vehicle

Abstract

To provide a vehicle capable of efficiently spraying an agricultural chemical.SOLUTION: Since a vehicle for spraying an agricultural chemical to an object to be controlled is provided with a photographing device for photographing the object to be controlled around the vehicle, a control device for determining whether or not spraying the agricultural chemical is necessary, and a spraying device for spraying the agricultural chemical, estimation and control of a pest can be performed at the same time. Further, since the area from the lowermost end to the uppermost end in the vertical direction of the object to be controlled around the vehicle is photographed by the photographing device, even when the agricultural chemical is sprayed to a tall object to be controlled such as a fruit tree, estimation and control of a pest can be performed at the same time by the vehicle. Thus, the agrochemical can be efficiently sprayed.SELECTED DRAWING: Figure 2

Description

[Technical Field]

[0001] The present invention relates to a vehicle for spraying pesticides on a pest control target. [Background technology]

[0002] Systems for spraying pesticides on control targets are well known. For example, the agricultural support system described in Patent Document 1 is one such system. Patent Document 1 discloses an agricultural support system equipped with a camera, a control device, and a spraying device. The camera in Patent Document 1 is a device mounted on an unmanned aerial vehicle such as a multicopter and photographs the inside of a farm. The control device in Patent Document 1 calculates the growth status of crops in the farm using images captured by the camera, estimates the occurrence of pests and diseases based on the calculation results, and creates an operation plan for spraying pesticides based on the estimation results. The spraying device in Patent Document 1 is a device that sprays pesticides on farms where the occurrence of pests and diseases is suspected. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Patent Publication No. 2021-106554 Summary of the Invention [Problem to be solved by the invention]

[0004] A system for spraying pesticides on pest control targets must have functions such as photographing the target, determining whether or not to spray pesticides, and spraying the pesticides. In the agricultural support system described in Patent Document 1, each device that realizes the above-mentioned functions is installed separately. In such an agricultural support system, after photographing the entire target farm, a determination is made as to whether or not to spray pesticides for each area within the farm, and pesticides are sprayed in the necessary areas based on the results of the determination. There is room for improvement in pesticide spraying. Additionally, if the target farm is a tall tree such as a fruit tree, setting the vertical position for photographing uniformly may result in inefficient pesticide spraying.

[0005] The present invention has been made in light of the above circumstances, and an object of the present invention is to provide a vehicle that can efficiently spray pesticides. [Means for solving the problem]

[0006] The gist of the first invention is (a) a vehicle for spraying pesticides on a pest control target, comprising: (b) a photographing device for photographing the pest control target around the vehicle; (c) a control device for determining whether or not spraying of the pesticide is necessary based on an image of the pest control target photographed; and (d) a spraying device for spraying the pesticide on the pest control target for which it has been determined that spraying of the pesticide is necessary; and (e) the photographing device photographs the pest control target around the vehicle from the lowest point to the highest point in the vertical direction. (f) when spraying the pesticide, the vehicle travels at an upper limit of the lower of the maximum vehicle speed within a predetermined vehicle speed range at which the image used to determine whether or not the pesticide needs to be sprayed is clearly displayed, and the maximum vehicle speed within a predetermined vehicle speed range at which the spraying device can appropriately spray the pesticide. The reason is that.

[0007] Another invention is that in the vehicle described in the first invention, the control device determines whether or not spraying of the pesticide is necessary by applying the image data used to determine whether or not spraying of the pesticide is necessary to a predetermined learning model that shows the relationship between the image data and whether or not spraying of the pesticide is necessary, which is realized by supervised learning using machine learning. [Effects of the Invention]

[0008] According to the first aspect of the present invention, a vehicle that sprays pesticides on control objects is equipped with a camera that photographs the control objects around the vehicle, a control device that determines whether pesticide spraying is necessary, and a sprayer that sprays the pesticide, so that pest estimation and pest control can be performed simultaneously. Furthermore, since the camera captures images of the control objects around the vehicle from the lowest to the highest point in the vertical direction, the vehicle can simultaneously estimate pests and pest control, even when spraying pesticides on tall control objects such as fruit trees. Therefore, pesticide spraying can be performed efficiently.

