WO2018235486A1 - Harvester - Google Patents

Abstract

This combine harvester is provided with: a threshing depth adjustment mechanism; a threshing depth control unit 620 which performs, using the threshing depth adjustment mechanism, threshing depth adjustment control on the basis of the length of reaped grain culms; a machine body position calculation unit 66 which calculates the machine body position; an image recognition module 5 which deduces a weed-growth region in a photographic image acquired by an photographing unit 70 and which outputs recognition output data representing the deduced weed-growth region; a weed position information generation unit 51 which generates, from the recognition output data and the machine body position when the photographic image was acquired, weed position information indicating the position of the weed-growth region on a map; and a work travel control unit for calculating the timing when weeds pass through the threshing depth adjustment mechanism and for executing weed ingress control while weeds are passing the threshing depth adjustment mechanism.

Description

Harvester

TECHNICAL FIELD The present invention relates to a harvester that harvests a grained paddy while traveling on a field.

Conventional harvesters are equipped with various functions to improve the performance of work or to improve the efficiency of work.
For example, as shown in Patent Document 1, in the conventional combine, a threshing depth detection sensor for detecting the tip position of the transport grain fed from the reaper to the threshing device is provided, and the detection information of the threshing depth detection sensor The threshing depth adjusting means for controlling the threshing depth in the threshing device is controlled based on the threshing device to thresh the transported grain at the target threshing depth.

In addition, in the field of crop harvesting work using a harvester in the field, it is necessary to pay attention to the locally different growth status of crops and the presence of obstacles including people. For this reason, for example, in the combine according to Patent Document 2, a television camera and an image processing apparatus for capturing a grain scale in front of the reaper are provided. The image processing apparatus detects a specific state of the grain scale by comparing the image from the television camera with an image indicating the establishment state of various grain scale stored in advance. For example, when the image processing apparatus detects that a part of the grain scale in front of the reaper is fallen, the scratching reel is tilted with the grain fall side down. As a result, the harvesting performance of the fallen cereal straw is improved. In the combine according to Patent Document 3 in which a television camera and an image processing apparatus are provided as in the combine according to Patent Document 2, the combine is a headland when it is detected that no grain is present in front of the reaper. The power transmission to the cutter is interrupted, the cutter is raised, and the traveling speed is reduced. As a result, the collision of the cutter with the wedge or the like during the turning on the headland is avoided, and the turning of the headland of the combine is smoothly performed.

Furthermore, in the combine according to Patent Document 4, a camera for photographing the rear of the machine body is mounted, and from the photographed image, a stock row of cereal straw cut by the reaper is recognized. Based on the difference between the recognized stock row and the aircraft, the orientation control of the vehicle is performed such that each branch passes between the strain rows and the stock rows.

Unexamined-Japanese-Patent No. 08-172867 gazette Japanese Patent Application Laid-Open No. 11-155340 Japanese Patent Application Laid-Open No. 11-137062 Unexamined-Japanese-Patent No. 2006-121952

The technology of threshing with appropriate threshing depth contributes to the improvement of threshing performance by adjusting the threshing depth using the threshing depth detection sensor when transporting the reaping cropped rice straw reaped by the reaper to the threshing device . However, depending on the field, there may be a mixture of weeds that extend longer than the grain canal around the grain canal, and this weed is cut away by a reaper and detected by a threshing depth detection sensor. As a result, the threshing depth adjusting means adjusts the threshing depth in the threshing device with the length of the weed, which causes a problem that the threshing performance of the original reaping grain crusher is reduced.

In view of such circumstances, there has been a demand for a combine that can avoid the reduction in the threshing performance of the cultivated cereal as much as possible even in a field where weeds are grown locally.

Moreover, in the combine by patent document 2 or patent document 3, the state of the grain crucible ahead of a reaper is detected using an image processing technique, and operation | movement of a working apparatus is controlled based on the detection result. However, the actual position on the map in the field of the detected cereals is not taken into consideration. For this reason, even if a grain crumb in a problem state is detected in work travel, and even if avoidance control is performed to solve the problem, the area showing the problem grain crumb condition and the combine are far apart In this case, there arises a problem that the avoidance control is too early or too late. In addition, it is also difficult to properly determine the timing at which the avoidance control should be ended.

In view of such actual circumstances, there is also a demand for a harvester that can effectively support harvesting work using an image captured by an imaging unit provided on a vehicle body.

In the combine according to Patent Documents 2, 3 and 4, the traveling device and the working device are controlled from the detection results obtained based on the photographed image acquired by the on-vehicle camera. However, in the case of photographing by a camera provided in a combine traveling on a field, various situations may occur in which the photographed image becomes inappropriate. As such a situation, for example, since the traveling surface of the field is not flat like a road, a sudden vibration is transmitted to the on-vehicle camera. In addition, there are frequent fluctuations in the direction of the sun with respect to the on-vehicle camera and the like because the work travels while changing the direction. Furthermore, in reaping work etc., soaring dust may be mentioned at the time of work traveling. If the work device or the traveling device is controlled based on an inappropriate photographed image, the work traveling becomes inappropriate, and a satisfactory result can not be obtained.

In order to suppress such a problem, there is a demand for a technology that can eliminate inappropriate photographed images as much as possible in crop work support using the photographed images.

An object of the present invention is to further improve the performance of various tasks and the efficiency of tasks in light of the above-mentioned needs.

A feature of a harvester according to an embodiment of the present invention is a combine harvester for harvesting grained rice husk while traveling in a field, and a mowing section for mowing the grained rice husk from the field, and a mowing grain gutter Based on the length of the reaping grain weir using the grain conveying apparatus for conveying from the reaper to the threshing apparatus, the threshing depth adjusting mechanism provided in the threshing grain conveying apparatus and the threshing depth adjusting mechanism A handling depth control unit that performs handling depth adjustment control, an aircraft position calculation unit that calculates an aircraft position that is map coordinates of an aircraft based on positioning data from a satellite positioning module, and the aircraft are provided at the time of harvesting work The imaging unit for imaging the field and the image data of the photographed image sequentially acquired sequentially by the imaging unit are input, the weed growth region in the photographed image is estimated, and the estimated weed growth region is determined. Output recognition output data to indicate A weed position information generation unit for generating weed position information indicating a position on the map of the weed growth area from the image recognition module at the time when the photographed image is acquired and the recognition output data; The work traveling control unit executes the weed entry control while the weeds cut in the weed growing area pass the processing depth adjustment mechanism and the weeds pass through the processing depth adjustment mechanism. It is in the point being

In the harvester according to one embodiment of the present invention, when weeds are present in the photographed image, the weed growth area is estimated by the image recognition module from the image data which is the photographed image. Furthermore, since the machine position indicated by the map coordinates at the time of acquisition of the photographed image is calculated by the machine position calculation unit, the weed growth area is calculated from the machine position and the recognition output data indicating the weed growth area. Weed position information indicating the position on the map of is generated. The aircraft position, which is the map coordinates of the aircraft, is calculated by the aircraft position calculation unit, so by considering the weed's harvesting position in the vehicle and the weed's transport time from the harvesting position to the handling depth adjustment mechanism, The timing at which weeds pass through the depth adjustment mechanism is determined. Based on this determined timing, while weeds pass through the depth adjustment mechanism, by performing special weed entry control, weeds are mixed in the field where weeds are grown locally. It is possible to suppress the reduction of the threshing performance for the standing cereals.

A feature of a harvester according to an embodiment of the present invention is a combine harvester for harvesting grained rice husk while traveling in a field, and a mowing section for mowing the grained rice husk from the field, and a mowing grain gutter Based on the length of the reaping grain weir using the grain conveying device for conveying from the reaper to the threshing device, the threshing depth adjusting mechanism provided in the threshing grain conveying device, and the threshing depth adjusting mechanism A handling depth control unit that performs handling depth adjustment control, an aircraft position calculation unit that calculates an aircraft position that is map coordinates of an aircraft based on positioning data from a satellite positioning module, and the aircraft are provided at the time of harvesting work The imaging unit for imaging the field and the image data of the photographed image sequentially acquired sequentially by the imaging unit are input, the weed growth region in the photographed image is estimated, and the estimated weed growth region is determined. Output recognition output data to indicate A weed position information generation unit for generating weed position information indicating a position on the map of the weed growth area from the image recognition module at the time when the photographed image is acquired and the recognition output data; While the mowing part passes through the weed growth area, a work traveling control part is provided which executes weed entry control.

In the harvester according to one embodiment of the present invention, weed entry control is performed while the reaper passes through the weed growth area. In this configuration, there is an advantage that the control is simplified since it is not necessary to consider the transport time of weeds and the like.

The timing when the reaper passes through the weed growth area is the position on the map calculated based on the positioning data from the satellite positioning module, the distance between the body position and the reaper, and the weed position information. It can be accurately determined from the position on the map of the weed growth area contained in. From this, in one of the preferred embodiments of the present invention, the timing at which the mowing section passes through the weed growth area is based on the weed position information and the machine position calculated by the body position calculating unit. Are configured to be determined.

When weeds taller than the planted grain weir enter the depth control mechanism, the depth control mechanism considers the weed to be a crop stem, so the depth control unit sets the depth to the weed length. It controls it together. As a result, the adjusted threshing depth is not suitable for the original reaper. Since the length of the harvest grain weir does not change so rapidly, when weeds enter the threshing depth adjustment mechanism, temporarily interrupt threshing depth control based on the threshing depth adjuster. It is preferable not to change the treatment depth. From this, in one of the preferred embodiments of the present invention, the execution of the weed approach control is configured to interrupt the depth adjustment control.

