JP2017169520A - Grape cultivation management method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a grape cultivation management method of capable of properly and efficiently performing management of pruning and new shoots and the like by recognizing the state of branches and sub branches and the number of new shoots in an overhead view image of the vine taken by a compact unmanned aircraft.SOLUTION: The method comprises: (a1) imaging while scanning in order a predetermined cultivation area composed of one or more vines; (a2) forming an overhead view image of the whole cultivation area from the images that were taken; (a3) forming an extracted image obtained by extracting the form of the crown of the main branches and sub branches of the vines with each main trunk as the center; (a4) dividing the extracted image into a plurality of pruning management areas; and (a5) specifying the branches for pruning in each of the pruning management areas.SELECTED DRAWING: Figure 1

Description

  The present invention relates to a grape cultivation management method for appropriately and efficiently performing pruning of vines and management of new treetops in the dormant season.

  As for the cultivation of grapes, pruning and other operations are necessary to prepare the crown during the dormancy period of the vines. By appropriately performing such operations as pruning, a new treetop is sprouting from the mother branch as a result of the tree crown, and a grape cluster can be greatly grown on the treetop (Patent Document 1).

  Conventionally, operations such as pruning have been carried out based on the experience and judgment of skilled growers by visually checking the branching of the vines from under the vines.

JP 2009-11296 A

  However, when the operation such as pruning is visually performed from under the vines, it is possible to confirm the branching, new treetops, and the like only within the range of view. For this reason, it is difficult to accurately grasp the crown form of the whole vine consisting of the main branch extending from the main trunk in multiple directions and the multiple sub-branches branching from this main branch. It is not easy to judge.

  In addition, the actual pruning work is often performed based on the experience and intuition of skilled growers, but the problem is that cultivation techniques are not passed down due to the aging of growers and the decrease in successors. There is. On the other hand, for technical guidance of cultivation, there are manuals using schematic diagrams of ideal grape canopies, but there are portions that are not suitable for actual grape cultivation and are often used as reference levels.

  Therefore, an object of the present invention is to recognize the shape of the main branches and sub-main branches, the number of new treetops, and the like from the overhead image view of the vines aerial photographed using a small unmanned aerial vehicle. It is providing the cultivation management method of the grape which can perform management of new shoots etc. appropriately and efficiently.

  In order to solve the above-mentioned problem, the invention according to claim 1 shoots while sequentially scanning a predetermined cultivation area composed of one or a plurality of vines, and forms an overhead image diagram of the entire cultivation area from the photographed image. Then, based on this bird's-eye view image diagram, an extracted image diagram is formed by extracting the crown form of the main branch and the sub-main branch centered on each main trunk of the vine, and this extracted image diagram is formed in a plurality of pruning management areas. A branch to be pruned is specified for each pruning management area.

  The invention which concerns on Claim 2 is image | photographed scanning the predetermined cultivation area which consists of one or several vines in order, forms the bird's-eye view image figure of the whole cultivation area | region from the image | photographed image, Based on this, the extracted image figure which extracted the crown form of the main branch and sub-main branch centering on each main trunk of the said vine is formed, and this extracted image figure and the recommended crown image which has a preset recommended crown form The branches to be pruned are identified by comparing with the figure.

  The invention which concerns on Claim 3 is image | photographed scanning the predetermined cultivation area which consists of one or several vines in order, forms the bird's-eye view image figure of the whole cultivation area | region from the image | photographed image, Based on this, an extracted image diagram in which the crown forms of the main branches and sub-main branches centering on each main trunk of the vine is extracted, and the distribution or number of new shoots is detected by color discrimination from the extracted image diagram.

  According to the invention according to claim 1 of the present invention, based on a bird's-eye view image of the entire grape cultivation region, the crown forms of the main branches and sub-main branches centered on each main trunk of the vine are schematically illustrated. By forming an extracted image diagram and dividing the extracted image diagram into a plurality of pruning management areas, it is possible to easily grasp the branching of the main branches and sub-main branches for each pruning management area. Can do. Accordingly, it is possible to specify branches to be pruned for each pruning management area, and it is possible to easily and reliably perform subsequent pruning work instructions and actual pruning work.

