JP7210009B2 - Deforestation survey data creation method and deforestation survey data creation device - Google Patents

Description

TECHNICAL FIELD The present invention relates to a felling situation investigation data creation method and a felling situation investigation data creation device for creating felling situation investigation data when trees are felled in a predetermined forest area.

The felling of trees in forests is an important work that affects forestry income. Gold is used to cover logging costs to keep forest owners and timber producers out of the red. On the other hand, from the perspective of appropriate execution of subsidies, forest managers in prefectures and municipalities and timber producers who performed logging work need to conduct surveys after logging to determine whether logging has been carried out as planned (logging status survey). However, at present, the logging situation survey is conducted manually.

Since such surveys of logging conditions are concentrated in winter at the end of the fiscal year when logging is completed, logging surveys are often conducted under snow in cold regions. For this reason, the ordering party, the forestry business entity, and the investigator should prepare for the logging survey such as snow removal, conduct the logging survey at the site, and prepare the logging survey materials. It requires a lot of work and a lot of effort.

In addition, in order to prepare the logging situation survey materials, it is necessary to conduct an on-site survey to check whether logging is being carried out evenly throughout the survey area without any concentration or unevenness, and to check whether the trees that have been felled (also called felled trees). It is necessary to calculate resource information related to felled trees, such as the number of trees, average breast height, average tree height, and volume. For this reason, the preparation of deforestation survey data requires a large amount of manpower and time, as well as a wealth of knowledge and experience.

In recent years, technology has been developed to create forest resource information on survey target forest areas based on laser measurement data obtained by irradiating laser light onto an area that includes a predetermined forest area (also called survey target forest area) from the sky. It has been proposed (see Patent Document 1, for example).

The forest resource information calculation method described in Patent Document 1 is based on three-dimensional laser point cloud data, and forest resource information such as crown diameter, breast height diameter, tree height, and timber volume obtained from breast height diameter and tree height of each tree. This is a forest resource information calculation method that enables highly accurate calculations.

JP 2018-84472 A

The forest resource calculation method described in Patent Document 1, as described above, collects forest resource information such as crown diameter, breast height diameter, tree height of each tree, and timber volume obtained from these tree heights and breast height diameters. It is possible to calculate with accuracy.

However, in the forest resource calculation method described in Patent Document 1, it is not possible to create data for investigating the state of felling when felling is performed in a predetermined forest area (data for investigating the state of felling). The logging survey data includes data that indicates whether logging is being done evenly over the entire area, without any concentration or bias, and resource information on felled trees (number of felled trees, average breast height, average tree height, timber volume, etc.). etc.) are included.

SUMMARY OF THE INVENTION Accordingly, the present invention has been made to solve the above problems. It is an object of the present invention to provide a harvesting survey data creation method and a harvesting survey data creation device capable of creating data without requiring

[1] A felling situation investigation data creation method of the present invention is a felling situation investigation data creation method for creating data for investigating the felling situation when trees are felled in a predetermined forest area, A pre-felling crown height image data acquisition processing step of acquiring pre-felling tree crown height image data in the predetermined forest area, and a post-felling tree crown height image data acquiring step of acquiring post-felling tree crown height image data in the predetermined forest area. A crown height in a crown region composed of the crowns of one or more felled trees is represented by a data acquisition processing step and a difference between the tree crown height image data before felling and the tree crown height image data after felling. a crown height image data difference processing step of creating tree canopy height difference image data, and removing noise existing in the tree canopy height difference image data created by the crown height image data difference processing, thereby removing noise. A noise removal processing step of creating noise-removed tree crown high-difference image data, and based on the noise-removed tree crown high-difference image data created by the noise removal processing, from the crown of the one or more felled trees a felled tree crown segmentation processing step of creating felled tree crown segmentation image data in which the tree crown region is segmented for each individual felled tree crown; and a felled tree resource information acquisition processing step of acquiring resource information about each of the felled trees.

The tree crown height image data of a predetermined forest area before felling and the tree crown height image data of the predetermined forest area after felling are created by obtaining the difference between the tree crown height image data before felling and the tree crown height image data after felling in the predetermined forest area. Noise present in the high-difference image data is removed to create noise-removed tree canopy high-difference image data. Then, based on the noise-removed tree crown high-difference image data, creating cut tree crown segmented image data in which a crown region composed of the crown of one or more cut trees is segmented for each crown of the cut tree, Based on the felled tree canopy segment image data, resource information about each felled tree is acquired.

For this reason, according to the felling situation investigation data creation method of the present invention, the felling situation investigation data for investigating the felling situation when trees are felled in a predetermined forest area can be generated using a large amount of manpower and time. can be created without needing As mentioned above, the logging status survey data includes data that indicates whether logging is being done evenly throughout the entire area without any concentration or unevenness, and resource information on felled trees (cutting trees). number of trees, average breast height, average tree height and volume, etc.).

Further, according to the felling situation survey data creation method of the present invention, the noise removal processing step of removing noise present in the tree crown height difference image data created by the tree crown height image data difference processing enables the tree crown height image The tree crown high-difference image data created by the data subtraction processing becomes the noise-removed tree crown high-difference image data, and using the noise-removed tree crown high-difference image data, a tree canopy consisting of the crown of one or more felled trees Cut-down tree canopy segmented image data is created by dividing the area into individual felled tree canopies. Therefore, the cut tree canopy segmented image data is a highly accurate reproduction of the canopy region of each cut tree. Therefore, by obtaining the resource information about each felled tree based on the felled tree crown segment image data that reproduces the crown area of each felled tree with high accuracy, the obtained resource information about the felled tree can be It will represent the resources of felled trees with high accuracy.

[2] In the felling situation survey data creation method of the present invention, the noise removal processing step includes extracting the maximum tree crown height from the tree crown height difference image data created by the tree crown height image data difference processing step. A crown height extraction processing step, and setting a threshold value of the crown height for making the crown candidate of the felled tree based on the maximum crown height, setting the threshold as the first threshold, and determining the first threshold in the crown height difference image data a felled tree crown candidate area extraction step of extracting an area having a crown height equal to or greater than a threshold as a felled tree crown candidate area and creating "crown height difference image data for which the felled tree crown candidate area has been extracted"; A threshold is set for the area of the tree crown candidate area, the set threshold is used as a second threshold, and the felled tree crown candidate area having a small area of which the area of the felled tree crown candidate area is less than the second threshold is defined as the above. A small-area felled tree crown candidate region removal processing step of removing from the "tree crown high-difference image data from which the felled tree crown candidate region has been extracted" to create "tree crown high-difference image data from which the small-area felled tree crown candidate region has been removed" Then, a threshold is set to a value obtained by dividing the longitudinal length of the felled tree crown candidate region by the transverse length of the felled tree crown candidate region, and the set threshold is used as a third threshold. and removes a thin felled tree crown candidate region having a felled tree crown candidate region equal to or greater than the third threshold value from the above-mentioned "crown height difference image data after removal of the small area felled tree crown candidate region" to obtain a "slender felled tree crown a slender tree crown candidate area removal processing step of creating "tree crown high difference image data after candidate area removal"; and a noise removal processing step in the felled tree canopy candidate area for creating "the tree canopy high difference image data in the felled tree canopy candidate area after removing the noise from the already removed tree canopy high difference image data". is preferred.

The noise removal processing step includes a maximum tree crown height extraction processing step, a felled tree crown candidate region extraction processing step, a small area felled tree crown candidate region removal processing step, a thin felled tree crown candidate region removal processing step, and noise in the felled tree crown candidate region. By sequentially performing the removal processing steps, it is possible to reliably remove various noises present in the canopy height difference image data created by the canopy height image data difference processing.

Here, the first threshold can be set as follows, for example. In other words, when the trees to be felled are arranged in order of height from the top, middle, and lower layers, the vertical height of the tree crown of these upper, middle, and lower layer trees (from the bottom of the crown) The height to the top of the tree crown) is about 1/2 of the tree height, and the crown height of middle and lower trees is about 1/2 of the maximum crown height (maxH) of the upper tree. Since the values are obtained, the first threshold value is set so that not only the upper layer tree but also the middle layer tree and the lower layer tree are included in the felled trees. For example, 1/5 of the maximum crown height (maxH) is set as the first threshold, and crown height image data having a crown height of 1/5 or more of the maximum crown height (maxH) is extracted from the crown height difference image data. and "the tree crown height difference image data for which the tree crown candidate region of the felled tree has been extracted". As a result, understory vegetation such as bushes below 1/5 of the maximum crown height (maxH) is removed as noise.

Also, the second threshold can be set as follows, for example. That is, since it is empirically obtained that the crown area of the understory tree to be felled is approximately ^10m2 , a smaller area (for example, 8m2) is set as the ^second threshold. . Then, the small-area felled tree crown candidate region having an area smaller than the second threshold is removed from the "tree crown high-difference image data for which the felled tree crown candidate region has been extracted." As a result, a felled tree canopy candidate region having an area smaller than that of the understory tree (for example, 8 m ² ) can be removed as noise.

Also, the third threshold can be set as follows, for example. That is, the size of the value obtained by dividing the length in the longitudinal direction (Lmax) of the crown of the tree to be felled in plan view by the length in the lateral direction (Lmin) Since it is empirically obtained that the height does not exceed 4, "4" is set as the third threshold. For example, if the third threshold is n, set "n = 4", and the thin felled tree crown candidate region that is equal to or greater than the third threshold (n = 4) is defined as a "small area felled tree crown candidate area-removed tree canopy height difference image data”. As a result, extremely long and narrow felled tree crown candidate regions can be removed as noise.