In addition, since weeds grow together with planted grain weirs, when weeding the weed growth area, the amount of processing including weeds and weeded grain weeds increases, and the weeding area, The load on working devices such as the grain transfer device and the threshing device is increased. In order to avoid such an overload, it is preferable to reduce the vehicle speed. From this, in one of the preferred embodiments of the present invention, the execution of the weed approach control is configured to reduce the vehicle speed.

A feature of a harvester according to an embodiment of the present invention is a harvester that harvests crops while traveling in a field, and calculates an aircraft position which is map coordinates of the vehicle based on positioning data from a satellite positioning module An image data of a photographed image sequentially obtained by the photographing unit which is provided in the aircraft position calculating unit, the photographing unit for photographing the field at the time of harvesting operation, provided in the unit, is input, and recognition in the photographed image is performed. An image recognition module for estimating an existence area in which an object exists, and outputting recognition output data including the existence area and an estimated probability when the existence area is estimated; and the body at the time when the photographed image is acquired And a recognition target position information generation unit for generating recognition target position information indicating a position on the map of the recognition target from the position and the recognition output data.

In a harvester according to an embodiment of the present invention, an image recognition module for estimating an existing area in which a recognition target exists in a captured image from image data which is a captured image is, for example, a neural network (including deep learning) or It is configured using techniques such as reinforcement learning. Moreover, since it is comprised so that the presumed probability at the time of estimating the said presence area | region can also be output simultaneously, the optimal control using the probability value also becomes possible. Furthermore, the machine position indicated by the map coordinates at the time the captured image is acquired is calculated by the machine position calculation unit, and the recognition target is calculated from the position of the machine and the recognition output data indicating the presence area of the recognition target. Recognized object position information indicating the position of the object on the map is generated. From such a configuration, an estimation probability for estimating the existence area of the recognition object, and a position of the recognition object on the map, and consequently, a harvesting operation support considering the distance between the recognition object and the harvester Control of the

In the photographed image of the field photographed by the photographing unit, the resolution for the recognition target far from the photographing unit is lower than the resolution for the recognition target near from the photographing unit due to the perspective.
From this, the recognition object appearing at a location far from the imaging unit has lower recognition reliability than the recognition object appearing at a location near the imaging unit. Correspondingly, in one of the preferred embodiments of the present invention, the estimation probability of the recognition target is reduced as the recognition target is located farther from the imaging unit in the photographed image. Is configured.

Since generation of recognition target object position information based on a captured image can be performed at high speed as compared with the vehicle speed of the harvester, a plurality of recognition target object position information regarding the same recognition target object is generated. For this reason, it is possible to calculate statistically the respective estimated probabilities included in a plurality of recognition target object position information regarding the same recognition target. Through this statistical operation, more reliable recognition object position information can be generated. The statistical operation here is arithmetic average, weighted average, intermediate value operation or the like, and is an operation such that a more reliable data value is derived from a plurality of data values. By using such a statistical operation for the estimated probability, it is possible to derive more appropriate recognition target object position information from a plurality of recognition target object position information.
From this, in one of the preferred embodiments of the present invention, a plurality of pieces of recognition target object position information are stored, and a plurality of pieces of stored recognition target object position information are included in the corresponding recognition output data. It is configured to be corrected based on the result of the statistical calculation of the estimated probability.

Although the photographed image acquired by the photographing unit provided in the machine becomes an input source of the image recognition module, there is a possibility that it becomes unsuitable for the recognition of the recognition object depending on the photographing condition. As a result, even if the recognition target object is recognized, the recognition probability of the recognition output data becomes low, so it is not preferable to use the recognition output data as it is for the subsequent processing. From this, in one of the preferred embodiments of the present invention, there is provided a data storage unit for temporarily and temporally storing the recognition output data sequentially output from the image recognition module as a recognition output data string; A determination unit is provided, and the data determination unit has the estimated probability lower than a predetermined level as compared with the estimated probability of the recognized output data that is sequentially related in the recognition output data sequence. The recognition output data is determined as improper recognition output data. The recognition output data determined to be unsuitable recognition output data may be deleted, or the estimation probability of the recognition output data may be replaced with an estimation probability interpolated with preceding and following recognition output data.

The objects of recognition that are important for harvesters that harvest crops while traveling on the field are people, fallen grain weirs, weeds, straw and so on. Recognizing a person who is an obstacle for work travel is an important factor for safe travel in a field because it serves as a trigger for obstacle avoidance control and obstacle alarm notification. It is important to perform high-quality harvest work, because it becomes control control of weed measures and weed control measures in work control. It is important to recognize the borderline of the field and to recognize the driving of the field on the field, etc. to recognize the ridge.

A harvester according to an embodiment of the present invention calculates an aircraft position calculation unit which calculates an aircraft position which is map coordinates of an aircraft based on positioning data from a satellite positioning module, and calculates a traveling locus of the aircraft from the vehicle position. The image data of the photographed image sequentially acquired by the photographing unit, which is provided on the machine, and which is provided on the machine to photograph the field at the time of harvesting work, is input, and An image recognition module for estimating a presence area in which a recognition target is present, and outputting recognition output data including the presence area and an estimated probability when the presence area is estimated; and a previously captured image based on the traveling locus Predicting the range in which the existing region should be located in the next captured image, and the estimated probability of the existing region overlapping the range in the next captured image, and A data determination unit that determines the recognition output data based on the next captured image as inappropriate recognition output data when the estimated probability of the existing area in the previous captured image is different by a predetermined allowable amount or more Prepare.

When a recognition object appears in front while the harvest machine is working and traveling, usually, a plurality of photographed images in which the recognition object is photographed are acquired. Therefore, the presence area of the recognition target is estimated for each of the plurality of photographed images, and recognition output data including the estimated probability is output. When the recognition object recognized in each of the plurality of photographed images is substantially a stationary object, each of the photographed objects is photographed according to the position of the machine at which each photographed image is photographed, that is, according to the traveling locus. It will be at different locations in the image. Therefore, based on the traveling locus, the range in which the region where the recognition target is present in the captured image should be located can be predicted.
When the existing region of the recognition target is not located in the range predicted in this way, or when it is located, its estimation probability is significantly different from the estimation probability based on other photographed images, Images taken are considered inappropriate. Also, if the captured image is inappropriate, the recognition output data output based on it is also considered to be inappropriate. From this, based on the traveling locus, the range in which the existing area in the previous captured image should be positioned in the next captured image is predicted, and the estimated probability of the existing area overlapping the range in the next captured image and in the previous captured image When the estimation probability of the existing area is different by a predetermined allowable amount or more, it is useful to determine recognition output data based on the next photographed image as inappropriate recognition output data.

In one of the preferred embodiments of the present invention, the estimation probability of the inappropriate recognition output data is obtained based on the estimated probability of the recognition output data that is sequentially before and after the inappropriate recognition output data. A data correction unit is provided that replaces the interpolation estimation probability. In this configuration, the estimation probability of recognition output data determined to be unsuitable recognition output data is interpolated and corrected with the estimation probability of recognition output data before and after that, thereby replacing inappropriate recognition output data with proper recognition output data, It can be used. This is more effective when the number of recognition output data regarding the same recognition target is small.

The objects of recognition that are important for harvesters that harvest crops while traveling on the field are people, fallen grain weirs, weeds, straw and so on. Recognizing a person who is an obstacle for work travel is an important factor for safe travel in a field because it serves as a trigger for obstacle avoidance control and obstacle alarm notification. It is important to perform high-quality harvest work, because it becomes control control of weed control and weed control in operation control. It is important to recognize the borderline of the field and to recognize the driving of the field on the field, etc. to recognize the ridge.

FIG. 1 is an overall side view of a combine according to a first embodiment. FIG. 2 is a side view of the reaper according to the first embodiment. FIG. 6 is an explanatory view of depth control in the first embodiment. FIG. 6 is a functional block diagram showing functions of a control system of the combine in the first embodiment. FIG. 8 is an explanatory view schematically showing a flow of generation of recognition output data by the image recognition module in the first embodiment. It is a flowchart which shows the flow of control which interrupts a handling depth control by calculating a weed position from the captured image containing the weed growth area | region in Embodiment 1, and an airframe position. It is a schematic diagram which shows the weed map obtained by mapping weed position information in Embodiment 1. FIG. FIG. 10 is an overall side view of a combine according to a second embodiment. FIG. 13 is a functional block diagram showing a control system of the combine in the second embodiment. FIG. 16 is an explanatory view schematically showing a flow of generation of recognition output data by the image recognition module in the second embodiment. FIG. 13 is a data flow chart showing the flow of data when generating recognition target object position information from a captured image in Embodiment 2. FIG. It is a schematic diagram which shows the weed map obtained by mapping weed position information as one of recognition target object positional information in Embodiment 2. FIG. FIG. 18 is an overall side view of a combine according to a third embodiment. FIG. 16 is a functional block diagram showing a control system of the combine in the third embodiment. FIG. 18 is an explanatory view schematically showing a flow of generation of recognition output data by the image recognition module in the third embodiment. 15 is a data flow chart showing the flow of data when generating recognition target object position information from a captured image in the third embodiment. FIG. 21 is an explanatory view schematically illustrating data processing for determining improper recognition output data in the third embodiment. It is a schematic diagram which shows the weed map obtained by mapping weed position information as one of recognition target object positional information in Embodiment 3. FIG.

Hereinafter, an embodiment of a combine as an example of a harvester according to the present invention will be described based on the drawings. In each embodiment, when defining the front-back direction of the airframe 1, it defines along the airframe advancing direction in a working state. The direction indicated by the symbol (F) in FIG. 1 is the front side of the vehicle, and the direction indicated by the symbol (B) in FIG. 1 is the rear side of the vehicle. When defining the left-right direction of the airframe 1, right and left are defined in the state seen in the advancing direction of the airframe.