[0009] According to the other invention, the image data used to determine whether or not pesticide spraying is necessary is applied to a predetermined learning model that shows the relationship between image data and whether or not pesticide spraying is necessary, which is realized by supervised learning using machine learning, thereby determining whether or not pesticide spraying is necessary. This makes it possible to determine whether or not pesticide spraying is necessary based on differences in appearance, similar to artificial determination methods. [Brief explanation of the drawings]

[0010] [Figure 1] 1 is a diagram illustrating a schematic configuration of a vehicle to which the present invention is applied, and is a diagram illustrating the appearance of the vehicle during work. [Figure 2] 1 is a diagram illustrating a schematic configuration of a vehicle to which the present invention is applied, and is a diagram illustrating various control functions and main parts of a control system in the vehicle. [Figure 3] FIG. 10 is a diagram illustrating an example of a color difference in a predetermined range of a target image. [Figure 4] FIG. 1( a ) is a diagram showing an example of a target image in which a predetermined area where pesticide spraying is required has been identified, and FIG. 1( b ) is a diagram showing an example of a corrected target image. [Figure 5] FIG. 10 is a diagram illustrating an example of a learning model. [Figure 6] 1 is a flowchart illustrating the main control operations of an electronic control device, and is a flowchart illustrating the control operations for efficiently spraying pesticides. DETAILED DESCRIPTION OF THE INVENTION

[0011] Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings. [Example]

[0012] 1 and 2 are diagrams illustrating the schematic configuration of a vehicle 10 to which the present invention is applied. Fig. 1 is a diagram illustrating the appearance of the vehicle 10 during work. Fig. 2 is a diagram illustrating various control functions and main parts of the control system in the vehicle 10.

[0013] The vehicle 10 is, for example, a known electric vehicle capable of autonomous driving. The vehicle 10 travels autonomously and unmanned along a predetermined route (see dashed line A in FIG. 1 ) within a farm 100. The vehicle 10 is a vehicle that sprays a pesticide 20 on a control target 102 within the farm 100. The farm 100 is a farm such as an orchard. The control target 102 is agricultural produce such as fruit trees grown within the farm 100. The agricultural produce includes, for example, fruits as well as leaves. The control target 102 is grown, for example, in rows lined up at predetermined intervals. The control target 102 is agricultural produce that is the target of pest prevention and extermination. Pests are diseases and harmful insects of agricultural produce. The pesticide 20 is a known agricultural chemical. The vehicle 10 may be an engine vehicle, may be a manned vehicle that travels autonomously, or may be a vehicle that can be manually driven.

[0014] The vehicle 10 is equipped with a photographing device 12, an electronic control device 14, a spraying device 16, and the like.

[0015] The photographing device 12 is installed, for example, at the front of the vehicle 10 in the forward / backward direction. The photographing device 12 is, for example, a monocular camera that photographs the area around the vehicle 10. The photographing device 12 photographs the area around the vehicle 10, particularly the control target 102 in front of and to the sides. The photographing device 12 outputs information on the target image PICs to the electronic control device 14 as image information Ipic. The target image PICs is the image PIC of the control target 102 photographed when spraying the pesticide 20, and is the image PIC used to determine whether or not to spray the pesticide 20 (see FIG. 4(a) described below).

[0016] The spraying device 16 is installed, for example, at the rear of the vehicle 10 in the forward / rearward traveling direction. The spraying device 16 includes, for example, a tank 22 for storing the pesticide 20 and a sprayer 24 for spraying the pesticide 20. The spraying device 16 sprays the pesticide 20 onto the control target 102.

[0017] The electronic control unit 14 is a control device, or controller, for the vehicle 10 that is involved in controlling the running of the vehicle 10, the photographing device 12, the spraying device 16, etc. The electronic control unit 14 is configured to include a so-called microcomputer equipped with, for example, a CPU, RAM, ROM, an input / output interface, etc. The CPU executes various controls of the vehicle 10 by performing signal processing according to programs stored in advance in the ROM while utilizing the temporary storage function of the RAM.