[3] In the felling situation survey data creation method of the present invention, in the felled tree resource information acquisition processing step, at least , acquisition of location information of each felled tree, crown area and/or crown diameter of each felled tree, height of each felled tree, breast-height diameter of each felled tree, and volume of each felled tree. preferably.

By acquiring such felled tree resource information, it is possible to easily create materials for investigating the state of felling when trees are felled in a predetermined forest area to be felled. Further, by displaying the felled tree resource information as a list on the display, detailed resource information of the felled trees in the predetermined forest area can be visually grasped.

[4] In the felled tree resource information creation method of the present invention, the felled tree resource amount is aggregated based on the resource information about the individual felled tree acquired in the felled tree resource information acquisition processing step. It is preferable to further include an amount aggregation processing step.

By performing such tally processing of the amount of felled tree resources, it is possible to create a tally table of the amount of felled tree resources. It is possible to easily create materials for investigating the logging situation when performing. In addition, by displaying the felled tree resource information as a list on the display, it is possible to visually grasp the outline of the felling situation.

[5] In the felling status survey data creation method of the present invention, the felled tree resource amount aggregation processing step includes: the number of felled trees, the average diameter at breast height of the felled trees, the average tree height of the felled trees, and the It is preferable that the method includes a process of counting the total volume of felled trees and calculating the cutting rate based on the number of trees and the cutting rate based on the volume of trees.

As a result, it is possible to easily grasp the state of felling carried out in a predetermined forest area. In addition, based on such a summary table of the amount of felled tree resources, it is possible to determine whether the number of trees felling rate and the timber pile felling rate, which are the basis for calculating the subsidy expenditure in the felling survey work, are at the predetermined felling rate. can be easily determined.

[6] In the felling status survey data creation method of the present invention, the crown area and/or crown diameter of each of the felled trees is obtained based on the number of pixels in the crown region of each of the felled trees. , the tree height of each of the felled trees is determined as the tree height of each of the felled trees by the maximum value of crown height image data corresponding to the crown of each of the felled trees classified by the above-mentioned felled tree crown classification processing; The diameter at breast height of a felled tree is obtained by a multiple regression formula from the crown area or crown diameter of each felled tree and the tree height of each felled tree. It is preferably determined based on the diameter and height of individual felled trees.

As a result, the crown area and/or crown diameter, tree height, breast height diameter and volume of the felled tree (cut tree) can be obtained even after the tree has been felled.

[7] In the felling situation survey data creation method of the present invention, the pre-felling tree canopy height image data obtained by the pre-felling tree crown height image data obtaining processing step corresponds to the pre-felling forest area. Meshed digital surface model data and digital elevation based on predetermined forest area image data before logging created based on three-dimensional point cloud data and geographical information corresponding to the predetermined forest area model data is created, and the difference between the digital surface layer model data and the digital elevation model data is taken for each mesh, and the deforestation acquired by the tree crown height image data acquisition processing step after deforestation The later canopy height image data is created based on the three-dimensional point cloud data corresponding to the predetermined forest area after logging and the geographical information corresponding to the predetermined forest area. It is created by creating meshed digital surface model data and digital elevation model data based on the area image data, and taking the difference between the digital surface model data and the digital elevation model data for each mesh. preferably.

As a result, the tree crown height image data before felling and the tree crown height image data after felling can be created with high accuracy for each mesh. In this case, the digital surface layer model data refers to data representing the altitude of the surface layer of trees and the like, and the digital altitude model data refers to data representing the altitude of the ground surface. Therefore, by taking the difference between the digital surface layer model data and the digital elevation model data, the tree crown height (the height of the tree crown from the ground surface) can be obtained for each mesh.

[8] In the felling situation survey data creation method of the present invention, the cut tree crown segmentation process uses an individual tree detection method to individually classify the crown regions of the cut tree. It is preferable to create felled tree crown segmented image data that is segmented for each tree crown of the felled tree.

The Individual Tree Detection Method uses the fact that the height of the crown at the edge of the crown is lower than the height of the crown at the top of the crown to classify the crown of the felled tree. By using the segmentation processing method, the tree canopy corresponding to each tree of the felled trees can be extracted with high accuracy.

[9] A felling situation investigation data creation device of the present invention is a felling situation investigation data creation device that creates data for investigating the felling situation when trees are felled in a predetermined forest area, A pre-felling crown height image data acquisition processing unit for acquiring pre-felling tree crown height image data in the predetermined forest area, and a post-felling tree crown height image data for acquiring post-felling tree crown height image data in the predetermined forest area. The crown height in the crown region composed of the crowns of one or more felled trees is represented by the data acquisition processing unit and the difference between the tree crown height image data before felling and the tree crown height image data after felling. and a crown height image data difference processing unit for creating tree crown height difference image data, and removing noise existing in the tree crown height difference image data created by the crown height image data difference processing. a noise removal processing unit that creates noise-removed tree crown high-difference image data; a felled tree crown segmentation processing unit that creates felled tree crown segmentation image data in which the tree crown region is segmented for each tree crown of each felled tree; and a felled tree resource information acquisition processing unit that acquires resource information about each of the felled trees.

The logging situation survey data creation device of the present invention is a device for performing each processing step of the logging situation survey data creation method described in [1] above, and the logging situation survey data creation device of the present invention is By having such a configuration, the same effect as the felling situation survey data creation method described in [1] above can be obtained.

[10] In the deforestation survey data creation device of the present invention, the noise removal processing unit includes a maximum height for extracting the maximum tree crown height from the tree crown height difference image data created by the tree crown height image data difference processing unit. A crown height extraction processing unit and a crown height threshold for setting a crown candidate of the felled tree based on the maximum crown height, the threshold being set as a first threshold, and the first threshold in the crown height difference image data a felled tree crown candidate area extraction processing unit that extracts an area having a crown height equal to or greater than a threshold as a felled tree crown candidate area and creates "crown height difference image data for which the felled tree crown candidate area has been extracted"; A threshold is set for the area of the tree crown candidate area, the set threshold is used as a second threshold, and the felled tree crown candidate area having a small area of which the area of the felled tree crown candidate area is less than the second threshold is defined as the above. A small-area felled tree crown candidate region removal processing unit that removes from the "tree crown high-difference image data after extraction of the felled tree crown candidate region" and creates "crown high-difference image data after the removal of the small-area felled tree crown candidate region" Then, a threshold is set to a value obtained by dividing the longitudinal length of the felled tree crown candidate region by the transverse length of the felled tree crown candidate region, and the set threshold is used as a third threshold. and removes a thin felled tree crown candidate region having a felled tree crown candidate region equal to or greater than the third threshold value from the above-mentioned "crown height difference image data after removal of the small area felled tree crown candidate region" to obtain a "slender felled tree crown a slender tree crown candidate area removal processing unit that creates "candidate area removed tree crown high difference image data"; and a noise removal processing unit in the felled tree canopy candidate area that removes the noise from the tree crown high difference image data that has been removed from the tree crown candidate area and creates "the tree canopy high difference image data that has undergone noise removal in the felled tree canopy candidate area." is preferred.

Since the noise removal processing unit performs each process included in the noise removal processing step described in [2] above, the same method as the logging situation survey data creation method described in [2] above is performed. effect is obtained.

[11] In the felling situation survey data creation device of the present invention, the cut tree resource information acquisition processing unit performs at least It is possible to perform processing to acquire position information of each felled tree, crown area and/or crown diameter of each felled tree, tree height of each felled tree, breast height diameter of each felled tree, and volume of each felled tree. preferable.

The cut tree resource information acquisition processing unit performs the same processing as the cut tree resource information acquisition processing step in the felling situation survey data creation method described in [3] above. The same effect as the data preparation method for logging situation survey can be obtained.

[12] In the felling situation survey data creation device of the present invention, the felled tree resources are aggregated based on the resource information relating to the individual felled trees acquired by the felled tree resource information acquisition processing unit. It is preferable to further have an amount aggregation processing unit.

Since the felled tree resource quantity aggregation processing unit performs the same processing as the felled tree resource quantity aggregation processing step in the felling situation survey data creation method described in [4] above, the above [4] The same effect as the method for creating data for felling survey described in 1) can be obtained.

[13] In the felling situation survey data creation device of the present invention, the felled tree resource amount tally processing unit includes the number of felled trees, the average diameter at breast height of the felled trees, the average tree height of the felled trees, and the felled trees. It is preferable to calculate the total timber volume of trees and calculate the cutting rate based on the number of trees and the cutting rate based on the timber volume.

The felled tree resource quantity aggregation processing unit performs the same processing as the felled tree resource quantity aggregation processing step in the felling situation survey data creation method described in [5] above. The same effect as the method for creating data for felling survey described in 1) can be obtained.

It is preferable that the felling survey data creation device of the present invention also has the characteristics described in [6], [7], and [8] of the felling survey data creation method of the present invention.