Embodiment 1
As shown in FIG. 1, in the combine according to the first embodiment, the reaper 2 is connected to the front of the machine body 1 including the pair of left and right crawler traveling devices 10 so as to be movable up and down around the horizontal axis X. A threshing device 11 and a grain tank 12 for storing grains are provided at the rear of the machine body 1 in a state of being aligned in the machine body width direction. A cabin 14 is provided at the front right side of the airframe 1 to cover the boarding driver, and a driving engine 15 is provided below the cabin 14.

As shown in FIG. 1, the threshing device 11 internally receives a reaping grain weir which is cut by the reworking unit 2 and transported backward, and the original origin of the weir by the threshing feed chain 111 and the pinching rail 112 The tip end side is threshed with a threshing cylinder 113 while holding and conveying. Then, the grain sorting process is performed on the threshing treatment product in the sorting unit provided below the threshing drum 113, and the grain sorted there is transported to the grain tank 12 and stored. Although not described in detail, a grain discharging device 13 for discharging grains stored in the grain tank 12 to the outside is provided.

The reaper 2 is provided with a plurality of raising devices 21 for causing fallen cereal grains, a clipper-type cutting device 22 for cutting the origin of the raised cereal grains, a grain feeding device 23, etc. There is. The grain feeder conveying device 23 is directed toward the beginning end of the threshing feed chain 111 of the threshing device 11 located on the rear side of the machine while gradually changing the cropped grain gutter in the vertical posture in which the stock origin is cut to a lying posture. Transport

The grain feeder conveying device 23 holds the stock origin of the combined feed conveying unit 231 which gathers and gathers a plurality of cutting grain rods cut by the cutting device 22 to the center in the cutting width direction, sandwiches the stock origin of the gathered grain collectors From the end of the stock holding and conveying device 232 to the threshing feed chain 111 to the threshing feed chain 111 It has a supply conveyance device 234 and the like that directs and guides it.

As shown in FIG. 2, the stock holder holding and conveying device 232 is swingably supported by the support frame of the reaper 2 around the horizontal axis. The stock holding and conveying device 232 is operated to swing up and down by the drive operation mechanism 235, and along with the swinging operation, the conveyance end portion is provided so as to change the position of the feeding and conveying device 234 in the scale length direction of the grain scale. ing. The drive operation mechanism 235 has a steering depth adjustment electric motor 236 (hereinafter referred to as a steering depth motor) as a drive source. The drive operation mechanism 235 includes an operation rod 237 which is pushed and pulled by the depth-of-feed motor 236.
The lower end portion of the operation rod 237 is pivotally connected to an intermediate portion of the stock origin holding and conveying device 232.

When the transport end portion of the stock origin nipping and conveying device 232 is separated from the feeding and conveying device 234, the stock origin nipping position of the harvesting grain by the supply transport device 234 is at the stock origin nipping position of the harvesting grain by the stock origin nipping and transporting device 232. On the other hand, it is changed to the tip side and is delivered to the feeding and conveying device 234. As a result, the entry depth (handling depth) of the harvest grain weir into the threshing device 11 is changed to a shallower (shallow side).

When the transport end portion of the stock origin gripping and conveying device 232 approaches the feeding and transporting device 234, the stock origin pinching position of the harvesting grain by the feed conveying device 234 is the stock origin pinching position of the harvesting grain by the stock origin gripping and transporting device 232 It is delivered to the feeding and conveying device 234 in a state where it is in the near position. As a result, the depth of threshing equipment 11 for the reaping grain can be changed to the deeper side (deep side).

By changing the posture of the stock origin holding and conveying device 232 in this manner, it is possible to change the handling depth of the harvesting grain cane with respect to the threshing device 11. In other words, the stock origin holding and conveying device 232 and the drive operation mechanism 235 constitute a threshing depth adjustment mechanism 3 capable of changing the threshing depth of the reaping grain gutter to the threshing device 11.

As shown in FIG. 3, the combine according to the first embodiment includes a contact-type scale length detecting device 30 for detecting the scale length of the reaping grain transferred by the grain transfer device 23 and a threshing depth control unit 620. It is equipped. The handling depth control unit 620 performs handling depth adjustment control for adjusting the handling depth based on the detection result of the eyelid length detection device 30. In this embodiment, the threshing depth control unit 620 controls the threshing depth motor 236 so that the threshing depth of the harvesting grain cane with respect to the threshing device 11 is maintained within the target setting range.

As shown in FIG. 3, the eyelid length detection device 30 is provided with a pair of swingable sensor arms 32, 33 on a main body case 31 as an apparatus main body portion formed in a substantially bottomless box shape opened downward. The pair of sensor arms 32 and 33 contact with the tip end of the reaping grain weir to detect the length of the beak. The pair of sensor arms 32 and 33 are provided in a state of being supported by the main body case 31 and hanging down toward the lower side in a state of being separated in the ridge length direction of the crop grain to be conveyed. Each sensor arm 32, 33 is pivotable back and forth (corresponding to the moving direction of cutting grain) around the horizontal axis center provided inside the main body case 31, and is biased to return to the downward reference posture Being supported in the

When the sensor arm 32, 33 is swung more than the set amount from the reference posture by bringing the reaper to be conveyed into contact with the base end side portion of the upper portion of each of the pair of sensor arms 32, 33, the sensor arm is turned on Detection switches 34 and 35 are provided which are turned off if the amount of oscillation from the reference attitude of 32 and 33 is less than the set amount.

The outputs of the pair of detection switches 34 and 35 are input to the depth control unit 620. The handling depth control unit 620 controls the operation of the handling depth motor 236 such that the detection switch 35 located on the tip side is in the off state, and the detection switch 34 located on the stock origin side is in the on state. Do.

That is, when the pair of detection switches 34 and 35 are both in the on state, the handling depth control unit 620 operates the handling depth motor 236 so that the stock origin holding conveyance device 232 moves to the shallow handling side. In addition, the handling depth control unit 620 operates the handling depth motor 236 so that the stock origin holding conveyance device 232 moves to the deep handling side when the pair of detection switches 34 and 35 are both in the OFF state. Furthermore, if the detection switch 35 located on the tip side of the pair of detection switches 34 and 35 is in the OFF state and the detection switch 34 located on the stock origin side is in the ON state, the handling depth control unit 620 The operation of the motor 236 is stopped and the state is maintained.

As shown in FIG. 1, at the front end of the ceiling portion of the cabin 14, an imaging unit 70 provided with a color camera is provided. The fore-and-aft extension of the imaging field of view of the imaging unit 70 substantially reaches the horizon from the front end region of the reaper 2. The breadth of the field of view in the field of view reaches about 10 m to several tens of meters. The photographed image acquired by the photographing unit 70 is converted into image data and sent to the control system of the combine.

The photographing unit 70 photographs a field at the time of harvesting work. The control system of the combine has a function of recognizing the weed growth area as a recognition object from the image data sent from the imaging unit 70. In FIG. 1, a group of normal planted cereals is indicated by a symbol Z0, and a weed growth area is indicated by a symbol Z1.

Furthermore, a satellite positioning module 80 is also provided on the ceiling of the cabin 14. The satellite positioning module 80 includes a satellite antenna for receiving global navigation satellite system (GNSS) signals (including GPS signals). In order to complement the satellite navigation by the satellite positioning module 80, an inertial navigation unit incorporating a gyro acceleration sensor and a magnetic direction sensor is incorporated in the satellite positioning module 80. Of course, the inertial navigation unit can be located elsewhere. In FIG. 1, the satellite positioning module 80 is disposed at the rear of the ceiling of the cabin 14 for convenience of drawing, but for example, the satellite positioning module 80 is positioned as close as possible to a position just above the center of left and right of the cutting device 22. It is preferable that the front end portion be disposed closer to the center of the airframe.

FIG. 4 shows a functional block diagram of a control system built inside the combine aircraft 1. The control system of this embodiment comprises an electronic control unit called a large number of ECUs, various operation devices, sensors, switches, and a wiring network such as a car-mounted LAN for data transmission between them. The notification device 91 is a device for notifying a driver or the like of a work traveling state and various warnings, and is a buzzer, a lamp, a speaker, a display, or the like. The communication unit 92 is used to exchange data with the cloud computer system 100 or the portable communication terminal 200 installed at a remote place. Here, the mobile communication terminal 200 is a tablet computer operated by a supervisor (including a driver) at the work travel site. The control unit 6 is a core element of this control system, and is shown as a collection of a plurality of ECUs. The positioning data acquired by the satellite positioning module 80 and the image data acquired by the imaging unit 70 are input to the control unit 6 through the wiring network.

The control unit 6 includes an output processing unit 6B and an input processing unit 6A as an input / output interface. The output processing unit 6B is connected to the vehicle travel device group 7A and the work device device group 7B. The vehicle travel device group 7A includes control devices related to vehicle travel, such as an engine control device, a shift control device, a braking control device, a steering control device, and the like. The working device group 7B includes a power control device and the like in the reaper 2, the threshing device 11, the grain discharging device 13, the grain conveying device 23, and the threshing depth adjusting mechanism 3.

A traveling system detection sensor group 8A, a working system detection sensor group 8B, and the like are connected to the input processing unit 6A. The traveling system detection sensor group 8A includes a sensor that detects the state of an engine speed adjuster, an accelerator pedal, a brake pedal, a gearshift operator, and the like. The working system detection sensor group 8B includes a cutting unit 2, a threshing device 11, a grain discharging device 13, and a device state in the grain conveying device 23 and a sensor for detecting the state of the grain or grain. Furthermore, the working system detection sensor group 8B also includes the detection switches 34 and 35 in the above-described handling depth adjustment mechanism 3.