  According to the invention which concerns on Claim 2, based on the extracted image figure obtained from the bird's-eye view image figure of the cultivation area | region of the whole grape, The crown form which consists of the main branch and sub-main branch which are shown by this extracted image figure, A branch to be pruned can be automatically detected by comparing with a recommended crown image figure having a canopy shape that is preset and appropriately pruned.

  According to the invention which concerns on Claim 3, the main trunk of one grape tree is selected from the extracted image figure which modeled the form of the main branch and sub-main branch centering on each main trunk of a vine, and this selected main trunk is selected. Is divided into a plurality of shoot management areas, and the distribution and number of shoots can be detected for each divided shoot management area. Thereby, compared with the appropriate number of shoots set for each shoot management area, the distribution location or number of shoots to be left can be selected, and the growth prediction of subsequent fruit bunches can be performed. it can.

It is a schematic flowchart of the cultivation method of the grape of this invention. It is a conceptual diagram which shows a mode that the cultivation area | region of grape is taken aerial by the drone from the sky. It is a bird's-eye view image figure which took aerial image of the cultivation field of grapes from the sky. It is an image figure which shows a part of said bird's-eye view image figure. It is the extraction image figure which extracted the main branch and the sub-main branch from the said bird's-eye view image figure. It is the extracted image figure divided | segmented into the some pruning management area. It is the extracted image figure which modeled in the said pruning management area. It is a recommended tree crown image figure in the pruning management area. It is an anticipation bird's-eye view image figure after pruning. It is the extracted image figure which modeled in the several shoots management area.

  Hereinafter, embodiments of a grape cultivation management method according to the present invention will be described in detail with reference to the accompanying drawings. The grape cultivation management method of the present invention relates to a pruning operation performed from the end of November to the end of February, which is a dormant period of vines, and this pruning operation is performed in a planned and efficient manner, and thereafter This is to predict the growth of new shoots and fruit bunches. The outline is aerial photography of the dormant vines where the bunches and leaves have fallen into branches only, and based on this aerial image, the location to be pruned and the management of new treetops, etc. Is to do.

  FIG. 1 shows an outline of a grape cultivation management method according to the present invention. In the first method, as shown in FIG. 2, first, the grape cultivation area 10 is sequentially photographed from the sky along a predetermined area E (a1), and the images of the areas E that are continuously photographed are connected. Thus, an overhead image diagram of the cultivation area 10 as shown in FIG. 3 is obtained (a2). Next, as shown in FIG. 4, by performing image processing on a predetermined area E of the overhead image diagram, an extracted image diagram obtained by extracting the main branch B and the sub-main branch C as shown in FIG. 5 is obtained (a3). . The extracted image diagram is divided into a plurality of pruning management areas (a4), and a location to be pruned is specified for each pruning management area (a5). This part to be pruned is performed by checking on the screen a part where the sub-main branches C closely overlap each other or the interval is short.

  The second method compares the extracted image diagram shown in FIG. 7A with the preset recommended tree image diagram shown in FIG. 7B after FIG. 1 (a1) to (a3) (a6), and pattern matching. The pruning point is specified by (a7). For example, a plurality of recommended crown image diagrams based on a canopy pattern that has been appropriately pruned are registered in advance in a database, and a recommendation that has a branching similar to the extracted image diagrams obtained by FIGS. 1 (a1) to (a3) A tree image is selected from the database. Then, by matching the extracted image diagram and the selected recommended tree crown image diagram, or by matching the pattern, the matching portion and the mismatching portion are detected, and the sub-branch at the mismatching portion is pruned. Can be identified.

  A pruning instruction image figure can be obtained by adding markings or the like to the extracted image figure for the pruning location confirmed by the extracted image figure or specified by image comparison.

  Hereinafter, based on FIG. 2 thru | or FIG. 9, the specific means for implement | achieving the cultivation management method shown in FIG. 1 is demonstrated. As shown in FIG. 2, a small unmanned aerial vehicle (drone) 11 that is capable of flying while turning at a constant altitude by radio control from the ground is used for photographing from above the grape cultivation region 10. A small camera 12 is mounted on the drone 11, and as shown in FIG. 3, an area E that partitions the sky above the grape cultivation region 10 in a grid pattern in the X-axis direction and the Y-axis direction. Take a picture of the crown of each vine. Each of the areas E can be easily identified by setting the intervals between the ledges of the grape shelves installed in the vines 13 or the intervals between the wire meshes assembled in a lattice shape. An overall bird's-eye view image as shown in FIG. 3 is obtained. The flight route and altitude of the drone 11 are determined so that photographing can be performed according to the size of each area E, and the mounted camera 12 is aimed. The flight of the drone 11 can be performed manually by the operator via the controller or while automatically maintaining the flight route and altitude by GPS.

