segment, the deepest point (X1, Y1, Z1) on the polyline or single line segment can be further extended vertically toward the bottom in the Z-axis direction, and the root point (X1, Y1, Z2) can be located at a water depth Z2 = Z(X1, Y1) ≦ Z1 in embodiment c.

[0056] The volume calculation unit 9 calculates the volume of each individual aquatic plant from the point cloud of each individual aquatic plant obtained by the classification unit 7 or a model of the point cloud as a polyline or a single line segment, and outputs the volume to the display unit 5.

[0057] When information on the area occupied by an individual aquatic plant is obtained in the classification unit 7 as a polyline by the polyline estimation unit 71 or as a single line segment by the upper cross section estimation unit 72 (hereinafter, the term "line segment information" is used to collectively refer to both the polyline and the single line segment), the line segment information and a predetermined three-dimensional model corresponding to the type of aquatic plant K(i,j) in the area Dist(i,j) in which the line segment information is located can be assumed to be arranged around the line segment information, and the volume of the aquatic plant can be calculated as the volume of this predetermined three-dimensional model.

[0058] FIG. 10 is a diagram showing a schematic diagram of the volume calculated by the volume calculation unit 9 for line segment information using a cylindrical model as an example of a predetermined three-dimensional model corresponding to the type K(i, j) of aquatic plant. For a polyline consisting of five line segments d1 to d5 as shown in Example EX01, the area occupied by the aquatic plant may be three-dimensionally modeled using cylinders V1 to V5, each of which has a central axis corresponding to each of the line segments d1 to d5 and a base circle with a radius corresponding to the length of each of the line segments d1 to d5, as shown in Example EX02, and the total volume of these cylinders V1 to V5 (if there are overlapping portions between the cylinders, this may be excluded) may be calculated as the volume.

[0059] Alternatively, as a simpler three-dimensional model, as shown in example EX03, the volume of a single cylinder V6 determined according to the depth distance between the topmost position (water surface side) and the bottommost position (bottom side) of a polyline consisting of five line segments d1 to d5 may be calculated by referring to these positions.

[0060] 10, a cylindrical model is used as an example of a predetermined 3D model corresponding to the type K(i,j) of aquatic plants, but a rectangular parallelepiped model or other models with more complex shapes may also be used. Any predetermined model that can be determined based on rules from line segment information and type K(i,j) can be prepared in advance, and its volume can be calculated.

[0061] On the other hand, if the classification unit 7 obtains information on the area occupied by individual aquatic plants as partial point clouds of the entire point cloud PG using the point cloud distance estimation unit 73 or the cluster estimation unit 74, the volume calculation unit 9 divides the individual point clouds pga into equal widths in the depth Z direction, as shown schematically in Figure 11, and calculates the volume as the sum of the product d*S of the interval d and the area S of each cross section.

[0062] In the example of Figure 11, as shown in example EX11, an individual point cloud pga is divided into six point cloud parts pga1 to pga6 by dividing it into sections of length d in the depth Z direction, and if the cross-sectional areas S1 to S6 represent the horizontal occupation range of each point cloud part, the volume V of the aquatic plant corresponding to the point cloud pga can be estimated as the sum V obtained by multiplying the length d by the cross-sectional areas S1 to S6. V=d*(S1+S2+S3+S4+S5+S6)

[0063] As shown in example EX13, the cross-sectional areas S1 to S6 of each point cloud portion pga1 to pga6 can be calculated by projecting the points contained in each point cloud portion pga1 to pga6 in the Z-axis direction (projecting onto the XY plane in the same manner as example EX63 in Figure 6), finding the outermost periphery of the points distributed on the projected XY plane, and then calculating each cross-sectional area S1 to S6 as the range enclosed by the outermost periphery.

[0064] 11 does not use information on the type K(i,j) of the aquatic plant in the corresponding area Dist(i,j) when calculating the volume V, but an effective volume calculation coefficient r(i,j) corresponding to the type K(i,j) may be set in advance, and the geometrically calculated volume V may be multiplied by the coefficient r(i,j) to obtain V*r(i,j), which is the volume of the corresponding aquatic plant. The coefficient r(i,j) may be a positive value that may be less than 1 or greater than 1.

[0065] As is clear from the above explanation, the volume calculated by the volume calculation unit 9 does not have to be modeled to match the actual volume of the aquatic plant as a physical object, but may be a model of the area in water occupied by the aquatic plant. That is, it is assumed that the actual volume of the aquatic plant as a physical object, for example, when seaweed is brought out of water, dried, and then compressed by applying force, becomes extremely small compared to the volume of the area it initially occupied in water. In this embodiment, however, a modeled volume may be calculated as the area in water occupied by the aquatic plant that receives sunlight and is capable of photosynthesis, as basic data for calculating the CO2 absorption amount.