[0018] The electronic control device 14 is supplied with image information Ipic from the photographing device 12. The electronic control device 14 is also supplied with various signals (e.g., vehicle speed V) based on detection values from various sensors (e.g., vehicle speed sensor 26) provided on the vehicle 10. The electronic control device 14 outputs a photographing command signal Spic to the photographing device 12 for photographing the pest control target 102 around the vehicle 10. The electronic control device 14 also outputs a spraying command signal Sspr to the spraying device 16 for spraying the pesticide 20 on the pest control target 102. The electronic control device 14 also outputs control command signals for automatic driving to each device (not shown) provided on the vehicle 10.

[0019] The photographing device 12 photographs the pest control object 102 around the vehicle 10 based on a photographing command signal Spic from the electronic control device 14, and outputs image information Ipic including information on the target image PICs to the electronic control device 14. The electronic control device 14 determines whether or not spraying of the pesticide 20 is necessary based on the target image PICs included in the acquired image information Ipic. If the electronic control device 14 determines that spraying of the pesticide 20 is necessary, it outputs a spraying command signal Sspr to the spraying device 16 to spray the pesticide 20 on the pest control object 102 corresponding to the target image PICs included in the acquired image information Ipic. The spraying device 16 sprays the pesticide 20 on the pest control object 102 determined to require spraying of the pesticide 20 based on the spraying command signal Sspr from the electronic control device 14.

[0020] However, if the control target 102 is a tall object such as a fruit tree, if the vertical position photographed by the photographing device 12 is set to a uniform height, there is a risk that the pesticide 20 cannot be sprayed efficiently.

[0021] Therefore, the photographing device 12 is equipped with a detection sensor 28, such as a lidar or infrared sensor, that detects the pest control object 102 around the vehicle 10. When the vehicle 10 is traveling while spraying the pesticide 20, the photographing device 12 uses the detection sensor 28 to detect the vertical length or height of the pest control object 102 around the vehicle 10. For example, when the vehicle 10 is traveling while spraying the pesticide 20, the photographing device 12 uses the detection sensor 28 to detect the position of the pest control object 102. Then, when the vehicle 10 is traveling while spraying the pesticide 20, the photographing device 12 photographs the pest control object 102 from the bottom to the top in the vertical direction.

[0022] The higher the vehicle speed V while the vehicle 10 is traveling, the more likely it is that the target image PICs will become unclear, making it difficult to determine whether or not the pesticide 20 needs to be sprayed. When spraying the pesticide 20, the vehicle 10 preferably travels within a predetermined vehicle speed range Vpic at which the target image PICs captured by the photographing device 12 is clearly captured. In particular, to increase the efficiency of spraying the pesticide 20, it is preferable to travel at the maximum vehicle speed Vpicmax within the vehicle speed range Vpic.

[0023] On the other hand, the higher the vehicle speed V while the vehicle 10 is traveling, the more difficult it is to spray the pesticide 20 onto the pest control target 102. When spraying the pesticide 20, the vehicle 10 preferably travels within a predetermined vehicle speed range Vspr that allows the spraying device 16 to appropriately spray the pesticide 20. In particular, to increase the efficiency of spraying the pesticide 20, it is preferable to travel at the maximum vehicle speed Vsprmax within the vehicle speed range Vspr.

[0024] In order to determine whether or not the pesticide 20 needs to be sprayed and to spray the pesticide 20 with maximum efficiency within a vehicle speed range in which the pesticide 20 can be appropriately sprayed, the lower of the maximum vehicle speeds Vpicmax and Vsprmax may be selected. That is, when spraying the pesticide 20, the vehicle 10 travels with an upper limit of the lower of the maximum vehicle speeds Vpicmax and Vsprmax (see FIG. 1). For example, the electronic control unit 14 outputs a control command signal to each device (not shown) provided in the vehicle 10 for automatic driving with the vehicle speed Vmin as the upper limit.