  FIG. 4 is an enlarged view of a part of the overhead image shown in FIG. In this image view, the crown form of a plurality of main branches B and sub-main branches C extending in multiple directions from the main trunk A of the vine 13 is reflected in a planar manner, but at the same time grows on the ground and this ground. The background of grass and scattered leaves is also reflected. When such a background is reflected, the form of the vine 13 becomes difficult to understand, and the background portion excluding the form of the branch of the vine 13 is converted into a single color by image processing. For example, as shown in FIG. 5, by converting the vine 13 into white and black binary data in which the background is black and the other background is black, the background is removed from the grass or the like that grows on the soil or the ground. An extracted image diagram in which only the forms of the main branch B and the sub-main branch C of the tree 13 are extracted can be obtained. For this reason, it is not necessary to cut the grass directly under the vine shelf immediately before photographing with the drone 11 or to spread a sheet of a single color different from the color of the vines 13.

  6, 7 </ b> A, and 7 </ b> B are schematic views illustrating a part of FIG. 5. In the first method, the vine 13 shown in FIG. 6 is divided into a plurality of pruning management areas E11 to E14, and the branching of the main branch B and the sub-main branch C is not confirmed for each management area. Identify pruning points from

  Moreover, in the second method, the extracted image diagram shown in FIG. 7A and the recommended tree image diagram shown in FIG. 7B are superimposed on the monitor, and the sub-main branches that do not appear in the recommended tree image diagram are pruned. As specified. Here, the sub-main branch C3 of the main branch B1, the sub-main branch C2 of the main branch B2, and the sub-main branch C2 of the main branch B3 shown in FIG. 7A are specified as pruning targets. This pruning location can be automatically detected by pattern matching the extracted image map and the recommended tree crown image map on a monitor.

  In the second method, as shown in FIG. 7A, the pruning point is specified for an area composed of a plurality of main branches B and a plurality of sub-main branches C that extend around one main trunk A of the vine. For example, by paying attention to a predetermined main branch in the extracted image diagram shown in FIG. 7A, it is also possible to specify a sub-main branch to be pruned based on the interval between a plurality of sub-main branches branching from the main branch.

  By connecting the images for each pruning management area where the pruning specific operation has been completed, it is possible to form a predicted bird's-eye view image after pruning as shown in FIG. When actually pruning a vine, it moves to the pruning management area corresponding to the coordinate axis shown in the overhead image diagram, and the sub-branch with the pruning mark etc. shown in the extracted image diagram Can be done by checking.

  As for the display of the portion to be pruned, if it can be discriminated on the monitor, such as adding an X mark or an enclosure mark to the sub-main branch specified as the pruning target, or performing color-coded or subtractive color display. Any display method may be used.

  In addition to the identification of the pruning location, it is possible to manage the new shoots that sprout on one vine. As shown in FIG. 9 which schematically shows the extracted image diagram of FIG. 5, this shoot management is divided into a plurality of shoot management areas E21, E22, and E23, and each shoot management area is divided. Is performed by detecting the distribution and number of shoots F. Like the mother branch D, the new treetop F appears as a minute protrusion on the surface of the main branch B, but has a different color. As a result, it is possible to discriminate by detecting a color difference between the main branch and the result.

  Each shoot management area is divided around the main trunk A of one vine 13, and the distribution state or number of the shoot F detected by the color discrimination is calculated for each divided shoot management area. The distribution state or the number of appropriate shoots F set for each divided shoot management area is compared. And based on this comparison result, the distribution location or number of shoots to be left is selected. As a result, it is possible to perform a pinching simulation or the like in each of the divided shoot management areas, and to predict the growth of subsequent fruit bunches.