The control unit 6 includes an image recognition module 5, a data processing module 50, a work travel control module 60 which is a work travel control unit, a machine position calculation unit 66, a notification unit 67, and a weed position calculation unit 68.

The notification unit 67 generates notification data on the basis of a command or the like received from each functional unit of the control unit 6, and gives the notification data to the notification device 91. The aircraft position calculation unit 66 calculates the aircraft position which is the map coordinates (or field coordinates) of the aircraft 1 based on the positioning data sequentially sent from the satellite positioning module 80. The weed position calculation unit 68 according to this embodiment calculates the position of the weed growth area, which is normally calculated by the data processing module 50, on the map of the weed growth area calculated by the body position calculation unit 66, and the grain. The timing at which the weeds harvested by the reaper 2 pass through the depth adjustment mechanism 3 is determined based on the transport speed of the straw transport device 23.

The combine according to this embodiment can travel in both automatic travel (automatic steering) and manual travel (manual steering). In addition to the travel control unit 61 and the work control unit 62, the work travel control module 60 includes an automatic work travel instruction unit 63 and a travel route setting unit 64. A traveling mode switch (not shown) is provided in the cabin 14 to select one of an automatic traveling mode in which the vehicle travels by automatic steering and a manual steering mode in which the vehicle travels by manual steering. By operating the travel mode switch, it is possible to shift from manual steering travel to automatic steering travel, or shift from automatic steering travel to manual steering travel.

The traveling control unit 61 has an engine control function, a steering control function, a vehicle speed control function, and the like, and provides a traveling control signal to the vehicle traveling device group 7A. The work control unit 62 provides a work control device group 7B with a work control signal in order to control the movement of the reaper 2, the threshing device 11, the grain discharging device 13, the grain feeding device 23, and the like. Furthermore, the operation control unit 62 includes the depth control unit 620 described with reference to FIG. 3.

When the manual steering mode is selected, the traveling control unit 61 generates a control signal based on an operation by the driver to control the vehicle traveling device group 7A. When the automatic steering mode is selected, the traveling control unit 61 controls the traveling equipment group 7A related to steering and the traveling equipment group 7A related to vehicle speed based on the automatic traveling instruction given by the automatic work traveling instruction unit 63. .

The travel route setting unit 64 expands the travel route for automatic travel created by any of the control unit 6, the mobile communication terminal 200, the cloud computer system 100, and the like in the memory. The travel route developed in the memory is sequentially used as a target travel route in automatic travel. Even if this traveling route is manual traveling, it is also possible to use the guidance for the combine to travel along the traveling route.

More specifically, the automatic work travel instruction unit 63 generates an automatic steering instruction and a vehicle speed instruction, and gives the travel control unit 61. The automatic steering command is generated so as to eliminate the azimuth deviation and the positional deviation between the traveling route set by the traveling route setting unit 64 and the vehicle position calculated by the machine position calculation unit 66. The vehicle speed command is generated based on a previously set vehicle speed value. Furthermore, the automatic work travel instruction unit 63 gives the work control unit 62 an operation device operation instruction according to the vehicle position and the traveling state of the vehicle.

The image recognition module 5 receives image data of a captured image sequentially acquired by the imaging unit 70 sequentially. The image recognition module 5 estimates an existing area in the photographed image in which a recognition target exists, and outputs recognition output data including an existing area and an estimated probability when the existing area is estimated as a recognition result. The image recognition module 5 is constructed using neural network technology adopting deep learning.

The flow of generation of recognition output data by the image recognition module 5 is shown in FIG. 5 and FIG. The image recognition module 5 receives pixel values of RGB image data as input values. In this embodiment, the putative verification object is a weed. Therefore, the recognition output data as the recognition result includes a weed growth area shown by a rectangle and an estimated probability when the weed growth area is estimated.

In FIG. 5, the estimation result is schematically shown, and the weed growth area is indicated by a rectangular frame denoted by reference numeral F1. The weed growth area is defined by four corner points respectively, but the coordinate positions on the photographed image of the four corner points of each such rectangle are also included in the estimation result. Of course, if the weed as a certification target is not estimated, the weed growth area is not output, and the estimation probability becomes zero.

In this embodiment, the image recognition module 5 sets internal parameters so that the probability of estimating the recognition object is reduced as the recognition object (weed) is located farther from the imaging unit 70 in the photographed image. doing. As a result, recognition of an object to be recognized in an imaging region whose resolution is low due to being far from the imaging unit 70 is made strict, and erroneous recognition is reduced.

The data processing module 50 processes recognition output data output from the image recognition module 5. As shown in FIGS. 4 and 6, the data processing module 50 of this embodiment includes a weed position information generation unit 51 and a statistical processing unit 52.

The weed position information generation unit 51 generates weed position information indicating the position on the map of the recognition object from the machine position at the time when the captured image is acquired and the recognition output data. The position on the map where weeds included in the recognition output data is converted from the coordinate position (camera coordinate position) on the photographed image of the four corner points of the rectangle indicating the weed to the coordinates on the map can get.

Since the photographing unit 70 acquires photographed images at predetermined time intervals, for example, 0.5 seconds, and inputs the image data to the image recognition module 5, the image recognition module 5 also recognizes and outputs recognition output data at the same time intervals. Output. Therefore, when weeds are included in the imaging field of view of the imaging unit 70, a plurality of recognition output data will include the presence regions for the same weeds. As a result, multiple weed location information for the same weed can be obtained. At that time, the estimation probability included in the recognition output data which is each source data, that is, the estimation probability of the weed presence area (weed growth area) included in the weed position information is the position between the imaging unit 70 and the weed Because the relationships are different, they often have different values.

Therefore, in this embodiment, such multiple weed position information is stored, and the estimated probability included in each of the stored multiple weed position information is calculated statistically. A representative value of the estimated probability group is determined using a statistical operation on the estimated probabilities of the plurality of recognition object position information. A plurality of pieces of recognition target object position information can be corrected to one piece of optimum recognition target object position information using the representative value. One example of such correction is to obtain the arithmetic mean value or weighted mean value or median value of each estimated probability as a reference value (representative value), and the logic of the existing area (weed growth area) having an estimated probability equal to or higher than the reference value. It is to calculate the sum and generate corrected weed position information which makes it the optimum existing area. Of course, other statistical operations can be used to generate reliable single weed position information.

As shown in FIG. 6, the weed position information indicating the position on the map of the weed growth area thus obtained is given to the weed position calculation unit 68. The weed position calculation unit 68 is also given an aircraft position which is map coordinates of the vehicle position calculated by the vehicle position calculation unit 66. The weed position calculation unit 68 determines the passing timing at which the weeds cut off by the harvesting unit 2 pass through the handling depth adjustment mechanism 3 based on the weed position information, the machine position, and the transfer speed of the grain feeding device 23. decide. While the weed passes the handling depth adjustment mechanism 3, the weed position calculation unit 68 gives a weed entry flag to the handling control unit 620.

The handling depth control unit 620 has a standard control mode and a weed approach control mode, and normally, the standard control mode is selected, and the above-mentioned handling depth control is executed during work travel. .
However, when the weed entry flag is given from the weed position calculation unit 68, the standard control mode is switched to the weed entry control mode, and weed entry control is executed. In this embodiment, when weed entry control is performed, the depth control is interrupted and the vehicle speed is also reduced. Of course, in the weed entry control, a configuration may be employed in which only one of interruption of the depth control and reduction of the vehicle speed is performed.

The weed position information generated by the weed position information generation unit 51 can be mapped as shown in FIG. 7 for easy visual display. FIG. 7 illustrates a weed map in which weed position information is mapped. When weed growth areas having different estimation probabilities are included in the weed position information, it is possible to display the weed growth areas in a form divided into patterns within a predetermined range of estimation probability values, as shown in FIG. .

Second Embodiment
As shown in FIG. 8, in the combine, a reaper 2002 is connected to the front of an airframe 2001 provided with a pair of left and right crawler traveling devices 2010 so as to be able to move up and down around the horizontal axis X. At the rear of the fuselage 2001, a threshing device 2011 and a grain tank 2012 for storing grains are provided in line with the width direction of the fuselage. A cabin 2014 is provided at the front right side of the fuselage 2001 to cover the boarding driver, and a driving engine 2015 is provided below the cabin 2014.

As shown in FIG. 8, the threshing device 2011 internally receives a reaping grain weir which is reaped by the reaping unit 2002 and transported backward, and the origin of the grain weir by the threshing feed chain 2111 and the clamping rail 2112. While holding and transporting, the tip side is threshed with a threshing cylinder 2113. Then, the grain sorting process is performed on the threshed product in the sorting unit provided below the threshing drum 2113, and the grain sorted there is transported to the grain tank 2012 and stored. Although not described in detail, a grain discharging device 2013 for discharging grains stored in the grain tank 2012 to the outside is provided.

The reaper 2002 is equipped with a plurality of raising devices 2021 for causing fallen cereal grains, a clipper-type cutting device 2022 for cutting the origin of the raised cereal grains, a grain feeding device 2023 and the like. There is. The grain feeder conveyance device 2023 is directed toward the beginning end of the threshing feed chain 2111 of the threshing device 2011 located on the rear side of the fuselage, while gradually changing the cropped grain gutter in the vertical posture in which the stock origin has been cut to a lying posture. Transport

The grain feeder conveying device 2023 holds a rear of the combined feeding unit 2231 which conveys a plurality of cut grain remnants cut by the cutting device 2022 while gathering them to the center in the cutting width direction, sandwiches the stock origin of the gathered grain collecting trouts The stock origin holding transport device 2232 for transporting to the tip, the ear tip holding transport device 2233 for locking and carrying the tip side of the harvesting grain pod, the end of the stock origin gripping transport device 2232 The stock origin of the harvesting grain stem to the threshing feed chain 2111 It has a supply conveyance device 2234 and the like for directing and guiding.