[0066] As described above, according to the embodiment of the present invention, the volume of aquatic plants can be calculated relatively simply and with a certain degree of accuracy as basic data for calculating the amount of CO2 absorption, using a point cloud obtained by measuring underwater objects, which can be obtained relatively easily.

[0067] Various supplementary examples, additional examples, alternative examples, etc. will be described below.

[0068] (1) The analysis device 2 (analysis unit 2) of the present invention can contribute to the enrichment of basic data for accurately estimating the amount of blue carbon by calculating the volume taking into account three-dimensional distribution. Therefore, by establishing the basis for appropriate credit trading, it will be possible to contribute to Goal 13 of the United Nations-led Sustainable Development Goals (SDGs), which is to “take urgent action to combat climate change and its impacts.”

[0069] (2) FIG. 12 is a diagram showing an example of the hardware configuration of a general computer device 80. All or any part of the analysis device 2 (analysis unit 2), the identification unit 3, the absorption amount DB 4, and the display unit 5 in the analysis system 100 can be realized as one or more computer devices 80 having such a configuration. When each unit (all or part) is realized by two or more computer devices 80, information required for processing may be transmitted and received via a network. The computer device 80 includes a CPU (central processing unit) 81 that executes predetermined instructions, a GPU (graphics processing unit) 82 as a dedicated processor that executes some or all of the CPU 81's execution instructions in place of or in cooperation with the CPU 81, a RAM 83 as a main storage device that provides a work area for the CPU 81 (and the GPU 82), a ROM 84 as an auxiliary storage device, a communication interface 85, a display 86 that displays and outputs images, an input interface 87 that accepts user input via a mouse, keyboard, touch panel, or the like, a speaker 88 that outputs audio, and a bus BS for transmitting and receiving data among these.

[0070] Each of the units 2, 3, 4, and 5 can be realized by a CPU 81 and / or a GPU 82 that reads from a ROM 84 and executes a predetermined program corresponding to the function of each unit. Both the CPU 81 and the GPU 82 are a type of computing device (processor). Here, when display-related processing is performed, a display 86 also operates in conjunction with the unit, when communication-related processing related to data transmission and reception is performed, a communication interface 85 also operates in conjunction with the unit, and when audio output-related processing is performed, a speaker 88 also operates in conjunction with the unit. [Explanation of symbols]

[0071] 100...analysis system, 1...measuring section, 2...analysis section (analysis device), 3...identification section, 4...absorption amount DB, 5...display section 6...Point cloud generation unit, 7...Classification unit, 8...Water bottom separation unit, 9...Volume calculation unit 71... Polyline estimation unit, 72... Upper cross section estimation unit, 73... Point cloud distance estimation unit, 74... Cluster estimation unit

Claims

1. a classification unit that applies classification to the point clouds measured for the underwater objects, and obtains sub-point clouds corresponding to the individual aquatic plants as a single object occupying a continuous range; and a volume calculation unit that calculates, for each of the partial point groups, a volume of an underwater area occupied by a corresponding aquatic plant.

2. The analysis device according to claim 1 , wherein the classification unit obtains each of the partial point clouds by performing clustering based on the density of points in the point cloud.

3. the classification unit further receives assignment of polylines to partial point clouds resulting from clustering based on the density of points in the point cloud; 3. The analysis device according to claim 2, wherein the volume calculation unit calculates the volume of the underwater area as the volume of a predetermined three-dimensional model corresponding to the assigned polyline.

4. The analysis device according to claim 2, characterized in that the classification unit further determines that, among the assigned polylines, those determined to be short or whose orientation is determined to be close to horizontal do not correspond to the individual.

5. The classification unit further slides a three-dimensional window of a predetermined size in the water depth direction for the partial point cloud resulting from clustering based on the density of points in the point cloud, and if the number of individual regions of the point cloud appearing in the cross section that is slid is multiple on the shallow side of the water depth and decreases from multiple to one on the deep side of the water depth, The analysis device described in claim 2, characterized in that the volume calculation unit calculates multiple volumes by assuming that each of the multiple individual areas on the shallow side of the water has an underwater range occupied by a corresponding aquatic plant as an individual.

6. If there are multiple clusters determined to be close to each other among the clusters resulting from clustering by the classification unit, The analysis device according to claim 2 , wherein the volume calculation unit calculates the volume of the underwater area by combining the plurality of clusters into one.

7. a water bottom separation unit for separating a water bottom point cloud corresponding to the water bottom from the point cloud; The analysis device according to claim 1 , wherein the classification unit applies the classification to the point cloud excluding the bottom point cloud from the point cloud.

8. The analysis device according to claim 1, characterized in that the volume calculation unit calculates the volume of the underwater range using the area of the horizontal occupied range determined for each range obtained by dividing the partial point cloud in the water depth direction.

9. 9. A program that causes a computer to function as the analysis device according to claim 1.

Images