It is a flowchart shown in order to demonstrate the data creation method for the felling situation investigation which concerns on embodiment. Fig. 10 is a flowchart for explaining a procedure for creating tree crown height image data before felling and tree crown height image data after felling; FIG. 2 is a diagram showing an aerial image before and after felling, which is taken from above the area including the forest area to be surveyed. FIG. 4 is a diagram showing a tree crown height image before felling and a tree crown height image after felling displayed on a display; FIG. 10 is a diagram showing a tree canopy height difference image displayed on the display; FIG. 10 is a diagram showing a “crown height difference image for which a felled tree crown candidate region has been extracted” displayed on the display; FIG. 10 is a diagram showing a “crown height difference image from which the small-area felled tree crown candidate region has been removed” displayed on the display; FIG. 10 is a diagram showing a tree crown height difference image from which an elongated felled tree tree crown candidate region has been removed, displayed on the display; FIG. 10 is a diagram showing a “crown height difference image from which noise has been removed in the felled tree crown candidate region” displayed on the display; FIG. 4 is a diagram showing a cut tree canopy segmented image displayed on the display; FIG. 4 is a diagram showing a list of felled tree resource information displayed on the display; FIG. 10 is a diagram showing a felled tree height classification image in which marks representing tree height classifications are attached to the crowns of individual felled trees at the accuracy verification site (within the frame line A2). FIG. 3 is a diagram showing a distribution image of felled trees in a survey target forest area (within the frame line A1). FIG. 10 is a diagram showing a tally table of felled tree resource amounts displayed on the display; BRIEF DESCRIPTION OF THE DRAWINGS It is a figure shown in order to demonstrate the felling condition investigation data creation apparatus 1 corresponding to the felling condition investigation data creation method which concerns on embodiment.

Embodiments of the present invention will be described below.

FIG. 1 is a flow chart for explaining a felling survey data creation method according to an embodiment. A method for creating data for felling survey according to an embodiment will be described with reference to the flowchart shown in FIG. In addition, a predetermined forest area for which data for felling survey is to be created according to the embodiment will be referred to as a survey target forest area. The survey target forest area is, for example, a forest area with tall planted trees such as Japanese larch, Japanese red pine, Japanese cedar, and cypress.

As shown in FIG. 1, the felling situation survey data creation method according to the embodiment includes pre-felling tree crown height image data acquisition processing (step S10) for acquiring pre-felling tree crown height image data in a forest area to be surveyed; Post-felling tree crown height image data acquisition processing (step S20) for acquiring post-felling tree crown height image data in the forest area to be surveyed, and obtaining the difference between the pre-felling tree crown height image data and the post-felling tree crown height image data. By doing so, crown height image data difference processing (step S30) for creating crown height difference image data representing the crown height in a crown region composed of the crown of one or more felled trees, and crown height image data difference processing. Noise removal processing (step S40) for removing noise present in the tree canopy high-difference image data created by and creating denoised tree canopy high-difference image data (step S40); Based on the noise-removed tree canopy high-difference image data, felled tree canopy segmentation image data is created by segmenting the canopy area consisting of the crown of one or more felled trees into individual tree canopies. processing (step S50); felled tree resource information acquisition processing (step S60) for acquiring resource information about individual felled trees based on the felled tree crown segmentation image data created by the felled tree crown segmentation processing (step S60); and a felled tree resource amount tallying process (step S70) for tallying the felled tree resource amount based on the resource information about each felled tree acquired by the resource information acquisition process (step S60).

In addition, in the tree crown height image data difference processing (step S30), when taking the difference between the tree crown height image data before felling and the tree crown height image data after felling, the geographical information obtained from the geographic information system (GIS) 100 Using information (in this case, boundary data of forest areas registered in forest management ledgers maintained by prefectures, etc.), the tree canopy height image data before logging and the canopy height image data after logging are combined. take the difference.

Next, the process of step S40 described above, that is, the noise removal process for creating the noise-removed tree canopy height difference image data will be described. The noise removal process (step S40) includes a maximum tree crown height extraction process (step S41) for extracting the maximum tree crown height from the tree crown height difference image data created by the tree crown height image data difference process (step S30); A threshold value (first threshold value) for tree crown height to be used as a tree crown candidate is set based on the maximum crown height. a tree crown candidate region” and creates “crown high-difference image data for which the felled tree crown candidate region has been extracted” (step S42); (assumed to be the second threshold), and a small area of the felled tree crown candidate area whose area is less than the second threshold is defined as "the extracted tree crown candidate area extracted tree crown height difference small-area felled tree crown candidate area removal processing (step S43) for creating "small-area felled tree crown candidate area removed tree crown high-difference image data" by removing from the image data", and A threshold (third threshold) is set to a value obtained by dividing the directional length by the length of the tree crown candidate area in the transverse direction, and the tree crown candidate area of the felled tree is equal to or greater than the third threshold. is removed from the above-mentioned "crown height difference image data after removal of the small area felled tree crown candidate area" to create "crown height difference image data after removal of the thin felled tree crown candidate area". Elongated tree crown candidate area removal processing (step S44), and noise in the felled tree candidate area existing inside the felled tree crown candidate area is removed from the "crown height difference image data after removal of the elongated tree crown candidate area", and noise removal processing in the felled tree canopy candidate region (step S45) for creating "crown height difference image data in which the noise in the felled tree canopy candidate region has been removed".

Note that the aforementioned “cut tree crown candidate region” is a tree crown region that can become a tree crown of a felled tree, and the cut tree crown candidate region is a single tree crown region formed by the crown of one tree. It also includes a crown area formed by connecting the crowns of multiple trees.

Next, each process (steps S10 to S70) in the flowchart shown in FIG. 1 will be described.

Tree crown height image data before felling acquired in the tree crown height image data acquisition process before felling (step S10) in FIG. 1 and tree crown height image data after felling acquired in the tree crown height image data acquisition process after felling (step S20) is based on pre-felling aerial image data obtained by photographing the area that includes the surveyed forest area from above, and is obtained by irradiating the area that includes the surveyed forest area with a laser beam from above. However, in the logging survey data creation method according to the embodiment, aerial photographs taken from above of the area including the forest area to be surveyed are taken before and after logging. It shall be created based on the image data.

FIG. 2 is a flowchart for explaining a procedure for creating tree crown height image data before felling and tree crown height image data after felling. FIG. 2(a) is a flowchart for explaining the procedure for creating tree crown height image data before felling, and FIG. 2(b) is a flowchart for explaining the procedure for creating tree crown height image data after felling. be. As shown in Fig. 2(a), the procedure for creating pre-felling tree canopy height image data is to create pre-felling aerial image data by taking an aerial photo of an area that includes the forest area to be surveyed before felling. Pre-felling aerial image data creation processing (step S11), and pre-felling three-dimensional point cloud data creation processing (step S12) for creating pre-felling three-dimensional point cloud data from the pre-felling aerial image data. , Pre-felling survey target forest area image data creation processing for creating pre-felling survey target forest area image data based on pre-felling three-dimensional point cloud data and geographical information obtained from the geographic information system (GIS) 100 (Step S13), based on the survey target forest area image data before logging, pre-logging tree crown height image data representing the crown height of trees existing in the survey target forest area is created. and tree crown height image data creation processing (step S14).

The procedure for creating post-felling tree canopy height image data is also basically the same as in FIG. That is, as shown in FIG. 2B, post-felling aerial image data creation processing (step S21) for creating post-felling aerial image data, and post-felling three-dimensional point cloud data creation processing (step S21). 3D point cloud data creation processing (step S22), survey target forest area image data creation processing after logging for creating survey target forest area image data after logging (step S23), and survey target forest area after logging and post-felling tree crown height image data creation processing (step S24) for creating post-felling tree crown height image data representing the crown height of trees existing in the tree.

Here, processing for creating three-dimensional point cloud data before and after felling from aerial image data before and after felling will be further described.

FIG. 3 is a diagram showing an aerial photographed image before and after felling of an area including a forest area to be surveyed taken from above. Fig. 3(a) shows an aerial image of the area including the survey target forest area before logging, taken from above, and Fig. 3(b) shows the area including the survey target forest area after logging. It is a figure which shows the aerial photography image after felling which image|photographed from the sky. The survey target forest area shown in Fig. 3 is a forest area mainly composed of Japanese larch. Fig. 3(a) is an aerial image taken in summer (early July), This is an aerial image taken in early November.

In FIG. 3, the area surrounded by the outer frame line A1 is the survey target forest area, and has an area of about 9.44 ha. In the felling survey data creation method according to the embodiment, the survey target forest area surrounded by the frame line A1 is assumed to be the target forest area for the harvest survey data creation. However, in order to make the logging situation survey data creation method according to the embodiment easier to understand, a specific area within the survey target forest area (the area surrounded by the frame line A1) is defined as an "accuracy verification area". An area surrounded by a line A2 is set, and a detailed explanation may be given in the accuracy verification area (area surrounded by the frame line A2) as necessary. The accuracy verification area (the area surrounded by the frame line A2) has an area of about 0.53 ha.

Since FIG. 3 is a monochrome image, changes in the color of the tree crown due to seasonal changes cannot be read in FIG. . That is, since FIG. 3A is an aerial image taken in early summer, the tree crown of larch is green, and FIG. 3B is an aerial image taken in late autumn, so the larch tree crown is From yellow to brown. Note that since FIG. 3B is an aerial image taken after felling, areas where trees have been felled (areas shown in black) are scattered in the survey target area.

Although an aircraft, a drone, or the like can be used to photograph an area including a survey target forest area from the air, a drone is used in the felling survey data creation method according to the embodiment.