  In the conventional dormancy vine management method, the skilled grower looks up from the bottom of the vine canopy so that the position and direction of the sub-branches branch from the main branch and between adjacent sub-branches. While confirming the mother branch as a result, the pruning and other work was done with the experience and intuition. However, since the work must be performed locally, it was not easy to prun in consideration of the balance of the whole vine. On the other hand, in the grape cultivation management method of the present invention, the whole image of the canopy of the vine is first grasped, and image processing is performed on a predetermined area from the first image, thereby pruning more accurately and easily. It is possible to detect the place to be. In addition, since it can be performed by image collation between the extracted image figure extracted from the overhead image figure and the recommended tree image figure appropriately pruned, it is possible to obtain a certain unified judgment, and beginners Even so, grape cultivation can be managed properly without major failure. Furthermore, by creating a database of bird's-eye view images of vines aerial photographed by drones and image-processed image maps for each area, etc. General technical guidance can be provided.

  In this invention, the whole image of the grape crown can be easily grasped by photographing the whole grape cultivation area in the air using a drone. After obtaining the bird's-eye view image by connecting the images for each area photographed by the drone, by performing image processing to extract the form of the main trunk, the main branch, and the sub-main branch from this bird's-eye view image, the result mother The position and number of branches and shoots can be detected easily and accurately. In addition, it is possible to automatically detect the location and number of sub-main branches, result mother branches, and shoots in a predetermined area from the image processed image. Further, the number of detected new treetops is calculated for each area of vines, and the number of new treetops set for each area and the distribution state thereof are compared in advance, whereby the fruit clusters attached to the new treetops Therefore, it is possible to carry out planned grape cultivation management, thereby contributing to the improvement of grape quality.

  In the present embodiment, the cultivation of grapes has been described as an example. However, the present invention can be applied to other fruit tree cultivation such as peaches accompanied by management and work such as pruning and new treetops.

A main trunk B main branch C sub-main branch D result mother branch E area F new treetop 10 cultivation area 11 drone (unmanned aerial vehicle)
12 Cameras 13 Grapes

Claims (8)

Photographed while sequentially scanning a predetermined cultivation area consisting of one or more vines,
Form an overhead image diagram of the entire cultivation area from the captured image,
Based on this bird's-eye view image diagram, form an extracted image diagram that extracts the main branch and sub-main branch crown shape centered on each main trunk of the vine,
This extracted image diagram is divided into a plurality of pruning management areas,
A grape cultivation management method, wherein a branch to be pruned is specified for each pruning management area. Photographed while sequentially scanning a predetermined cultivation area consisting of one or more vines,
Form an overhead image diagram of the entire cultivation area from the captured image,
Based on this bird's-eye view image diagram, form an extracted image diagram that extracts the main branch and sub-main branch crown shape centered on each main trunk of the vine,
A grape cultivation management method characterized by identifying a branch to be pruned by comparing the extracted image diagram with a recommended crown image diagram having a preset recommended crown shape. Photographed while sequentially scanning a predetermined cultivation area consisting of one or more vines,
Form an overhead image diagram of the entire cultivation area from the captured image,
Based on this bird's-eye view image diagram, form an extracted image diagram that extracts the main branch and sub-main branch crown shape centered on each main trunk of the vine,
A grape cultivation management method characterized by detecting the distribution or number of new shoots by color discrimination from the extracted image map.   The grape cultivation management method according to claim 3, wherein the detection of the distribution or number of shoots by the color discrimination is performed based on a color difference between the shoots and a result mother branch.   The grape cultivation management method according to claim 2, wherein each of the extracted image figure and the recommended crown image figure has a pattern recognition of each crown shape, and a pruned portion is displayed on the extracted image figure. The extracted image diagram is divided into a plurality of shoot management areas,
The distribution or number of shoots detected by color discrimination for each divided shoot management area is compared with the distribution or number of appropriate shoots set in advance for each of the above-mentioned shoot management areas, and is left. The method for cultivating and managing grapes according to claim 3, wherein the distribution or number of shoots is selected.   The grape cultivation management method according to claim 1 or 2, wherein the photographing is performed along a plurality of grid-like areas that divide the predetermined cultivation area in the X-axis direction and the Y-axis direction.   The method for cultivating grapes according to claim 1 or 2, wherein the photographing is performed by an unmanned aerial vehicle equipped with a camera and flying while turning over the cultivation area by remote control.

Images