[0025] The determination of whether or not the pesticide 20 needs to be sprayed by the electronic control device 14 will now be described in detail. For example, the electronic control device 14 performs predetermined image processing on the target image PICs to calculate a color difference Dt between the color of the control object 102 (especially the leaves) within the predetermined range AR and the normal color. The normal color may be, for example, a predetermined color, the average color of the control object 102 within the farm 100, or the average color of the control object 102 outside the predetermined range AR of the target image PICs. The electronic control device 14 determines whether or not the pesticide 20 needs to be sprayed based on whether or not the color difference Dt within the predetermined range AR of the target image PICs is equal to or greater than the necessity determination value THn. Determining whether or not the pesticide 20 needs to be sprayed based on whether or not the color difference Dt is equal to or greater than the necessity determination value THn is equivalent to estimating whether or not a pest has occurred based on whether or not the color difference Dt is equal to or greater than the necessity determination value THn. The necessity determination value THn is a predetermined pesticide spray necessity determination value for determining whether the color difference Dt is large enough to require the spraying of the pesticide 20. When the electronic control device 14 determines that the color difference Dt in the predetermined range AR is equal to or greater than the necessity determination value THn, it identifies the predetermined range AR as one in which the spraying of the pesticide 20 is required.

[0026] Fig. 3 is a diagram showing an example of the color difference Dt in a predetermined range AR of the target image PICs. Fig. 4(a) is a diagram showing an example of the target image PICs in which a predetermined range AR where spraying of the pesticide 20 is required is identified. In Fig. 3, the color difference Dt in the third predetermined range AR3 is set to be equal to or greater than the necessity determination value THn. As shown in Fig. 4(a), the third predetermined range AR3 is identified as the range where spraying of the pesticide 20 is required.

[0027] The situation ST when the control target 102 is photographed by the photographing device 12 is not the same. There are multiple types of situation ST at the time of photographing, such as the brightness around the vehicle 10, the type of control target 102, the growth status of the control target 102, the health status of the control target 102, and the location of the farm 100. The brightness around the vehicle 10 is brightness that varies depending on, for example, the season, time of day, weather conditions, shadows, etc.

[0028] The electronic control unit 14 stores a predetermined reference image PICb as the image PIC when the control target 102 is photographed by the photographing device 12 in a predetermined situation STf. The electronic control unit 14 corrects the target image PICs based on the difference between the predetermined situation STf when the control target 102, which is the basis for the predetermined reference image PICb, was photographed and the situation STs when the control target 102, which is the basis for the target image PICs, was photographed. For example, the electronic control unit 14 corrects the target image PICs based on the difference between the same type of situation STs and the predetermined situation STf using a predetermined image correction process. In this embodiment, the corrected target image PICs is referred to as the corrected target image PICsc (see FIG. 4(b)). The electronic control unit 14 determines whether or not spraying of the pesticide 20 is necessary based on the corrected target image PICsc. For example, the electronic control unit 14 determines whether or not spraying of the pesticide 20 is necessary based on whether or not the color difference Dt within a predetermined range AR of the corrected target image PICsc is equal to or greater than the necessity determination value THn. In this case, the electronic control unit 14 only needs to store the predetermined condition STf necessary for correcting the target image PICs, and does not need to store the predetermined reference image PICb. Alternatively, the electronic control unit 14 may determine whether or not spraying of the pesticide 20 is necessary based on the difference between the predetermined reference image PICb and the corrected target image PICsc. For example, the electronic control unit 14 may determine whether or not spraying of the pesticide 20 is necessary based on whether the color difference Dt between the color in the predetermined reference image PICb and the color in the corrected target image PICsc is equal to or greater than the necessity determination value THn.

[0029] The electronic control device 14 may use, for example, a learning model 30 based on machine learning to determine whether or not it is necessary to spray the pesticide 20. For example, the electronic control device 14 determines whether or not it is necessary to spray the pesticide 20 by applying the target images PICs to the learning model 30. The learning model 30 is a pre-defined trained model that indicates the relationship between the image PIC data and whether or not it is necessary to spray the pesticide 20. The learning model 30 is realized by supervised learning based on machine learning, using the image PIC data and whether or not it is necessary to spray the pesticide 20 as training data. Whether or not it is necessary to spray the pesticide 20 may be determined by whether or not there is an outbreak of pests or diseases.