At the front end of the ceiling of the cabin 2014, a photographing unit 2070 provided with a color camera is provided.
In this embodiment, the fore-and-aft extension of the imaging field of view of the imaging unit 2070 substantially reaches the horizon from the front end region of the mowing unit 2002. The breadth of the field of view in the field of view reaches about 10 m to several tens of meters. The photographed image acquired by the photographing unit 2070 is converted into image data, and is sent to the control system of the combine.

The photographing unit 2070 photographs a field at the time of harvesting work, but various objects exist in the field as objects to be photographed. The control system of the combine has a function of recognizing a specific object as a recognition target object from the image data sent from the imaging unit 2070. As such a recognition object, in FIG. 8, a normal open-celled flock group indicated by code Z0, a weed group extended higher than an open-celled hail indicated by code Z1, an lodging indicated by code Z2. A group of cereals, a person indicated by a symbol Z3 is schematically shown.

A satellite positioning module 2080 is also provided on the ceiling of the cabin 2014. The satellite positioning module 2080 includes a satellite antenna for receiving global navigation satellite system (GNSS) signals (including GPS signals). In order to complement the satellite navigation by the satellite positioning module 2080, an inertial navigation unit incorporating a gyro acceleration sensor and a magnetic direction sensor is incorporated in the satellite positioning module 2080. Of course, the inertial navigation unit can be located elsewhere. In FIG. 8, the satellite positioning module 2080 is disposed at the rear of the ceiling of the cabin 2014 for convenience of drawing, but for example, the satellite positioning module 2080 is positioned as close as possible to a position just above the center of left and right of the cutting device 2022. It is preferable that the front end portion be disposed closer to the center of the airframe.

A functional block diagram of the control system of the combine is shown in FIG. The control system of this embodiment comprises an electronic control unit called a large number of ECUs, various operation devices, sensors, switches, and a wiring network such as a car-mounted LAN for data transmission between them. The notification device 2091 is a device for notifying a driver or the like of a work traveling state and various warnings, and is a buzzer, a lamp, a speaker, a display, or the like. The communication unit 2092 is used to exchange data with the cloud computer system 2100 or the portable communication terminal 2200 installed at a remote place. Here, the mobile communication terminal 2200 is a tablet computer operated by a supervisor (including a driver) at the work travel site.
The control unit 2006 is a core element of this control system, and is shown as a collection of a plurality of ECUs. The positioning data from the satellite positioning module 2080 and the image data from the imaging unit 2070 are input to the control unit 2006 through the wired network.

The control unit 2006 includes an output processing unit 2006B and an input processing unit 2006A as an input / output interface. The output processing unit 2006B is connected to the vehicle travel device group 2007A and the work device device group 2007B. The vehicle travel device group 2007A includes control devices related to vehicle travel, such as an engine control device, a transmission control device, a braking control device, a steering control device, and the like. The work device group 2007B includes a power control device and the like in the harvesting unit 2002, the threshing device 2011, the grain discharging device 2013, and the grain conveying device 2023.

A traveling system detection sensor group 2008A, a reaper 2008B, and the like are connected to the input processing unit 2006A. The traveling system detection sensor group 2008A includes a sensor that detects the state of an engine speed adjuster, an accelerator pedal, a brake pedal, a gearshift operator, and the like. The reaping unit 2008B includes a reaping unit 2002, a threshing device 2011, a grain discharging device 2013, and a device state in the grain conveying device 2023 and a sensor that detects the state of the grain or grain.

The control unit 2006 includes a work travel control module 2060, an image recognition module 2005, a data processing module 2050, a machine position calculation unit 2066, and a notification unit 2067.

The notification unit 2067 generates notification data based on an instruction or the like received from each functional unit of the control unit 2006, and gives the notification data to the notification device 2091. The aircraft position calculation unit 2066 calculates the aircraft position which is the map coordinates (or field coordinates) of the aircraft 2001 based on the positioning data sequentially sent from the satellite positioning module 2080.

The combine according to this embodiment can travel in both automatic travel (automatic steering) and manual travel (manual steering). The work travel control module 2060 includes an automatic work travel instruction unit 2063 and a travel path setting unit 2064 in addition to the travel control unit 2061 and the work control unit 2062. A travel mode switch (not shown) is provided in the cabin 2014 for selecting one of an automatic travel mode in which automatic travel is performed and a manual steering mode in which manual travel is performed. By operating the travel mode switch, it is possible to shift from manual steering travel to automatic steering travel, or shift from automatic steering travel to manual steering travel.

The traveling control unit 2061 has an engine control function, a steering control function, a vehicle speed control function, and the like, and supplies a traveling control signal to the vehicle traveling device group 2007A. The work control unit 2062 provides a work control device group 2007B with a work control signal in order to control the movement of the reaper 2002, the threshing device 2011, the grain discharging device 2013, the grain conveying device 2023, and the like.

When the manual steering mode is selected, the traveling control unit 2061 generates a control signal based on an operation by the driver to control the vehicle traveling device group 2007A. When the automatic steering mode is selected, the traveling control unit 2061 controls the vehicle traveling equipment group 2007A related to steering and the vehicle traveling equipment group 2007A related to vehicle speed based on the automatic traveling instruction given by the automatic work traveling instruction unit 2063. .

The travel route setting unit 2064 deploys, in the memory, a travel route for automatic travel created by any of the control unit 2006, the mobile communication terminal 2200, the cloud computer system 2100, and the like. The travel route developed in the memory is sequentially used as a target travel route in automatic travel. Even if this traveling route is manual traveling, it is also possible to use the guidance for the combine to travel along the traveling route.

More specifically, the automatic work travel instruction unit 2063 generates an automatic steering instruction and a vehicle speed instruction, and supplies the instruction to the travel control unit 2061. The automatic steering command is generated so as to eliminate the azimuth deviation and the positional deviation between the traveling route set by the traveling route setting unit 2064 and the vehicle position calculated by the machine position calculation unit 2066. The vehicle speed command is generated based on a previously set vehicle speed value. Furthermore, the automatic work travel instruction unit 2063 gives the work control unit 2062 a work device operation instruction according to the vehicle position and the traveling state of the vehicle.

The image recognition module 2005 receives image data of a captured image sequentially acquired by the imaging unit 2070 sequentially. The image recognition module 2005 estimates an existing area in the photographed image in which a recognition target exists, and outputs recognition output data including an existing area and an estimated probability when the existing area is estimated as a recognition result. The image recognition module 2005 is constructed using neural network technology that employs deep learning.

The flow of generation of recognition output data by the image recognition module 2005 is shown in FIG. 10 and FIG. The image recognition module 2005 receives pixel values of RGB image data as input values. In the illustrated example of FIG. 10, the authentication objects to be estimated are a weed and a fallen cereal weir and a person. Therefore, in the recognition output data as the recognition result, the present region of weeds (hereinafter referred to as weed region) and its estimated probability, the present region of the fallen kernel (hereinafter referred to as the fallen cereal region) and its estimated probability, It includes the existence area of a person (hereinafter referred to as a person area) and its estimated probability.

In FIG. 10, the estimation result is schematically shown, the weed area is shown by a rectangular frame with a reference F1, the fallen grain area is shown with a rectangular frame with a reference F2, and the human area is It is indicated by a rectangular frame given a code F3. Each region is linked to its estimated probability. The weed area, the fallen grain area and the human area are each defined by four corner points, but the coordinate positions on the photographed image of the four corner points of each rectangle are also included in the estimation result. Of course, if the authentication target is not estimated, the existence region of the authentication target is not output, and the estimation probability becomes zero.

In this embodiment, the image recognition module 2005 sets internal parameters so that the probability of estimating the recognition target decreases as the recognition target is located farther from the imaging unit 2070 in the captured image. . As a result, recognition of an object to be recognized in an imaging region whose resolution is low because it is far from the imaging unit 2070 is made stricter, and erroneous recognition is reduced.

The data processing module 2050 processes recognition output data output from the image recognition module 2005. As shown in FIGS. 9 and 11, in the data processing module 2050 of this embodiment, a recognition target object position information generation unit 2051, a statistical processing unit 2052, a data storage unit 2053, a data determination unit 2054, and a data correction unit 2055 are provided. include.

The recognition target position information generation unit 2051 generates recognition target position information indicating the position on the map of the recognition target from the machine position at the time when the captured image is acquired and the recognition output data. The position on the map where the recognition target (weed, fallen grain moth, person) is included in the recognition output data is a rectangle that indicates the presence area of the certification object (weed area, fallen grain moth area, person area) The coordinate positions (camera coordinate positions) on the photographed image of the four corner points of are obtained by converting the coordinates on the map.

The imaging unit 2070 acquires captured images at predetermined time intervals, for example, 0.5 seconds, and inputs the image data to the image recognition module 2005. Therefore, the image recognition module 2005 also recognizes and outputs recognition output data at the same time intervals. Output. Therefore, when the recognition target is included in the shooting field of view of the imaging unit 2070, the plurality of recognition output data includes the existence area for the same recognition target. As a result, multiple pieces of recognition target object position information for the same recognition target can be obtained. At that time, the estimation probability included in the recognition output data which is each original data, that is, the estimation probability of the existence region of the recognition target object included in the recognition target object position information is between the photographing unit 2070 and the recognition target object. Because the positional relationship is different, they often have different values.