In this way, when the aerial image data before logging and the aerial image data after logging are created by photographing the area including the forest area to be surveyed from above, the aerial image data corresponding to each aerial image is generated. 3D point cloud data before and after felling are created from the tree (steps S12 and S22). In addition, the creation of 3D point cloud data is performed by capturing the area including the survey target area with a camera mounted on a drone, overlapping the shooting locations by about 80%, for example, and shifting the image data. From the photographed image data), a map-corrected orthoimage without distortion is created by SFM (Structure from Motion), which is a known 3D shape restoration technology, and 3D point cloud data is created from the created orthoimage. be able to.

Specifically, several hundred images of the entire area, including the target forest area before and after logging, were taken with a camera mounted on a drone, while overlapping about 80% and shifting. , a distortion-free map-corrected orthoimage is created by SFM. Based on the orthoimages created in this way, it is possible to create three-dimensional point cloud data showing the height before and after felling. The SFM is described in the following publication 1.

Publication 1: Linda G. Shapiro; George C. Stockman (2001). Computer Vision. Prentice Hall.

Image data of the survey target forest area surrounded by the frame line A1 in FIGS. , to create survey target forest area image data before and after felling (steps S13 and S23). That is, the survey target forest area image data before and after logging is obtained from the three-dimensional point cloud data before and after logging using geographical information obtained from the geographic information system (GIS) 100. The image data of the area (surveyed forest area) is extracted and used as the surveyed forest area image data before and after felling.

Based on the survey target forest area image data before and after felling created in this way, tree crown height image data representing the crown height of trees existing in the survey target forest area (crown height before and after cutting) image data) is created (steps S14 and S24).

The pre-cutting and post-cutting tree canopy height image data is created by using the pre-cutting and post-cutting survey target forest area images created by the pre-cutting and post-cutting survey target forest area image data creation processing (step S13 and step S23). Based on the data, crown height image data representing the height of the tree crown (the height of the crown from the ground surface) existing in the survey target forest area before and after felling is created.

Specifically, the tree canopy height image data before felling is combined with meshed digital surface layer model data (data representing the elevation of the surface layer of trees, etc.) based on the survey target forest area image data created before felling. Canopy height image data representing tree crown height in each mesh by creating digital elevation model data (data representing the elevation of the ground surface) and taking the difference between the digital surface model data and the digital elevation model data in each mesh. to create

On the other hand, the tree canopy height image data after logging is based on the image data of the target forest area for survey created after logging. Data (data representing the elevation of the ground surface) is created, and by taking the difference between the digital surface model data and the digital elevation model data for each mesh, crown height image data showing the tree crown height for each mesh is created. do.

Here, the mesh size can be set based on the ground resolution of the survey target forest area image. , Conversely, if the mesh size is too small, the amount of data to be processed increases. For this reason, it is preferable to set the mesh size to an optimum size in consideration of the area of the surveyed forest area, the state of the forest in the surveyed forest area, the accuracy of the calculated forest resource information, etc. For example, 50 cm A mesh size of about 50 cm can be set, and a smaller mesh size can also be set.

Note that the pre-felling tree crown height image data creation process (step S14) and the post-felling tree crown height image data creation process (step S24) are performed for each mesh. , the crown height of each tree can be obtained for each 50 cm x 50 cm mesh.

By storing the pre-cutting tree crown height image data and the post-cutting tree crown height image data created in this manner in an information processing device such as a computer, a tree crown height image data corresponding to the pre-cutting tree crown height image data can be obtained. The post-felling tree crown height image corresponding to the post-felling tree crown height image data and the post-felling tree crown height image data can be displayed on the display at any time.

FIG. 4 is a diagram showing a tree crown height image before felling and a tree crown height image after felling displayed on the display. Note that FIG. 4A is a tree crown height image before cutting, and FIG. 4B is a tree crown height image after cutting. In the frame lines A1 and A2 of FIG. 4, the black portions are portions where no trees exist (portions such as the ground where the crown height is zero). In addition, since FIG. 4(b) is a canopy height image after felling, the parts where trees have been felled (parts shown in black) are scattered more than in FIG. 4(a). .

Subsequently, the felling situation survey data creation process according to the embodiment is performed along the flowchart shown in FIG. 1 described above. That is, pre-felling tree crown height image data acquisition processing (step S10) for acquiring pre-felling tree crown height image data stored in an information processing device such as a computer, and post-felling tree crown height image data acquisition processing for acquiring post-felling tree crown height image data. High image data acquisition processing (step S20) is performed respectively, and processing after crown height image data difference processing (step S30) for obtaining the difference between the tree crown height image data before felling and the tree crown height image data after felling is performed. The processes after the crown height image data difference process (step S30) will be described in detail below.

In the tree crown height image data difference processing (step S30), the tree crown height image data before felling and the tree crown height image data after felling are superimposed using geographical information obtained from the geographic information system (GIS) 100 to create an image. By performing difference processing between trees, crown height difference image data representing the difference between the tree crown height image data before felling and the tree crown height image data after felling is created. The tree crown height difference image data expresses the tree crown height in a tree crown area composed of the tree crown of one or more felled trees. Here, "the crown area consisting of the crowns of one or more felled trees" may be a crown area consisting of the crowns of one felled tree, or the crowns of a plurality of felled trees are connected in a plane. This means that there is also a case of a tree canopy region with a large area.

By storing the canopy height difference image data created by the canopy height difference image data difference processing (step 30) in an information processing device such as a computer, a canopy height difference image corresponding to the canopy height difference image data can be generated at any time. , can be displayed on the display.

FIG. 5 is a diagram showing a tree crown height difference image displayed on the display. FIG. 5(a) shows a crown height difference image of the entire survey target area (the area surrounded by the frame line A1), and FIG. 5(b) shows the accuracy verification area (frame line A2) in FIG. FIG. 10 is a diagram showing an enlarged image in which the enclosed area) and its surroundings are enlarged.

Here, focusing on the accuracy verification area (the area surrounded by the frame line A2), as shown in FIG. are represented in black, and deforested areas (regions with large changes) are represented in white. Note that the black area (hereinafter also referred to as the black area) is an area where there is no change. In some cases. A white area (hereinafter also referred to as a white area) is an area that can be a tree crown candidate for a felled tree (cut tree). Note that the white area may be a crown area consisting of the crown of a single felled tree, or may be a crown area in which the crowns of a plurality of felled trees are connected in a plane.

The canopy height difference image data created by the canopy height image data difference processing (step 30) contains various noises caused by the tree crown height image data difference processing. Various noises caused by differential processing of canopy height image data will be referred to as "noise due to differential processing of canopy height image data". The "noise due to difference processing of tree crown height image data" is removed by noise removal processing (step S40), which will be described later.

In addition, in FIG. 5, gray areas are generated. This is due to the difference in the timing of the photographing, the difference in the angle of the camera when taking the aerial photograph before and after felling, etc. These are also included in "noise due to difference processing of tree canopy height image data". .

In the crown height difference image obtained by taking the difference between the tree crown height image data before felling and the tree crown height image data after felling, the area of the felled tree (the white area in FIG. 5(b)) is , has information on the crown height of the tree that has been felled (the felled tree).

In this way, in the crown height image data difference processing (step S30), difference processing is performed between the crown height image data before felling and the crown height image data after felling to create crown height difference image data. Subsequently, various noises present in the tree canopy height difference image data, that is, "noise due to tree canopy height image data difference processing" are removed by noise removal processing (step S40), and "noise due to tree canopy height image data difference processing" is removed. Create "noise-removed tree canopy height difference image data".

The noise removal process (step S40) is to create "noise-removed tree canopy height difference image data" from which "noise due to tree crown height image data difference processing" has been removed. First, the maximum tree crown height extraction process (step S41) is performed. This maximum crown height extraction process (step S41) extracts the maximum tree crown height (maxH) from the crown height difference image data created by the crown height image data difference process (step S30). It can be said that the maximum crown height (maxH) extracted here is the height of the tallest tree among the trees present in the survey target forest area (within the frame line A1).

Subsequently, a cut tree crown candidate region extraction process (step S42) is performed. In the cut tree crown candidate region extraction processing, a threshold value (first threshold value) of the crown height for making a tree crown candidate of the cut tree is set based on the maximum crown height, and the first threshold value is set in the crown height difference image data. This is a process of extracting an area having the above tree crown height as a "cut tree crown candidate area" and creating "crown height difference image data for which the cut tree crown candidate area has been extracted". Here, the first threshold is set as follows.

In other words, when the trees to be felled are the upper, middle, and lower layer trees in descending order of tree height, the first threshold is the height of the tree crown in the vertical direction of these upper, middle, and lower layer trees. (height from the bottom of the crown to the top of the crown) is approximately 1/2 of the tree height. is obtained as an empirical value, a threshold is set as the first threshold such that not only the upper-layer tree but also the middle-layer tree and the lower-layer tree are included in the felled trees.

Here, 1/5 of the maximum crown height (maxH) is set as the first threshold, and crown height image data having a crown height of 1/5 or more of the maximum crown height (maxH) is extracted from the crown height difference image data. Then, the extracted tree canopy candidate region is extracted as "canopy height difference image data". As a result, understory vegetation such as bushes below 1/5 of the maximum crown height (maxH) is removed as noise. Note that the first threshold is not limited to 1/5 of the maximum crown height (maxH), and an optimum value can be set as appropriate.