[0030] FIG. 5 is a diagram illustrating an example of a learning model 30. In FIG. 5, the learning model 30 is a neural network based on image data PIC and the situation ST at the time the target object 102, which is the basis of the image PIC, was photographed. The learning model 30 can be configured by modeling a group of living neurons using software (a computer program) or hardware consisting of a combination of electronic elements. The learning model 30 has a multi-layer structure consisting of an input layer consisting of i neuronal elements (neurons) Pi1, a middle layer consisting of j neuronal elements Pj2, and an output layer consisting of k neuronal elements Pk3. The middle layer may also have a multi-layer structure. Furthermore, the learning model 30 includes a transfer element Dij having a weighting value Wij and a transfer element Djk having a weighting value Wjk to transmit the state of the neuronal elements from the input layer to the output layer. The transfer element Dij is a transfer element that connects the i neuron elements Pi1 and the j neuron elements Pj2, respectively. The transfer element Djk is a transfer element that connects the j neuron elements Pj2 and the k neuron elements Pk3, respectively.

[0031] The learning model 30 is a necessity determination system that determines whether or not the pesticide 20 needs to be sprayed. This necessity determination system also serves as a pest estimation system. In the learning model 30, the weighting values Wij and Wjk are machine-learned using a predetermined algorithm. In supervised learning in the learning model 30, training data, i.e., training signals, identified in the vehicle 10 are used. In the learning model 30 of FIG. 5, for example, data on the image PIC and data on the situation ST when the pest control target 102 was photographed are provided as training signals for the input layer. In the learning model 30 of FIG. 5, the result of the determination of whether or not the pesticide 20 needs to be sprayed, i.e., the result of the estimation of the presence or absence of pests, is provided as training signals for the output layer. The electronic control device 14 determines whether or not the pesticide 20 needs to be sprayed by applying the target images PICs and the situation STs when the pest control target 102, which is the basis of the target images PICs, to the learning model 30 of FIG. 5. When the learning model 30 in FIG. 5 is used, the situation ST when the control target 102 is photographed is also learned, so there is no need to correct the target images PICs.

[0032] The learning model 30 may be a pre-trained model that indicates the relationship between the image PIC data, the necessity of spraying the pesticide 20, the type of pests and diseases that have occurred in the control target 102, and the extent of damage caused by the pests and diseases. In other words, the learning model 30 may further indicate the relationship between the image PIC data and the type of pests and diseases that have occurred in the control target 102 and the extent of damage caused by the pests and diseases. In this case, the teacher signal for the output layer composed of k neuron elements Pk3 in FIG. 5 includes the determination result of the necessity of spraying the pesticide 20, as well as the determination result of the type of pests and diseases that have occurred in the control target 102 and the extent of damage caused by the pests and diseases. The type of pests and diseases is, for example, the name of a disease such as "downy mildew." The extent of damage caused by the pests and diseases is, for example, "small," "medium," or "large." The determination result of the type of pests and the extent of damage caused by the pests and diseases is provided, for example, in the output layer where it is determined that the pesticide 20 is required to be sprayed. The electronic control device 14 applies the target images PICs to the learning model 30 to determine whether spraying of the pesticide 20 is necessary, as well as to determine the type of pest and the extent of damage caused by the pest. The electronic control device 14 determines the type and required amount of the pesticide 20 based on the determined type of pest and the extent of damage caused by the pest. The type of pesticide 20 is, for example, Bordeaux mixture, which corresponds to "downy mildew." The required amount of the pesticide 20 is, for example, a dilution ratio and a spray rate. The spray command signal Sspr output by the electronic control device 14 includes not only spraying the pesticide 20 on the control target 102, but also the type and required amount of the pesticide 20. Based on the spray command signal Sspr from the electronic control device 14, the spraying device 16 sprays the required amount of the pesticide 20 corresponding to the pest and the pest on the control target 102 for which spraying of the pesticide 20 is determined to be necessary.