Therefore, in this embodiment, such plural pieces of recognition target object position information are stored, and an estimated probability included in each of the stored plural pieces of recognition target object position information is calculated statistically.
The statistical processing unit 2052 obtains a representative value of the estimated probability group by using a statistical operation on estimated probabilities of a plurality of pieces of recognition object position information. A plurality of pieces of recognition target object position information can be corrected to one piece of optimum recognition target object position information (recognition target object correction position information) using the representative value. One example of such correction is to obtain the arithmetic mean value or weighted mean value or median value of each estimated probability as a reference value (representative value), and obtain the logical sum of the existing regions having the estimated probability equal to or higher than the reference value It is to generate correction recognition target object position information in which the Of course, it is also possible to generate one reliable object position information with high reliability using statistical operations other than this. That is, the plurality of recognition target object position information is corrected based on the result of the statistical calculation of the estimation probability included in the recognition output data corresponding to the recognition target object position information.

Recognition object position information (weed position information, fallen grain weir position information, person position) indicating the position on the map of the existence area (weed area, fallen grain weir area, person area) of the certification object thus obtained By using the information), at the time of recognition of weeds, fallen grain weirs, and persons, traveling work control and warning notification set in advance are performed.

As described above, in the present invention, the estimated probability output from the image recognition module 2005 is important for generation of final recognition target object position information. Therefore, in this embodiment, the recognition output data is evaluated by examining the reliability of the estimation probability of the recognition output data output from the image recognition module 2005 and determining the recognition output data having the unreliable estimation probability as the inappropriate recognition output data. A function is included in data processing module 2050.

The recognition output data evaluation function is realized by a data storage unit 2053, a data determination unit 2054, and a data correction unit 2055, as shown in FIGS. The data storage unit 2053 temporarily and temporally stores recognition output data sequentially output from the image recognition module 2005 as a recognition output data sequence. The data determination unit 2054 compares the estimated probability of the recognition output data to be determined with the estimated probability of the recognition output data sequentially related to the recognition output data in the recognition output data sequence. Recognition output data having an estimated probability lower than the level is determined as inappropriate recognition output data.

Since the photographed image used to generate the recognition output data is photographed during the traveling operation of the combine, the photographed image may become unclear due to the sudden movement of the combine and the like. Furthermore, since the combine changes its direction near the eyelid, the shooting direction may be suddenly changed. Along with this, there may be a mixture of shooting with the ordinary light and shooting with the back light for the same object to be recognized. Although such fluctuation in imaging conditions may cause recognition output data having an unexpectedly low estimation probability, extraction of inappropriate recognition output data which is such recognition output data is performed by the recognition output data evaluation function described above. It is possible.

When the number of recognition output data is small, the estimation probability of the recognition output data determined to be unsuitable recognition output data may be interpolated and corrected with the estimation probability of recognition output data before and after that. As a result, the inappropriate recognition output data is appropriately corrected and can be used, and the number of data can be secured. The data correction unit 2055 performs such interpolation correction. Of course, when the recognition output data determined to be unsuitable recognition output data is discarded, the interpolation correction becomes unnecessary, so the data correction unit 2055 is omitted.

Note that each recognition target object position information generated by the recognition target object position information generation unit 2051 can be mapped as shown in FIG. 12 for easy visual understanding. FIG. 12 illustrates a weed map in which weed position information is mapped. When the weed location information includes weed presence regions having different estimation probabilities, it is possible to display the weed presence regions in a form divided in a predetermined range of estimation probability values, as shown in FIG. .

Third Embodiment
As shown in FIG. 13, in the combine according to the third embodiment, a reaper 3002 is connected to a front portion of an airframe 3001 including a pair of left and right crawler traveling devices 3010 so as to be movable up and down around a horizontal axis X. A threshing device 3011 and a grain tank 3012 for storing grains are provided at the rear of the machine body 3001 in a state of being aligned in the machine body width direction. A cabin 3014 is provided at the front right side of the fuselage 3001 to cover the boarding operation unit, and a driving engine 3015 is provided below the cabin 3014.

As shown in FIG. 13, the threshing device 3011 receives internally the cutting grain weir which is cut off by the cutting unit 3002 and transported backward, and the origin of the grain weir by the threshing feed chain 3111 and the clamping rail 3112 While holding and transporting, the tip side is threshed by a threshing cylinder 3113. Then, the grain sorting process is performed on the threshed product in the sorting unit provided below the threshing drum 3113, and the grain sorted there is transported to the grain tank 3012 and stored. Further, although not described in detail, a grain discharging device 3013 for discharging grains stored in the grain tank 3012 to the outside is provided.

The reaper 3002 is provided with a plurality of raising devices 3021 for causing fallen cereal grains, a clipper-type cutting device 3022 for cutting the origin of the raised cereal grains, a grain feeding device 3023 and the like. There is. The grain feeder conveying device 3023 is directed toward the beginning end of the threshing feed chain 3111 of the threshing device 3011 located on the rear side of the fuselage while gradually changing the cropped grain gutter in the vertical posture in which the stock origin has been cut to a lying posture. Transport

The grain feeder conveying device 3023 holds a rear of the confluence conveying unit 3231 which gathers and gathers a plurality of cut grain remnants cut by the cutting device 3022 to the center in the cutting width direction, sandwiches the stock origin of the gathered grain relics From the end of the stock holding and conveying device 3232 and the stock holding and conveying device 3232 to the threshing feed chain 3111 It is provided with a feeding and conveying device 3234 and the like that directs it to guide it.

At the front end of the ceiling of the cabin 3014, a photographing unit 3070 provided with a color camera is provided.
In this embodiment, the extension in the front-rear direction of the field of view of the imaging unit 3070 substantially reaches the horizon from the front end region of the reaper 3002. The breadth of the field of view in the field of view reaches about 10 m to several tens of meters. The photographed images sequentially acquired by the photographing unit 3070 are converted into image data and sent to the control system of the combine.

The photographing unit 3070 photographs a field at the time of harvesting work, but various objects exist in the field as objects to be photographed. The control system of the combine has a function of recognizing a specific object as a recognition target object from the image data sent from the imaging unit 3070. As such a recognition object, in FIG. 13, a normal open-cell crop group indicated by symbol Z0, a weed group extended higher than the open-cell granule indicated by symbol Z1, and an lodging indicated by symbol Z2 A group of cereals, a person indicated by a symbol Z3 is schematically shown.

A satellite positioning module 3080 is also provided on the ceiling of the cabin 3014. The satellite positioning module 3080 includes a satellite antenna for receiving global navigation satellite system (GNSS) signals (including GPS signals). In order to complement the satellite navigation by the satellite positioning module 3080, an inertial navigation unit incorporating a gyro acceleration sensor and a magnetic direction sensor is incorporated in the satellite positioning module 3080. Of course, the inertial navigation unit can be located elsewhere. In FIG. 13, the satellite positioning module 3080 is disposed at the rear of the ceiling of the cabin 3014 for convenience of drawing, but for example, the satellite positioning module 3080 is positioned as close as possible to a position just above the center of the left and right of the cutting device 3022. It is preferable that the front end portion be disposed closer to the center of the airframe.

FIG. 14 shows a functional block diagram of the control system of the combine. The control system of this embodiment comprises an electronic control unit called a large number of ECUs, various operation devices, sensors, switches, and a wiring network such as a car-mounted LAN for data transmission between them. The notification device 3091 is a device for notifying a driver or the like of a work traveling state and various warnings, and is a buzzer, a lamp, a speaker, a display or the like. The communication unit 3092 is used to exchange data with the cloud computer system 3100 or the portable communication terminal 3200 installed at a remote location. Here, the mobile communication terminal 3200 is a tablet computer operated by a supervisor (including a driver) at the work travel site.
The control unit 3006 is a core element of this control system, and is shown as a collection of a plurality of ECUs. The positioning data from the satellite positioning module 3080 and the image data from the imaging unit 3070 are input to the control unit 3006 through the wired network.

The control unit 3006 includes an output processing unit 3006B and an input processing unit 3006A as an input / output interface. The output processing unit 3006B is connected to the vehicle travel device group 3007A and the work device device group 3007B. The vehicle travel device group 3007A includes control devices related to vehicle travel, such as an engine control device, a shift control device, a braking control device, and a steering control device. The work device group 3007 B includes power control devices and the like in the harvesting unit 3002, the threshing device 3011, the grain discharging device 3013, and the grain feeding device 3023.

A traveling system detection sensor group 3008A, a working system detection sensor group 3008B, and the like are connected to the input processing unit 3006A. The traveling system detection sensor group 3008A includes a sensor that detects the state of an engine speed adjuster, an accelerator pedal, a brake pedal, a gearshift operator, and the like. The work system detection sensor group 3008 B includes a sensor for detecting the state of the harvester 3002, the threshing device 3011, the grain discharging device 3013, and the device state in the grain conveying device 3023 and the state of the grain or grain.

The control unit 3006 includes a work travel control module 3060, an image recognition module 3005, a data processing module 3050, a machine position calculation unit 3066, a notification unit 3067, and a travel locus calculation unit 3068.

The notification unit 3067 generates notification data based on an instruction or the like received from each functional unit of the control unit 3006, and gives the notification data to the notification device 3091. The aircraft position calculation unit 3066 calculates the aircraft position which is the map coordinates (or field coordinates) of the aircraft 3001 based on the positioning data sequentially sent from the satellite positioning module 3080. The traveling locus calculation unit 3068 calculates a traveling locus of the vehicle body 3001 from the successive vehicle position calculated by the vehicle body position calculating unit 3066. Furthermore, the traveling locus calculation unit 3068 can add the predicted traveling locus immediately ahead of the vehicle 3001 to the traveling locus based on the latest vehicle position and the steering angle or target traveling route by the steering control device at that time. .