By performing such a tree crown height candidate region extraction process, the pixel value of a region less than ⅕ of the maximum tree crown height value (maxH) in the tree crown height difference image data is set to zero. As a result, it is possible to create tree crown height image data in which the felled tree crown candidate region is extracted, that is, "tree crown height difference image data for which the felled tree crown candidate region has been extracted". By storing the "tree crown height difference image data for which the felled tree crown candidate region has been extracted" created in this manner in an information processing device such as a computer, the "crown height difference image data for which the felled tree crown candidate region has been extracted" can be stored. The "canopy height difference image of the felled tree crown candidate region extracted" corresponding to the "image data" can be displayed on the display at any time.

FIG. 6 is a diagram showing a "canopy height difference image of a harvested tree crown candidate region extracted" displayed on the display. By extracting the candidate for the height of the tree crown from the tree crown height difference image data (see Fig. 5(b)), the lower layer vegetation less than 1/5 of the maximum crown height value (maxH) has a pixel value of zero (black) as noise. It is said that Therefore, in FIG. 6, the white area is the felled tree canopy area including the upper tree, the middle tree, and the lower tree. It should be noted that even if such a cut tree crown candidate region extraction process is performed, noise due to the tree crown height image data differential process still remains.

Next, from the "tree crown high difference image data for which the felled tree crown candidate area has been extracted", small areas of the felled tree canopy candidate area whose area is equal to or less than a predetermined area are removed, A small-area felled tree crown candidate area removal process (step S43) is performed to create "crown height difference image data from which the area-cut tree crown candidate area has been removed". Although there are many small crown regions in FIG. 6, for example, a fine white region N1 can be exemplified.

Specifically, the small-area felled tree crown candidate region removal processing is performed by removing the tree crown candidate region (the white region in FIG. 6) of the felled tree crown candidate region (white region in FIG. , a process of removing a large number of fine small-area felled tree crown candidate regions (for example, fine region N1 in FIG. 6) as small-area tree crown noise.

Whether or not the felled tree canopy candidate area is small is determined by inputting a value smaller than the crown area of the understory trees existing in the survey target forest area as a threshold (second threshold). It is determined whether or not the area of the candidate area is less than the second threshold. Specifically, since it is empirically obtained that the crown area of the understory tree to be felled is about 10 m ² , a smaller area (for example, 8 m ² ) is the second threshold value. set as Then, the small-area felled tree crown candidate region having an area smaller than the second threshold is removed from the "tree crown high-difference image data for which the felled tree crown candidate region has been extracted." As a result, a felled tree canopy candidate region having an area smaller than that of the understory tree (for example, 8 m ² ) can be removed as noise. Note that the second threshold is not limited to 8 m ² , and an optimum value can be set as appropriate.

In this way, the tree crown high-difference image data obtained by removing the felled tree crown candidate region smaller than the second threshold as a small-area felled tree crown candidate region from the "tree crown high-difference image data for which the felled tree crown candidate region has been extracted" is Canopy height difference image data from which the small-area felled tree canopy candidate region has been removed.

By storing the "tree crown height difference image data from which the small area felled tree crown candidate area has been removed" in an information processing device such as a computer, the "tree crown height difference image data from which the small area felled tree crown candidate area has been removed" can be displayed on the display at any time.

FIG. 7 is a diagram showing a “crown height difference image from which the small-area felled tree crown candidate region has been removed” displayed on the display. The "crown height difference image from which the small area felled tree crown candidate area has been removed" shown in FIG. Areas (for example, fine white areas N1 in FIG. 6) are removed.

In this way, when the crown height difference image data from which the small area crown area has been removed is created by the small area tree crown area removal process (step S43), subsequently, the longitudinal length of the felled tree canopy area is determined. Elongated tree crown regions with a value equal to or greater than a threshold value (third threshold value) obtained by dividing the tree crown candidate region in the transverse direction are referred to as "crown high-difference image data after small-area tree crown region removal". , and an elongated canopy area removing process (step S44) is performed to create "canopy height difference image data after removing the elongated canopy area".

A plurality of elongated tree canopies exist in FIG. 7, and for example, an elongated white region N2 can be exemplified. For example, even if the crown is not a felled tree, the difference in the shooting angle and the difference in the shooting position between the aerial image taken before and after felling can cause the height of the tree crown to increase. This may occur when image data difference processing (step S30) is performed. Therefore, processing is performed to exclude such elongated tree crowns from the tree crown candidates for the felled trees.

Specifically, the slender tree crown region removal processing is performed by removing the slender tree crown candidate region (the white region in FIG. 7) in the “crown height difference image from which the small area felled tree crown candidate region has been removed” shown in FIG. Processing is performed to remove the tree canopy candidate area of the felled tree (for example, the long and narrow white area N2 in FIG. 7) as the long and narrow canopy noise.

It should be noted that the determination of whether or not the tree crown is elongated is performed by comparing the longitudinal length of the felled tree crown candidate region to the short length Based on the value obtained by dividing by the height, it is determined whether or not it is an elongated felled tree canopy region. Specifically, it is obtained by dividing the length in the longitudinal direction (Lmax) of the crown of the tree to be felled in plan view by the length in the lateral direction (Lmin). Since it is empirically obtained that the magnitude of the value is unlikely to be 4 or more, for example, "n=4" is set when the third threshold is n. Then, a thin felled tree crown candidate area equal to or larger than the third threshold value (n=4) is removed from the "small area felled tree crown candidate area removed tree crown high-difference image data". As a result, extremely long and narrow felled tree crown candidate regions can be removed as noise. Note that the third threshold value n is not limited to n=4, and an optimum value can be set as appropriate.

By performing such processing for removing the elongated felled tree crown candidate area, it is possible to create "crown height difference image data from which the elongated felled tree crown candidate area has been removed." By storing the created "tree crown height difference image data after removal of the elongated tree crown candidate area" in an information processing device such as a computer, the elongated felled tree crown corresponding to the removed elongated tree crown candidate area image data can be obtained. The canopy height image from which the candidate regions have been removed can be displayed on the display at any time.

FIG. 8 is a diagram showing a tree crown height difference image from which the elongated felled tree tree crown candidate region has been removed, displayed on the display. The tree crown high-difference image from which the slender tree crown has been removed shown in FIG. has been removed as

In this way, when the thin felled tree crown candidate area removal process (step S45) creates the "tree crown high-difference image data after the thin felled tree crown candidate area has been removed". , removes noise existing in the felled tree crown candidate region (inside the white region) (referred to as noise N3 in the felled tree crown candidate region) from the "crown height difference image data after removal of the thin felled tree crown candidate region". Then, noise removal processing within the felled tree canopy candidate region is performed (step S45).

Specifically, the noise removal processing in the felled tree crown candidate region is performed by removing the noise in the felled tree crown candidate region (white region in FIG. Among these processes, the area (black area) having a pixel value of zero existing in the felled tree canopy candidate area (inside the white area) is removed as the felled tree canopy candidate area noise N3. A black area with a pixel value of zero existing in the felled tree canopy candidate area (white area), that is, noise N3 in the felled tree canopy candidate area, is surrounded by the felled tree canopy candidate area (white area). It indicates a black area with a pixel value of zero that exists like an island.

Thus, the noise N3 in the harvested tree canopy candidate area is assumed to be a black area surrounded by the harvested tree canopy candidate area (white area) like an island and having a pixel value of zero. For this reason, even if there is a black region (region with zero pixel values) that looks as if there is a black region with zero pixel values in the felled tree crown candidate region (inside the white region), For a black area (for example, black area D in FIG. 8) that is connected to a black area that exists outside the felled tree crown candidate area (white area), the black area (for example, The black area D) is not used as noise in the felled tree crown candidate area N3.

The noise N3 in the felled tree canopy candidate area is generated when the ground is photographed through the space, such as when there is a gap in the tree crown when the survey target forest area is photographed from above. be. In other words, when considering a tree, if there is a space in the crown of the tree, the ground will be captured through the space in the tree crown in the image taken before the tree is cut down. When the difference between the photographed image and the photographed image of the tree after felling is taken, the pixel value becomes zero (black) because the crown height does not change in the part corresponding to the space. Such noise N3 in the felled tree canopy region can be said to be noise derived from gaps that existed in the canopy of the felled tree. The noise in the felled tree crown area N3 is originally an area that can be considered as the crown of the tree. Therefore, noise N3 in the felled tree canopy candidate region (black region with zero pixel value) existing in the felled tree canopy candidate region (inside the white region in FIG. 8) is filled with white.

In addition, in the felled tree crown candidate region, there is not only noise derived from gaps in the crown of the felled tree, but also the noise from the crown of one tree and the crowns of trees adjacent to the one tree. There may be some black areas with zero pixel values (for example, black areas denoted by reference numeral N4 in FIG. 8) due to the space that originally exists between them.

When the noise removal process in the felled tree crown candidate area (step S44) is performed, the black area N4 having a pixel value of zero due to the space that originally exists between the tree crowns is also painted in white. However (see FIG. 9), the black area N4 can be reproduced as a space in the felled tree canopy segmentation process (step S50), which will be described later.

By performing such noise removal processing in the felled tree canopy candidate region, it is possible to create "crown height difference image data in which noise in the felled tree crown candidate region has been removed". By storing the created "noise-removed crown height difference image data in the felled tree crown candidate area" in an information processing device such as a computer, the "noise-removed tree crown height in the felled tree crown candidate area" The "canopy height difference image after noise removal in the felled tree crown candidate region" corresponding to the "difference image data" can be displayed on the display at any time.