[0033] The electronic control device 14 may use the learning model 30 to estimate trends such as the timing and location of pest and disease occurrence (see the area surrounded by the dashed line B in FIG. 1). The electronic control device 14 stores the estimated trend results as data. The electronic control device 14 may use the estimated results to correct, for example, the amount of pesticide 20 to be sprayed the next time the pesticide 20 is sprayed. This makes it possible to grasp the trends of pests and diseases occurring on the control target 102. It is also possible to add shading to the amount of pesticide 20 to be sprayed.

[0034] FIG. 6 is a flowchart illustrating the main control operations of the electronic control unit 14, which are performed to efficiently spray the pesticide 20, and are executed repeatedly, for example, while the vehicle is traveling.

[0035] In Figure 6, in step (hereinafter, step will be omitted) S10, target images PICs included in the image information Ipic from the photographing device 12 are acquired. Next, in S20, it is determined whether or not spraying of the pesticide 20 is necessary based on the target images PICs. If the determination in S20 is negative, this routine is terminated. If the determination in S20 is positive, in S30, a spray command signal Sspr is output to the spraying device 16 to spray the pesticide 20 on the pest control target 102 corresponding to the acquired target images PICs.

[0036] As described above, according to this embodiment, the vehicle 10 is equipped with the camera 12, the electronic control device 14, and the spraying device 16, so that it is possible to simultaneously estimate and control pests. Because the vehicle 10 is not an air vehicle, energy consumption associated with an increase in the vehicle 10's payload is reduced, for example. Furthermore, because the camera 12 captures images of the control target 102 from the bottom to the top in the vertical direction, the vehicle 10 can simultaneously estimate and control pests, even when spraying pesticide 20 on tall control targets 102 such as fruit trees. In this case, it is possible to find pests that occur at low positions in the tall control targets 102, for example. Therefore, the pesticide 20 can be sprayed efficiently.

[0037] Furthermore, according to this embodiment, when spraying the pesticide 20, the vehicle 10 is driven with an upper limit of the lower of the maximum vehicle speeds Vpicmax and Vsprmax, Vmin. This allows for the most efficient spraying of the pesticide 20 and the most accurate target image PICs, which allows for determining whether or not the pesticide 20 needs to be sprayed. Because the vehicle 10 is not an aircraft, energy efficiency is improved even when working at low speeds, for example.

[0038] Furthermore, according to this embodiment, the electronic control unit 14 corrects the target image PICs based on the difference between the predetermined situation STf and the situation STs when the control target 102, which is the basis of the target image PICs, was photographed. Furthermore, the electronic control unit 14 determines whether or not spraying of the pesticide 20 is necessary based on the corrected target image PICsc. This makes it possible to obtain a target image PICs that is less affected by the situation STs, such as the brightness around the vehicle 10 and the type of control target 102, and thereby allows for appropriate estimation and control of pests.

[0039] Furthermore, according to this embodiment, the electronic control device 14 applies the target images PICs to the learning model 30 to determine whether or not spraying of the pesticide 20 is necessary, so that it is possible to determine whether or not spraying of the pesticide 20 is necessary based on differences in appearance, which is similar to an artificial judgment method.

[0040] Furthermore, according to this embodiment, the electronic control device 14 applies the target images PICs to the learning model 30 to determine the type of pest or disease and the extent of damage caused by the pest, and then determines the type and required amount of pesticide 20 based on the type of pest or disease and the extent of damage caused by the pest or disease. This allows the type of pesticide and dilution ratio that is effective against the pest or disease to be selected. Furthermore, if the type of pest or disease does not match the type of pesticide 20 loaded on the vehicle 10, for example, spraying of the pesticide 20 can be stopped and the amount of pesticide 20 used can be reduced.

[0041] Although the embodiments of the present invention have been described in detail above with reference to the drawings, the present invention can also be applied to other embodiments.

[0042] For example, in the above-described embodiment, the color difference Dt is used to determine whether or not spraying of the pesticide 20 is necessary, but this is not limited to this. For example, the growth status or health condition of the control target 102 may be used to determine whether or not spraying of the pesticide 20 is necessary. In this way, various methods are possible for determining whether or not spraying of the pesticide 20 is necessary.