The combine according to this embodiment can travel in both automatic travel (automatic steering) and manual travel (manual steering). The work travel control module 3060 is provided with an automatic work travel instruction unit 3063 and a travel route setting unit 3064 in addition to the travel control unit 3061 and the work control unit 3062. A travel mode switch (not shown) is provided in the cabin 3014 to select one of an automatic travel mode for traveling by automatic steering and a manual steering mode for traveling by manual steering. By operating the travel mode switch, it is possible to shift from manual steering travel to automatic steering travel, or shift from automatic steering travel to manual steering travel.

The traveling control unit 3061 has an engine control function, a steering control function, a vehicle speed control function, and the like, and provides a traveling control signal to the vehicle traveling device group 3007A. The work control unit 3062 gives a work control signal to the work device group 3007B to control the movement of the reaper 3002, the threshing device 3011, the grain discharging device 3013, the grain feeding device 3023, and the like.

When the manual steering mode is selected, the traveling control unit 3061 generates a control signal based on the operation by the driver to control the vehicle traveling device group 3007A. When the automatic steering mode is selected, the traveling control unit 3061 controls the vehicle traveling device group 3007A related to steering and the vehicle traveling device group 3007A related to vehicle speed based on the automatic traveling instruction given by the automatic work traveling instruction unit 3063. .

The travel route setting unit 3064 develops a travel route for automatic travel created by any of the control unit 3006, the mobile communication terminal 3200, the cloud computer system 3100, and the like in a memory. The travel route developed in the memory is sequentially used as a target travel route in automatic travel. Even if this traveling route is manual traveling, it is also possible to use the guidance for the combine to travel along the traveling route.

More specifically, the automatic work travel instruction unit 3063 generates an automatic steering instruction and a vehicle speed instruction, and gives the travel control unit 3061. The automatic steering command is generated so as to eliminate the azimuth deviation and the positional deviation between the traveling route set by the traveling route setting unit 3064 and the vehicle position calculated by the machine position calculation unit 3066. The vehicle speed command is generated based on a previously set vehicle speed value. Furthermore, the automatic work travel instruction unit 3063 gives the work control unit 3062 a work device operation instruction according to the vehicle position and the traveling state of the vehicle.

The image recognition module 3005 receives image data of a photographed image sequentially acquired by the photographing unit 3070 sequentially. The image recognition module 3005 estimates an existing area in the photographed image in which the recognition target exists, and outputs recognition output data including an existing area and an estimated probability when the existing area is estimated as a recognition result. The image recognition module 3005 is constructed using neural network technology that employs deep learning.

The flow of generation of recognition output data by the image recognition module 3005 is shown in FIG. 15 and FIG. Pixel values of RGB image data are input to the image recognition module 3005 as input values. In the illustrated example of FIG. 15, the authentication objects to be estimated are weeds and fallen cereals and persons. Therefore, in the recognition output data as the recognition result, the present region of weeds (hereinafter referred to as weed region) and its estimated probability, the present region of the fallen kernel (hereinafter referred to as the fallen cereal region) and its estimated probability, It includes the existence area of a person (hereinafter referred to as a person area) and its estimated probability.

In FIG. 15, the estimation result is schematically shown, the weed area is shown by a rectangular frame with a reference F1, the fallen grain area is shown with a rectangular frame with a reference F2, and the human area is It is indicated by a rectangular frame given a code F3. Each region is linked to its estimated probability. The weed area, the fallen grain area and the human area are each defined by four corner points, but the coordinate positions on the photographed image of the four corner points of each rectangle are also included in the estimation result. Of course, if the authentication target is not estimated, the existence region of the authentication target is not output, and the estimation probability becomes zero.

The presence area of the recognition target and the estimation probability output from the image recognition module 3005 as recognition output data are important for the generation of final recognition target position information. However, in the case of shooting with a camera provided in a combine traveling on a field, the shooting is performed due to sudden shaking of the shooting unit 3070, instantaneous backlighting, dust crossing within the shooting field of the shooting unit 3070, etc. Images may be inappropriate. In that case, recognition output data having an unexpectedly low estimation probability will be output. Alternatively, the recognition target can not be estimated, and recognition output data with an estimated probability of zero is output. Therefore, the data processing module 3050 that processes the recognition output data output from the image recognition module 3005 checks the reliability of the recognition output data output from the image recognition module 3005 as a pre-processing function to recognize the unreliability A recognition output data evaluation function is provided to determine output data as inappropriate recognition output data.

This recognition output data evaluation function is realized by the data storage unit 3053 of the data processing module 3050, the data determination unit 3054 and the data correction unit 3055 as shown in FIG. The data storage unit 3053 temporarily and temporally stores recognition output data sequentially output from the image recognition module 3005 as a recognition output data sequence. The data determination unit 3054 is a range within which the existing area in the previously captured image estimated by the image recognition module 3005 should be positioned in the next captured image based on the travel locus calculated by the travel locus calculation unit 3068 (hereinafter also referred to as a prediction range). Predict). Furthermore, the data determination unit 3054 compares the estimated probability of the existing area overlapping the prediction range in the next captured image with the estimated probability of the existing area in the previous captured image. In this comparison, when the two estimated probabilities differ by more than a predetermined allowable amount, recognition output data based on the next captured image is determined as unsuitable recognition output data. Alternatively, a configuration may be adopted in which recognition output data for which the estimation probability is equal to or less than a predetermined value (for example, 0.5 or less) is determined as inappropriate recognition output data.

Data processing for determining the inappropriate recognition output data will be described below with reference to FIG. In this example, the object to be estimated is a weed.
(Step 01) The image recognition module 3005 recognizes the weed area surrounded by the rectangular frame indicated by the symbol F1 from the captured image acquired by the imaging unit 3070 (# 01a), and the estimated probability is 0.75 (# 01 b). If the estimated probability exceeds 0.7, it is considered as high confidence. Next, based on the position of the rectangular frame in the photographed image and the traveling locus calculated by the traveling locus calculation unit 3068, an expected range PA in which the weed region should be located in the photographed image to be acquired next is calculated. (# 01c).

(Step 02) Next, the image recognition module 3005 recognizes the weed area surrounded by the rectangular frame indicated by the code F1 from the photographed image acquired by the photographing unit 3070 (# 02a), and the estimated probability is It is 0.8 (# 02 b). The recognized weed areas overlap with the expected range PA. Based on the position of the rectangular frame and the travel locus, the expected range PA where the weed region should be located in the captured image to be acquired next is calculated (# 02c).

(Step 03) Further, based on the next captured image (# 03a), the image recognition module 3005 estimates a presence area where weeds are present, but the estimation probability is 0.2 (# 03b). As described above, when the estimation probability is lower than 0.5, the reliability of the position of the rectangular frame indicating the weed area also becomes low. Therefore, for calculating the prediction range PA here, it is preferable to use the position of the rectangular frame in step 02 and the traveling locus instead of using the position of the rectangular frame of the recognition output data with low reliability (# 03c).

(Step 04) Further, based on the next captured image (# 04a), the presence area where weeds are present is estimated by the image recognition module 3005, and the weed area surrounded by the rectangular frame indicated by the code F1 is It is recognized, and its estimated probability is 0.9 (# 04 b). The recognized weed areas overlap with the expected range PA.

When the image recognition module 3005 outputs the estimated output data as described above, the data determination unit 3054 lowers the estimated probability in step 03 from 0.8, which is the previous estimated probability, to 0.2. The estimated output data in step 03 is determined to be unsuitable recognition output data because the reduction amount is different than a predetermined allowable amount (for example, the change rate is 50% or more).

When the data determination unit 3054 determines that the estimated output data in step 03 is inappropriate recognition output data, the data correction unit 3055 uses the estimation probability of proper recognition output data before and after that to determine improper recognition output data. Calculate the interpolation estimation probability for Here, the arithmetic mean is used as the simplest interpolation operation, and an interpolation estimation probability of 0.85 is obtained. By replacing the estimation probability of inadequate recognition output data with the obtained interpolation estimation probability, the inadequate recognition output data can be used as proper recognition output data in the subsequent processing.

As described above, in this embodiment, when the number of recognition output data is small, the estimation probability of the recognition output data determined to be unsuitable recognition output data is interpolated and corrected by the estimation probability of the recognition output data before and after that. it can. As a result, the inappropriate recognition output data is appropriately corrected and can be used, and the number of data can be secured. Of course, when the recognition output data determined to be unsuitable recognition output data is discarded, the interpolation correction becomes unnecessary, so the data correction unit 3055 is omitted.

The data processing module 3050 of this embodiment further has a function of generating recognition target object position information from the recognition output data evaluated by the recognition output data evaluation function described above.
This function is realized by the recognition target object position information generation unit 3051 and the statistical processing unit 3052 of the data processing module 3050.

The recognition target object position information is information indicating the position of the recognition target on the map. The recognition target object position information generation unit 3051 generates recognition target object position information from the machine position at the time when the captured image is acquired and the recognition output data. The position on the map where the recognition target (weed, fallen grain moth, person) is included in the recognition output data is a rectangle that indicates the presence area of the certification object (weed area, fallen grain moth area, person area) The coordinate positions (camera coordinate positions) on the photographed image of the four corner points of are obtained by converting the coordinates on the map.