FIG. 9 is a diagram showing a “crown height difference image from which noise has been removed in the felled tree crown candidate region” displayed on the display. The noise-removed image within the felled tree canopy region shown in FIG. 9 is derived from the image with the elongated tree canopy region removed (within the frame line A1) shown in FIG. is zero (noise N3 in the felled tree crown candidate region and black region N4 with zero pixel value due to the space originally existing between tree crowns) is removed as noise in the felled tree crown candidate region. ing.

In this way, the tree crown height image (see FIG. 9) that has undergone noise removal processing in the felled tree crown candidate region (step S45) and has been subjected to the noise removal processing in the felled tree crown candidate region (see FIG. 9) is subjected to the cut tree crown height candidate region extraction processing. The small-area felled tree crown candidate area removal process (step S43) is performed from the "cropped tree crown height candidate area extracted tree crown height difference image data" for which (step S42) has been performed, and further, the slender tree crown candidate removal process ( Step S44) has been performed, and in addition, the tree crown high-image difference image has been subjected to noise removal processing in the felled tree crown candidate region (step S45). Therefore, the "noise-removed crown height difference image data in the felled tree crown candidate region" created by the noise removal processing in the felled tree crown candidate region (step S45) is a felled tree crown height image from which various noises have been removed. data.

When the "noise-removed tree crown high-difference image data in the felled tree crown candidate region" is created as described above, the tree crown high-difference image data in the felled tree canopy candidate region after noise removal has been performed. A tree canopy segmentation process is performed (step S50). In the cut tree crown segmentation process (step S50), the cut tree crown candidate area (white area) existing in the "crown height difference image" (see FIG. 9) from which "noise due to the crown height image data difference processing" is removed. are divided into crowns of individual trees to create cut tree crown division image data.

That is, in the tree crown candidate region (white region) in the "noise-removed tree crown height difference image in the tree crown candidate region" shown in FIG. It is not classified as individual tree crowns because it is represented as a state in which it is covered. Therefore, processing is performed to classify the felled tree canopy candidate region in the "noise-removed tree crown height difference image in the felled tree canopy region" shown in FIG.

Specifically, the cut tree crown segmentation process uses "crown height difference image data with noise removed in the cut tree crown candidate region" and "crown height difference image data" representing the tree crown height. do. That is, the "noise-removed tree crown height difference image data in the felled tree canopy candidate region" is binary image data as shown in FIG. Therefore, it does not have information on crown height. For this reason, when performing the tree crown classification processing for a felled tree, "crown height difference image data" having information on the tree crown height as shown in the "enlarged image of the tree crown height difference image" in FIG. Refer to it and perform the cut tree canopy division processing.

The cut tree crown segmentation process for classifying the tree crown candidate region in the "tree crown high-difference image from which noise has been removed in the cut tree crown candidate region" shown in FIG. The individual tree detection method for single tree crown extraction described in Reference 2 can be used. In this case, the harvested tree canopy classification process is performed using the individual tree detection method for extracting the canopy of a single tree, utilizing the fact that the canopy height at the edge of the canopy is lower than the canopy height at the apex of the canopy.

Publicly known document 2: J Hyypp M Inkinen, ``Detecting and estimating attributes for single trees using laser scanner'', The photogrammetric journal of Finland 16 (2), 1999, 27-42

By performing the cut tree crown division processing (step S50), cut tree crown division image data in which the crown of each cut tree is divided for each cut tree is created. By storing the cut tree crown segment image data in the information processing device, the cut tree crown segment image corresponding to the cut tree crown image data can be displayed on the display at any time.

FIG. 10 is a diagram showing a cut tree canopy segmented image displayed on the display. The cut tree crown segmented image shown in FIG. 10 is obtained by segmenting the crown of each tree. In FIG. 10, one divided white area (for example, area C1) is the crown of one tree.

In this way, in the cut tree crown classification process (step S50), when the cut tree crown classification image data corresponding to each tree (cut tree) is created, the resource information (cut tree resource information) on each cut tree is generated. ) is performed (step S60).

The felled tree resource information acquisition process (step S60) includes at least position information (x-coordinate and y-coordinate ), the crown area and/or crown diameter of each felled tree, the diameter at breast height (DBH) of each felled tree, the height of each felled tree, and the volume of each felled tree, as felled tree resource information. is Then, the cut tree resource information is stored in an information processing device such as a computer as a database in which the number (label number) assigned to each cut tree, tree species, etc. are associated with each other. The felled tree resource information stored in the information processing device can be displayed on the display at any time.

In addition, as resource information on individual trees before felling in the forest area surveyed, the location information of each tree, the crown area and/or crown diameter of each tree, the height of each tree, the diameter at breast height of each tree and When the timber volume of each tree is obtained as pre-cutting resource information for each individual tree, the pre-cutting resource information may be referenced to obtain the resource information for each felled tree. Alternatively, the resource information about each felled tree (individual felled tree) may be obtained by calculation or the like after felling.
Here, it is assumed that the resource information about each felled tree (individual felled tree) is obtained by calculation or the like after felling.

A method for obtaining resource information about individual felled trees (cut trees) after felling will be described. The tree height of each felled tree can be obtained by taking the maximum value of the crown height image data corresponding to the crown of each felled tree classified by the felled tree crown classification process as the tree height of the tree.

Moreover, the crown area can be obtained by accumulating the number of pixels in the crown area of each felled tree, and the crown diameter can also be obtained based on the number of pixels in the crown area.

In addition, the diameter at breast height can be determined from the crown area or crown diameter, and the tree height using a multiple regression equation. When determining the diameter at breast height using the multiple regression equation, select about 10 standard trees for each tree species in a field survey, and measure the crown area, crown diameter, and tree height of the selected standard trees to calculate the variables of the multiple regression equation. It is preferable to ask

By the way, the "crown diameter" indicates the size of the horizontal spread of the tree crown when viewed from above, and the "breast height diameter" indicates the thickness of the trunk of the tree. , Since the crown and the trunk of a tree are not actually a perfect circle, it cannot be strictly called a "diameter", but in this specification, the "crown diameter ”, and “crown diameter” is used to represent the thickness of the trunk.

Also, the volume can be obtained from a two-variable volume formula of diameter at breast height and tree height. Note that a quick reference chart for determining the timber volume from the diameter at breast height and the tree height is known, and the timber volume can be obtained from such a quick reference chart. The quick reference table is described in the following documents (publicly known documents 3 and 4).

Publicly Known Document 3: Forestry Agency Planning Division, Tree Trunk Volume Table East Japan Edition, Japan Forestry Research Institute, page 334, 2003 Publicly Known Document 4: Forestry Agency Planning Division, Tree Trunk Volume Table West Japan Edition, Japan Forestry Research Institute, Page 320, 1970

FIG. 11 is a diagram showing a list of felled tree resource information displayed on the display. In FIG. 11, the felled tree resource information includes a serial number such as 1, 2, 3, . x-coordinate and y-coordinate), tree species, diameter at breast height (DBH), tree height, and volume. The felled tree resource information to be displayed on the display may include tree crown diameter, tree crown area, etc., in addition to the above, if necessary. In addition, the serial numbers such as 1, 2, 3, . The number of target trees can be obtained. By displaying the felled tree resource information shown in FIG. 11 as a list on the display, it is possible to visually grasp the detailed resource information of the felled trees in the survey target forest area.

In addition, by displaying the tree height of each felled tree in association with each crown of the felled tree crown classification image (see Fig. 11) on the display, the resource overview of each felled tree can be visually grasped. can. The tree height of each felled tree may be a numerical value that represents the tree height itself, but for example, a mark representing the tree height division that divides the tree height into multiple stages is attached to the crown of each felled tree. The cut tree height classification image may be displayed on the display.

FIG. 12 is a diagram showing a felled tree height classification image in which marks representing tree height classifications are added to the crowns of individual felled trees in the accuracy verification area (within the frame line A2). In addition, in FIG. 12, the circle marks (accurately not circular, but written as circle marks) shown in individual tree crowns are attached to the positions of the tree apexes of individual felled trees. The size of the circle represents the tree height (m) for each stage. Here, the tree height is divided into four stages, such as 18.6m to 20.0m, 20.1m to 25.0m, 25.1m to 30.0m, and 30.1m or more.

Since FIG. 12 is a monochrome image, it is difficult to grasp the tree height distribution of the felled trees in FIG. Since they are displayed as marks, the tree height distribution of the felled trees can be easily grasped.

FIG. 13 is a diagram showing a distribution image of felled trees in the survey target forest area (within the frame line A1). In addition, the circled tree crowns in FIG. 13 indicate trees cut down in the survey target forest area (inside the frame line A1). By displaying the felled tree distribution image shown in FIG. 13 on the display, it is possible to visually grasp at which position the tree in the survey target forest area was felled. As a result, for example, it is possible to easily determine whether or not the felled trees are distributed throughout the survey target forest area without concentration or unevenness, and whether there are any uncut trees.

Since FIG. 13 is a monochrome image, it is difficult to grasp the distribution of felled trees in FIG. Because it is displayed as , the distribution of felled trees can be easily grasped.