[0043] Furthermore, in the above-described embodiment, the photographing device 12 is equipped with the detection sensor 28 and detects the vertical length or height of the pest control target 102, but this is not limited to this embodiment. For example, the photographing device 12 may process the image PIC to recognize the photographing range and photograph the vertical length or height of the pest control target 102 from the bottom to the top. In other words, the photographing device 12 does not necessarily need to detect the vertical length or height of the pest control target 102 around the vehicle 10. In this case, the photographing device 12 does not need to be equipped with the detection sensor 28. In short, it is sufficient for the photographing device 12 to be able to photograph at least the vertical length or height of the pest control target 102 from the bottom to the top while the vehicle 10 is traveling when spraying the pesticide 20.

[0044] Furthermore, in the above-described embodiment, the spraying device 16 is installed at the rear of the vehicle 10, but this is not limiting. For example, the spraying device 16 may be installed on a dolly connected to the rear of the vehicle 10 body in the forward and backward traveling direction. This dolly is pulled by the vehicle 10 body and travels integrally with the vehicle 10 body, and constitutes a part of the vehicle 10. Therefore, the spraying device 16 installed on this dolly is provided on the vehicle 10.

[0045] Furthermore, in the above-described embodiment, the installation of the photographing device 12 is not limited to the front of the vehicle 10, and the installation of the spraying device 16 is not limited to the rear of the vehicle 10. In short, it is sufficient that the configuration is such that the pesticide 20 can be sprayed on the control target 102 that is determined to require spraying of the pesticide 20.

[0046] Furthermore, in the above-described embodiment, if the learning model 30 is a predetermined trained model that shows the relationship between the data of the image PIC and whether or not to spray the pesticide 20, the teacher signal for the input layer in Figure 5 does not need to include data on the situation ST.

[0047] It should be noted that the above is merely one embodiment, and the present invention can be embodied in various forms with various modifications and improvements based on the knowledge of those skilled in the art. [Explanation of symbols]

[0048] 10: Vehicle 12: Camera device 14: Electronic control device (control device) 16: Spraying device 20: Pesticide 30: Learning model 102: Control target object PIC: Image PICb: Predetermined reference image PICs: Target image (image used to determine whether or not pesticide spraying is necessary) PICsc: Corrected target image (corrected image)

Claims

1. A vehicle that sprays pesticides on pest control targets, an imaging device that images the control target around the vehicle; a control device that determines whether or not spraying of the pesticide is necessary based on an image of the pest control target; A spraying device that sprays the pesticide to the control object that is determined to require spraying of the pesticide; and The photographing device photographs the control target around the vehicle from the lowest point to the highest point in the vertical direction, When spraying the pesticide, the vehicle is characterized in that it travels at an upper limit of the lower of the maximum vehicle speed within a predetermined vehicle speed range at which the image used to determine whether or not the pesticide needs to be sprayed is clearly displayed, and the maximum vehicle speed within a predetermined vehicle speed range at which the spraying device can properly spray the pesticide.

2. The vehicle described in claim 1, characterized in that the control device corrects the image based on a specified situation when the object to be controlled, which serves as the basis for a specified reference image, was photographed and a situation when the object to be controlled, which serves as the basis for the image used to determine whether or not the pesticide needs to be sprayed, was photographed, and determines whether or not the pesticide needs to be sprayed based on the corrected image.

3. The vehicle described in claim 1 or 2, characterized in that the control device determines whether or not spraying of the pesticide is necessary by applying the image data used to determine whether or not spraying of the pesticide is necessary to a predetermined learning model that shows the relationship between the image data and whether or not spraying of the pesticide is necessary, which is realized by supervised learning using machine learning.

4. The learning model further indicates a relationship between the image data and the type of pest that has occurred in the control target and the progress of damage caused by the pest, The vehicle described in claim 3, characterized in that the control device determines the type of pest and the extent of the damage by applying the image data used to determine whether or not the pesticide needs to be sprayed to the learning model, and determines the type and required amount of the pesticide based on the type of pest and the extent of the damage.

Images