Since the imaging unit 3070 acquires captured images at predetermined time intervals, for example, 0.5 seconds, and inputs the image data to the image recognition module 3005, the image recognition module 3005 also recognizes and outputs recognition output data at the same time intervals. Output. Therefore, when the recognition target is included in the shooting field of view of the imaging unit 3070, the plurality of recognition output data includes the existence region for the same recognition target. As a result, multiple pieces of recognition target object position information for the same recognition target can be obtained. At that time, the estimation probability included in the recognition output data which is each original data, that is, the estimation probability of the existing region of the recognition target included in the recognition target object position information, is between the photographing unit 3070 and the recognition target Because the positional relationship is different, they often have different values.

Therefore, in this embodiment, as shown in FIG. 16, a plurality of such recognition target object position information is stored, and an estimated probability included in each of the stored plurality of recognition target object position information is statistically calculated. Ru. The statistical processing unit 3052 obtains a representative value of the estimated probability group by using a statistical operation on the estimated probabilities of a plurality of recognition target object position information. A plurality of pieces of recognition target object position information can be corrected to one piece of optimum recognition target object position information (recognition target object correction position information) using the representative value. One example of such correction is to obtain the arithmetic mean value or weighted mean value or median value of each estimated probability as a reference value (representative value), and obtain the logical sum of the existing regions having the estimated probability equal to or higher than the reference value It is to generate correction recognition target object position information in which the Of course, it is also possible to generate one reliable object position information with high reliability using statistical operations other than this. That is, the plurality of recognition target object position information is corrected based on the result of the statistical calculation of the estimation probability included in the recognition output data corresponding to the recognition target object position information.

Recognition object position information (weed position information, fallen grain weir position information, person position) indicating the position on the map of the existence area (weed area, fallen grain weir area, person area) of the certification object thus obtained By using the information), at the time of recognition of weeds, fallen grain weirs, and persons, traveling work control and warning notification set in advance are performed.

Note that each recognition target object position information generated by the recognition target object position information generation unit 3051 can be mapped as shown in FIG. 18 for display that is easy to understand visually. FIG. 18 exemplifies a weed map in which weed position information is mapped. If the weed location information includes a weed presence region having a different estimation probability, it is possible to display the weed presence region in a form divided in a predetermined range of the estimation probability value, as shown in FIG. .

The configurations disclosed in the above-described embodiments (including the other embodiments, the same applies hereinafter) can be applied in combination with the configurations disclosed in the other embodiments as long as no contradiction arises. The embodiment disclosed in the present specification is an exemplification, and the embodiment of the present invention is not limited to this, and can be appropriately modified within the scope of the object of the present invention.

[Another embodiment]
(1) In the embodiment described above, the weed group which is stretched higher than the planted grain weir is set as the recognition object to be recognized by the image recognition module (5, 2005, 3005). , Falling cereals or people may also be set. At that time, the work travel control module (60, 2060, 3060) is configured to perform the necessary control in response to the recognition of the fallen cereal group or person.
(2) In the embodiment described above, the image recognition module (5, 2005, 3005) is constructed using deep learning type neural network technology. Alternatively, an image recognition module (5, 2005, 3005) constructed using other machine learning techniques may be employed.
(3) In the embodiment described above, the image recognition module (5, 2005, 3005), the data processing module (50, 2050, 3050), and the weed position calculation unit 68 correspond to the control unit (6, 2006, 3006) of the combine. Although incorporated, some or all of them can be configured in a control unit independent of the combine, such as a mobile communication terminal (200, 2200, 3200).
(4) The functional units shown in FIG. 4, FIG. 9, FIG. 14 and FIG. 16 are divided mainly for the purpose of explanation. In practice, each functional unit may be integrated with another functional unit or may be further divided into a plurality of functional units.

The present invention is applicable not only to combine harvesters such as rice and wheat, but also to combine harvesters that harvest other crops such as corn and harvesters that harvest carrots and the like.

3: Handle depth adjustment mechanism 5: Image recognition module 6: Control unit 23: Grain conveyor device 236: Electric motor for adjustment (motor)
237: operation rod 30: length detection device 34: detection switch 35: detection switch 50: data processing module 51: weed position information generation unit 52: statistical processing unit 60: work traveling control module (work traveling control unit)
61: traveling control unit 63: automatic work traveling instruction unit 620: grain depth control unit 66: machine position calculating unit 68: weed position calculating unit 70: imaging unit 80: satellite positioning module 91: notification device 92: communication unit 2053: Data storage unit 2054: data determination unit 2055: data correction unit 2051: recognition target object position information generation unit

Claims (13)

1. It is a harvester that harvests cropped rice husk while traveling on the field,
   A reaper section for reaping the crop from the field;
   A grain conveying device for conveying a reaper from the reaper to a threshing device;
   A handling depth adjustment mechanism provided in the cereal gutter transport device;
   A handling depth control unit that performs handling depth adjustment control based on the length of the reaping grain using the handling depth adjustment mechanism;
   An aircraft position calculation unit that calculates an aircraft position, which is map coordinates of an aircraft, based on positioning data from a satellite positioning module;
   An imaging unit provided on the machine body and imaging the field during harvesting work;
   Image recognition in which image data of a captured image sequentially acquired sequentially by the imaging unit is input, a weed growth region in the captured image is estimated, and recognition output data indicating the estimated weed growth region is output Modules,
   A weed position information generation unit that generates weed position information indicating a position on the map of the weed growth area from the position of the body at the time when the photographed image is acquired and the recognition output data;
   A work travel control unit that executes weed entry control while the weeds cut in the weed growth region pass the threshing depth adjustment mechanism and the weeds pass through the threshing depth adjustment mechanism;
   Harvester equipped with.
2. It is a harvester that harvests cropped rice husk while traveling on the field,
   A reaper section for reaping the crop from the field;
   A grain conveying device for conveying a reaper from the reaper to a threshing device;
   A handling depth adjustment mechanism provided in the cereal gutter transport device;
   A handling depth control unit that performs handling depth adjustment control based on the length of the reaping grain using the handling depth adjustment mechanism;
   An aircraft position calculation unit that calculates an aircraft position, which is map coordinates of an aircraft, based on positioning data from a satellite positioning module;
   An imaging unit provided on the machine body and imaging the field during harvesting work;
   Image recognition in which image data of a captured image sequentially acquired sequentially by the imaging unit is input, a weed growth region in the captured image is estimated, and recognition output data indicating the estimated weed growth region is output Modules,
   A weed position information generation unit that generates weed position information indicating a position on the map of the weed growth area from the position of the body at the time when the photographed image is acquired and the recognition output data;
   A work traveling control unit that executes weed entry control while the mowing unit passes through the weed growth area;
   Harvester equipped with.
3. The harvester according to claim 2, wherein the timing at which the harvesting section passes the weed growth area is determined based on the weed position information and the machine position calculated by the machine position calculation unit.
4. The harvester according to any one of claims 1 to 3, wherein the depth adjustment control is interrupted by execution of the weed entry control.
5. The harvester according to any one of claims 1 to 3, wherein the vehicle speed is reduced by the execution of the weed entry control.
6. It is a harvester that harvests crops while traveling in the field,
   An aircraft position calculation unit that calculates an aircraft position, which is map coordinates of an aircraft, based on positioning data from a satellite positioning module;
   An imaging unit provided on the airframe for imaging a field during harvesting work;
   When the image data of the captured image sequentially acquired sequentially by the imaging unit is input, the existing region where the recognition target exists in the captured image is estimated, and the existing region and the existing region are estimated An image recognition module that outputs recognition output data including an estimated probability of
   A recognition target position information generation unit for generating recognition target position information indicating the position on the map of the recognition target from the position of the machine at the time when the photographed image is acquired and the recognition output data; Harvester.
7. The harvester according to claim 6, wherein the estimation probability of the recognition object is reduced as the recognition object is located farther from the imaging unit in the photographed image.
8. A plurality of pieces of recognition target object position information are stored, and the plurality of pieces of recognition target object position information stored are corrected based on the result of statistical calculation of the estimated probability included in the corresponding recognition output data. A harvester according to claim 6 or 7.
9. A data storage unit that temporarily and temporally stores the recognition output data sequentially output from the image recognition module as a recognition output data sequence; and a data determination unit.
   The data determination unit improperly recognizes and outputs the recognition output data having the estimated probability lower than a predetermined level as compared to the estimated probability of the recognition output data sequentially related to each other in the recognition output data sequence. The harvester according to any one of claims 6 to 8, which is determined as data.
10. The harvester according to any one of claims 6 to 9, wherein the object to be recognized is at least one of a person, a fallen grain weir, a weed, and a weir.
11. It is a harvester that harvests crops while traveling in the field,
    An aircraft position calculation unit that calculates an aircraft position, which is map coordinates of an aircraft, based on positioning data from a satellite positioning module;
    A travel locus calculation unit that calculates a travel locus of the vehicle from the position of the vehicle;
    An imaging unit provided on the airframe for imaging a field during harvesting work;
    When the image data of the captured image sequentially acquired sequentially by the imaging unit is input, the existing region where the recognition target exists in the captured image is estimated, and the existing region and the existing region are estimated An image recognition module that outputs recognition output data including an estimated probability of
    Based on the traveling locus, a range where the existing area in the previous captured image is to be located in the next captured image is predicted, and the estimated probability of the existing area overlapping the range in the next captured image, and the previous captured image A data determination unit that determines the recognition output data based on the next captured image as improper recognition output data when the estimated probability of the existing area in the area is different by a predetermined allowable amount or more;
    Harvester equipped with.
12. A data correction unit is provided which replaces the estimated probability of the inappropriate recognition output data with an interpolation estimation probability obtained based on the estimated probability of the recognition output data that is sequentially before and after the inappropriate recognition output data. The harvester according to claim 11.
13. The harvester according to claim 11 or 12, wherein the object to be recognized is at least one of a person, a fallen grain weir, a weed, and a weir.
    

Images