The tree height classification image shown in FIG. 12 and the felled tree distribution image shown in FIG. 13 can be displayed on the display. , can be visually grasped. For example, it is possible to easily determine whether the felled trees are distributed throughout the survey area, and whether they are concentrated or uneven. In addition, when actually going to the site, by bringing a printout of each of these information or a mobile terminal that can display them on site and using GPS together, even non-experts can easily You can check the trees that have been felled.

When the harvested tree resource information is obtained as described above, the harvested tree resource amount is tallied (step S70). The felled tree resource amount counting process is based on the individual felled tree resource information (for example, see FIG. 11) acquired by the felled tree resource information acquisition process (step S60). To be able to compare with the amount of resources in the forest area surveyed in 1).

In the felled tree resources totaling step, the number of felled trees, the average diameter at breast height of the felled trees, the average height of the felled trees, and the total timber volume of the felled trees are aggregated, and the felling rate based on the number of trees and the felling rate based on the timber volume are calculated. It contains the processing to do.

Specifically, the number of trees existing in the surveyed forest area before felling, the number of felled trees (cut trees), and the average diameter at breast height (DBH: cm) of the trees existing in the surveyed forest area before felling and the average diameter at breast height (DBH: cm) of felled trees (cut trees), the average tree height (m) of trees existing in the surveyed forest area before felling, and the average tree height (m ), the total volume of trees existing in the surveyed forest area before logging (m ³ ) and the total volume of felled trees (cut trees) (m ³ ), the number of trees existing in the surveyed forest area before logging and the ratio (%) of the volume of trees existing in the surveyed forest area before felling are calculated and tabulated. As a result, the tabulated results can be displayed on the display at any time as a list.

FIG. 14 is a diagram showing a tally table of the amount of felled tree resources displayed on the display. According to the summary table of the amount of felled trees shown in Fig. 145, the number of trees felled in the survey target forest area (cut trees), the average diameter at breast height of the cut trees, the average height of the cut trees, and the total timber volume of the cut trees. , the number of trees before felling, the average diameter at breast height, the average tree height, and the total volume of trees in the target forest area can be compared. In addition, as shown in FIG. 14, the felled tree resource amount summary table also records the felling rate (the number of trees felling rate and the timber pile felling rate).

According to the summary table of the amount of felled tree resources shown in FIG. 14, it is possible to easily grasp the state of felling (investigation) carried out in the survey target forest area. In addition, based on the summary table of the amount of felled tree resources shown in Fig. 14, the number cutting rate and the timber cutting rate, which are the basis for calculating the subsidy expenditure in the logging survey work, can be set to the appropriate cutting rate (for example, the number cutting rate 30% or less, and the logging rate is less than 35%). According to the summary table of the amount of felled tree resources shown in FIG. Since it satisfies the requirements, it can be determined to be appropriate.

As described above, according to the felling situation survey data creation method according to the embodiment, the tree crown height image data difference processing for obtaining the difference between the tree crown height image data before felling and the tree crown height image data after felling is performed. to create crown height difference image data representing the canopy area of the felled tree, and noise due to the crown height image data difference processing is removed by noise removal processing (step S40), and maximum crown height extraction processing (step S41). , felled tree crown candidate region extraction processing (step S42), small area felled tree crown candidate region removal processing (step S43), slender tree crown candidate region removal processing (step S44), and noise removal processing in the felled tree crown candidate region (step S45 ) to create noise-removed tree canopy height difference image data.

Then, using the noise-removed tree crown high-difference image data, felled tree crown segmented image data in which the tree crown region of the felled tree is segmented for each tree crown is created. As a result, each felled tree can be reproduced with high precision, and the felled tree resource information is acquired based on the felled tree canopy segmented image data that reproduces each felled tree with high precision. As a result, felled tree resource information can be obtained with high accuracy. In addition, since the amount of felled tree resources is aggregated based on the information on felled tree resources obtained with high accuracy, the aggregate result of the amount of felled tree resources ( For example, see FIG. 14.) is a highly reliable tally result of the amount of felled tree resources.

In addition, the cut tree resource information acquired by the cut tree resource information acquisition process, the tally result of the felled tree resource amount obtained by the tally processing of the felled tree resource amount, and the tree height classification image of the felled tree shown in FIG. The felled tree distribution image and the like shown in FIG. 13 can be used as felling situation investigation data for investigating the felling situation when trees are felled in a predetermined forest area.

The felling situation investigation data creation method according to the embodiment has been described above. Subsequently, the felling situation investigation data creation device corresponding to the felling situation investigation data creation method according to the embodiment will be explained.

FIG. 15 is a diagram for explaining the harvesting survey data creation device 1 corresponding to the harvesting survey data creation method according to the embodiment. As shown in FIG. 15, the felling situation survey data creation device 1 includes a storage unit 10 storing tree canopy height image data before felling to be acquired in step S10 in FIG. A storage unit 20 storing tree crown height image data after felling to be performed, a tree crown height image data difference processing unit 30 performing the processing of step S30 in FIG. 1, and a noise removal processing performing the processing of step S40 in FIG. 1, a felled tree crown classification processing unit 50 that performs the processing of step S50 in FIG. 1, a harvested tree resource information acquisition processing unit 60 that performs the processing of step S60 in FIG. 1, and a processing of step S70 in FIG. and a tally processing unit 70 for the amount of felled tree resources.

The noise removal processing unit 40 in FIG. 15 includes a maximum crown height extraction unit 41 that performs the processing of step S41 in FIG. The small area felled tree crown candidate area removal processing unit 43 that performs the process of step S43 in , the elongated felled tree crown candidate area removal processing unit 44 that performs the process of step S44 in FIG. 1, and the process of step S45 in FIG. A noise removal processing unit 45 in the felled tree canopy candidate area is included.

Each component of the harvesting survey data creation device 1 configured as described above performs processing according to the procedure shown in the flowchart shown in FIG. Similar to the above method, it is possible to obtain the felled tree resource information, the tally result of the felled tree resource amount, the height classification image of the felled trees shown in FIG. 12, the felled tree distribution image shown in FIG. can be used as felling situation survey data for investigating the felling situation when trees are felled in a predetermined forest area.

It should be noted that the functions of the respective components (see FIG. 15) included in the felling survey data creation device 1 shown in FIG. 15 are installed as computer programs, and given to the respective components. By doing so, the functions of each component are executed on the computer software.

It should be noted that the present invention is not limited to the above-described embodiments, and various modifications can be made without departing from the gist of the present invention. For example, modified implementations as shown below are also possible.

(1) Examples of trees existing in the survey target forest area include coniferous planted trees such as larch, Japanese red pine, cedar, and cypress. However, it is not limited to coniferous trees. For example, if there is a forest in which beech and oak grow together in a predetermined area, the same can be applied to the beech and oak.

(2) In the above embodiment, a drone is used to obtain aerial image data, but the invention is not limited to drones. For example, an aircraft may be used to obtain aerial image data. .

(3) In the above embodiment, three-dimensional point cloud data for obtaining tree canopy height image data is created by creating an orthoimage from the aerial image data and creating it from the created orthoimage. Three-dimensional point cloud data may be obtained based on laser measurement data obtained by irradiating laser light.

(4) In the above embodiment, the individual tree detection method is used for the cut tree canopy division processing, but other methods can also be used. For example, the Watershed algorithm, which is one of the area division methods, can be used.

(5) In the above-described embodiment, the felled tree resource information acquisition process performed in step S60 is a case where resource information on each felled tree (individual felled tree) is obtained by calculation or the like after felling. However, as described above, the resource information on individual trees before felling in the forest area surveyed includes location information of individual trees, crown area and/or crown diameter of individual trees, height of individual trees , If the diameter at breast height and timber volume of each tree are required as pre-cutting resource information for each tree, resource information for each felled tree is obtained by referring to the pre-cutting resource information. may be obtained. In this case, based on the felled tree crown segmentation image data created by the felled tree crown segmentation process in step S50, the positional information of each felled tree is obtained, and based on the positional information, each felled tree Obtain tree resource information from pre-logging forest resource information.

Reference numerals 1... Data creation device for felling survey, 10... Tree crown height image data storage unit before logging, 20... Tree crown height image data storage unit after felling, 30... Tree crown height image data difference processing Section 40 Noise removal processing section 41 Maximum tree crown height extraction processing section 42 Trimmed tree crown candidate region extraction processing section 43 Small area felled tree crown candidate region removal processing section 44 Elongated tree crown candidate area removal processing unit 45 Cut tree crown candidate area noise removal processing unit 50 Cut tree crown classification processing unit 60 Cut tree resource information acquisition processing unit 70 . . . Aggregation processing unit for felled tree resource amount, S10 .. Tree crown height image data acquisition process before felling, S20 .. Tree crown height image data acquisition process after felling, S30 . , S40... noise removal process, S41... maximum tree crown height extraction process, S42... felled tree crown candidate area extraction process, S43... small area felled tree crown candidate area removal process, S44... slender tree crown Candidate area removal process, S45... Noise removal process in felled tree canopy candidate area, S50... Felled tree crown classification process, 60... Felled tree resource information acquisition process, S70... Aggregation of felled tree resources amount process

Claims (13)

A felling situation investigation data creation method for creating data for investigating the felling situation when trees are felled in a predetermined forest area,
A pre-felling tree crown height image data acquisition step of acquiring pre-felling tree crown height image data in the predetermined forest area;
A post-felling tree crown height image data acquisition step of acquiring post-felling tree crown height image data in the predetermined forest area;
A canopy height difference image representing the canopy height in a canopy region composed of the canopy of one or more felled trees by taking the difference between the canopy height image data before felling and the canopy height image data after felling. a crown height image data difference processing step for creating data;
a noise removal processing step of removing noise present in the tree canopy height difference image data created by the tree canopy height image data difference processing to create denoised tree canopy height difference image data;
Cut tree crown segmentation, in which a tree crown region composed of the tree crown of the one or more cut trees is segmented for each tree crown based on the noise-removed tree crown high-difference image data created by the noise removal process. A felled tree canopy segmentation processing step for creating image data;
a felled tree resource information acquisition step of acquiring resource information about each of the felled trees based on the felled tree crown segmentation image data created by the felled tree crown segmentation process;
A data creation method for logging situation survey, characterized by having In the felling situation survey data creation method according to claim 1,
In the noise removal processing step,
A maximum crown height extraction processing step of extracting the maximum tree crown height from the tree crown height difference image data created by the tree crown height image data difference processing step;
A threshold value of the crown height for making the tree crown candidate for the felled tree is set based on the maximum crown height, the threshold value is set as a first threshold value, and the tree crown height difference image data has a crown height equal to or higher than the first threshold value. A felled tree crown candidate region extraction processing step of extracting the region as a felled tree crown candidate region and creating "crown height difference image data for which the felled tree crown candidate region has been extracted";
setting a threshold value for the area of the felled tree crown candidate area, setting the set threshold value as a second threshold value, and a small felled tree crown candidate area in which the area of the felled tree crown candidate area is less than the second threshold value; A small-area felled tree crown candidate area for creating "crown high-difference image data after removal of the small-area felled tree crown candidate area" by removing the area from the above-mentioned "cropped tree crown candidate area-extracted crown high-difference image data" a removal processing step;
setting a threshold to a value obtained by dividing the longitudinal length of the felled tree crown candidate region by the transverse length of the felled tree crown candidate region, and using the set threshold as a third threshold; Elongated felled tree crown candidate areas having a felled tree crown candidate area equal to or greater than the third threshold value are removed from the "crown high-difference image data after removal of the small-area felled tree crown candidate area", and the "elongated felled tree crown candidate area" is removed. a slender canopy candidate area removal processing step for creating "removed tree crown height difference image data";
The noise in the felled tree candidate area existing inside the felled tree crown candidate area is removed from the "crown high-difference image data after removal of the thin felled tree crown candidate area", and the "noise in the felled tree canopy candidate area has been removed." a noise removal processing step in the felled tree canopy candidate region for creating the canopy high-difference image data of
A data creation method for deforestation survey, characterized in that the data is included. In the felling situation survey data creation method according to claim 1 or 2,
The cut tree resource information acquisition processing step includes at least position information of each cut tree, crown area of each cut tree, based on the cut tree crown section image data created by the cut tree crown section processing step. and/or a method for creating data for felling survey, characterized by including a process of obtaining crown diameter, height of each felled tree, breast-height diameter of each felled tree, and volume of each felled tree. In the logging situation survey data creation method according to claim 3,
The cutting situation survey, further comprising a felled tree resource quantity totaling processing step of totaling the felled tree resource quantity based on the resource information relating to the individual felled trees acquired in the felled tree resource information acquisition processing step. Data creation method for In the logging situation survey data creation method according to claim 4,
In the felled tree resource quantity totaling step, the number of felled trees, the average diameter at breast height of the felled trees, the average height of the felled trees, and the total volume of the felled trees are totaled, and the felling rate and A method for creating data for deforestation survey, characterized by including a process of calculating a deforestation rate based on the volume of lumber. In the felling situation survey data creation method according to any one of claims 3 to 5,
The crown area and/or crown diameter of each cut tree is obtained based on the number of pixels in the crown area of each cut tree,
The tree height of each of the felled trees is obtained as the tree height of the felled tree by the maximum value of the crown height image data corresponding to the crown of each of the felled trees classified by the felled tree crown classification process,
The breast height diameter of each cut tree is obtained by a multiple regression equation from the crown area or crown diameter of each cut tree and the height of each cut tree,
The volume of each cut tree is obtained based on the breast height diameter of each cut tree and the height of each cut tree,
A method for creating data for deforestation survey, characterized by: In the felling situation survey data creation method according to any one of claims 1 to 6,
The pre-felling tree crown height image data obtained by the pre-felling tree crown height image data obtaining processing step is
Based on the predetermined forest area image data before logging created based on the three-dimensional point cloud data corresponding to the predetermined forest area before logging and the geographical information corresponding to the predetermined forest area, mesh It is created by creating digital surface model data and digital elevation model data, and taking the difference between the digital surface model data and the digital elevation model data for each mesh,
The post-felling tree crown height image data acquired by the post-felling tree crown height image data acquisition processing step is
Based on the predetermined forest area image data after logging created based on the three-dimensional point cloud data corresponding to the predetermined forest area after logging and the geographical information corresponding to the predetermined forest area, mesh created by creating digital surface model data and digital elevation model data, and taking the difference between the digital surface model data and the digital elevation model data for each mesh,
A method for creating data for deforestation survey, characterized by: In the felling situation survey data creation method according to any one of claims 1 to 7,
The cut tree crown segmentation process uses an individual tree detection method to generate cut tree crown segment image data in which the tree crown region of the cut tree is segmented into individual tree crowns. A method for creating data for deforestation survey, characterized in that: A logging situation investigation data creation device for creating data for investigating the logging situation when trees are felled in a predetermined forest area,
A pre-felling tree crown height image data acquisition processing unit that acquires pre-felling tree crown height image data in the predetermined forest area;
a post-felling tree crown height image data acquisition processing unit that acquires post-felling tree crown height image data in the predetermined forest area;
A canopy height difference image representing the canopy height in a canopy region composed of the canopy of one or more felled trees by taking the difference between the canopy height image data before felling and the canopy height image data after felling. a crown height image data difference processing unit that creates data;
a noise removal processing unit that removes noise present in the tree crown height difference image data created by the tree crown height image data difference processing, and creates denoised tree crown height difference image data from which noise has been removed;
Cut tree crown segmentation, in which a tree crown region composed of the tree crown of the one or more cut trees is segmented for each tree crown based on the noise-removed tree crown high-difference image data created by the noise removal process. a felled tree crown segmentation processing unit that creates image data;
a felled tree resource information acquisition processing unit that acquires resource information about each of the felled trees based on the felled tree crown segmentation image data created by the felled tree crown segmentation process;
A deforestation survey data creation device, comprising: In the deforestation survey data creation device according to claim 9,
In the noise removal processing unit,
A maximum crown height extraction processing unit for extracting the maximum tree crown height from the tree crown height difference image data created by the tree crown height image data difference processing unit;
A threshold value of the crown height for making the tree crown candidate for the felled tree is set based on the maximum crown height, the threshold value is set as a first threshold value, and the tree crown height difference image data has a crown height equal to or higher than the first threshold value. a felled tree canopy candidate area extraction processing unit that extracts an area as a felled tree canopy candidate area and creates "crown height difference image data for which the felled tree canopy candidate area has been extracted";
setting a threshold value for the area of the felled tree crown candidate area, setting the set threshold value as a second threshold value, and a small felled tree crown candidate area in which the area of the felled tree crown candidate area is less than the second threshold value; A small-area felled tree crown candidate area for creating "crown high-difference image data after removal of the small-area felled tree crown candidate area" by removing the area from the above-mentioned "cropped tree crown candidate area-extracted crown high-difference image data" a removal processing unit;
setting a threshold to a value obtained by dividing the longitudinal length of the felled tree crown candidate region by the transverse length of the felled tree crown candidate region, and using the set threshold as a third threshold; Elongated felled tree crown candidate areas having a felled tree crown candidate area equal to or greater than the third threshold value are removed from the "crown high-difference image data after removal of the small-area felled tree crown candidate area" to obtain the "elongated felled tree crown candidate area." a slender tree crown candidate region removal processing unit that creates "the removed tree crown height difference image data";
Remove noise in the felled tree candidate area existing inside the felled tree crown candidate area from the "crown high difference image data after removal of the thin felled tree crown candidate area" to obtain "noise removed in the felled tree crown candidate area" a noise removal processing unit in the felled tree canopy candidate region that creates the canopy height difference image data of
A deforestation survey data creation device, characterized in that In the deforestation survey data creation device according to claim 9 or 10,
The felled tree resource information acquisition processing unit provides at least position information of each felled tree, crown area of each felled tree, and / Or a felling situation investigation data creation device characterized by performing a process of acquiring a crown diameter, a tree height of each felled tree, a breast-height diameter of each felled tree, and a lumber volume of each felled tree. In the felling situation investigation data creation device according to claim 11,
A felling situation survey, further comprising: a felled tree resource quantity aggregation processing unit that aggregates the felled tree resource quantity based on the resource information relating to the individual felled trees acquired by the felled tree resource information acquisition processing unit. Data preparation device for In the logging situation survey data creation device according to claim 12,
The felled tree resource quantity aggregation processing unit aggregates the number of felled trees, the average diameter at breast height of the felled trees, the average tree height of the felled trees, and the total volume of the felled trees, and the felling rate based on the number of trees and the A deforestation survey data creation device characterized by performing processing for calculating a deforestation rate based on timber volume.

Images