WO2024105405A1 - Chariot de cueillette de fruits - Google Patents

Chariot de cueillette de fruits Download PDF

Info

Publication number
WO2024105405A1
WO2024105405A1 PCT/GB2023/053011 GB2023053011W WO2024105405A1 WO 2024105405 A1 WO2024105405 A1 WO 2024105405A1 GB 2023053011 W GB2023053011 W GB 2023053011W WO 2024105405 A1 WO2024105405 A1 WO 2024105405A1
Authority
WO
WIPO (PCT)
Prior art keywords
trolley
picking trolley
picking
fruit
sensors
Prior art date
Application number
PCT/GB2023/053011
Other languages
English (en)
Inventor
Edward Herbert
Mathew Cook
Duncan Robertson
Francis TULLY
Thomas Dean
Julian STORTT
Anders Johansson
Robert KARPINSKI
Chris PADBURY
Original Assignee
Dogtooth Technologies Limited
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Priority claimed from GBGB2217136.7A external-priority patent/GB202217136D0/en
Priority claimed from GBGB2301007.7A external-priority patent/GB202301007D0/en
Application filed by Dogtooth Technologies Limited filed Critical Dogtooth Technologies Limited
Publication of WO2024105405A1 publication Critical patent/WO2024105405A1/fr

Links

Classifications

    • A—HUMAN NECESSITIES
    • A01—AGRICULTURE; FORESTRY; ANIMAL HUSBANDRY; HUNTING; TRAPPING; FISHING
    • A01D—HARVESTING; MOWING
    • A01D46/00—Picking of fruits, vegetables, hops, or the like; Devices for shaking trees or shrubs
    • A01D46/30—Robotic devices for individually picking crops
    • A—HUMAN NECESSITIES
    • A01—AGRICULTURE; FORESTRY; ANIMAL HUSBANDRY; HUNTING; TRAPPING; FISHING
    • A01D—HARVESTING; MOWING
    • A01D46/00—Picking of fruits, vegetables, hops, or the like; Devices for shaking trees or shrubs

Definitions

  • the field of the invention relates to a fruit picking trolley or robot.
  • a first aspect is a fruit, or other crop, picking trolley configured to include trays or other containers for holding picked fruit or crops, in which the trolley includes an electric traction motor system and/or is configured to be moved manually by an operator; and in which the trolley is instrumented with one or more sensors.
  • the trolley can include a sensor mounting system to which different sensors, or multiple combinations of sensors, can be attached.
  • the sensor mounting system enables different sensors to be connected to common resources, such as a power supply, data storage and data ports.
  • the picking trolley may also include any or more of the following features:
  • the picking trolley is instrumented with one or more sensors capable of gathering data about crop status or condition.
  • the picking trolley is equipped with sensors that provide data to a crop yield monitoring and/or forecasting system.
  • the picking trolley is instrumented with one or more cameras (e.g. stereo cameras and depth sensing cameras) capable of determining the number or proportion of fruits at different stages of maturity.
  • cameras e.g. stereo cameras and depth sensing cameras
  • the picking trolley is instrumented with one or more sensors capable of recording information about the status of crop status or condition in the course of normal use of the trolley for a different purpose such as carrying picked fruit.
  • the picking trolley is instrumented with one or more sensors capable of providing useful information about the growing environment, such as temperature, humidity, incident light levels.
  • the picking trolley is instrumented with one or more cameras (e.g. stereo cameras and depth sensing cameras) capable of determining the amount of foliage.
  • the picking trolley is equipped with weight sensors or scales to keep track of how much fruit the trolley contains, either as a whole or per tray or per individual punnet.
  • the picking trolley is equipped with weight sensors or scales and a feedback system configured to give an indication to the operator of the total weight of fruit or the weight of fruit in each punnet, as an aid to filling punnets to a requisite minimum weight.
  • the picking trolley is instrumented with sensors so that it is capable of real time measurement of both the weight of fruit picked and its position, e.g. for a particular section of a particular crop row.
  • the picking trolley is equipped with separate weight sensors or scales for weighing damaged and saleable fruit picked, so that wastage as well as yield can be determined or estimated.
  • the picking trolley is instrumented with one or more sensors capable of recording information about the status of farm infrastructure.
  • the picking trolley is instrumented with one or more sensors that generate data is geo-referenced.
  • the picking trolley is equipped with a position sensor for determining its position in real time (such as GPS), to give an indication to the user of the need for a particular action when it reaches a particular position and/or to relate other sensor measurements to particular locations on a farm (such as crop rows).
  • a position sensor for determining its position in real time (such as GPS), to give an indication to the user of the need for a particular action when it reaches a particular position and/or to relate other sensor measurements to particular locations on a farm (such as crop rows).
  • the picking trolley is equipped with a position sensor (e.g. optical encoder or GPS receiver) to allow tracking of distance travelled vs. time, e.g. to log information about the movement of the trolley, to give real time feedback to the operator, and/or to allow geo-referencing of data captured by other sensors.
  • a position sensor e.g. optical encoder or GPS receiver
  • the picking trolley is instrumented with one or more sensors that generate data are time-referenced.
  • the picking trolley is configured to warn the user when the oldest picked fruit in the trolley is older than a configurable time limit.
  • the picking trolley and/or operator is associated with a unique ID so that captured data can be associated with that specific trolley and/or operator.
  • the picking trolley is instrumented with a data capture module capable of recording information about the productivity of individual users.
  • the picking trolley is instrumented with one or more sensors, e.g. cameras, that are packaged for attachment to the trolley within a removable sensor module.
  • sensors e.g. cameras
  • the picking trolley configured with removable actuators, for example actuators capable of applying pesticide where needed, or releasing insects for pollination or pest control.
  • the picking trolley may also include any one or more of the features of the second aspect of the invention.
  • a second aspect of the invention is a fruit or other crop picking trolley configured to include trays or other containers for holding picked fruit or crops, in which the trolley includes an electric traction motor system configured to enable hands-free movement of the trolley and is also configured to be manually moved by an operator.
  • the picking trolley may also include any or more of the following features:
  • the picking trolley is equipped with a padded lever in the handle to control the speed of the traction motor.
  • the picking trolley is configured to move at a pre-set speed or with a pre-set distance from the operator.
  • the picking trolley is configured to move automatically along a row or path to avoid colliding with objects at the side of the path (e.g. plants).
  • the picking trolley may also include any one or more of the features of the first aspect of the invention.
  • a third aspect of the invention is a fruit, or other crop, autonomous picking robot, the robot being configured to navigate autonomously using a computer vision guidance system, in which the robot is instrumented with one or more sensors capable of gathering data about crop status or condition.
  • the picking robot may also include any one or more of the features defined above. BRIEF DESCRIPTION OF THE FIGURES
  • Figure 1A is a perspective view of a motorized trolley, filled with trays.
  • Figure IB is a top view of a motorized trolley, filled with trays.
  • Figure 1C is a perspective view of a motorized trolley, filled with trays.
  • Figure ID is a front view of a motorized trolley, filled with trays.
  • Figure IE is a side view of a motorized trolley, filled with trays.
  • Figure IF is a back view of a motorized trolley, filled with trays.
  • Figure 1G is a perspective view of a motorized trolley, filled with trays.
  • Figure 1H is a bottom view of a motorized trolley, filled with trays.
  • Figure II is a perspective view of a motorized trolley, filled with trays.
  • Figure 2A is a perspective view of a motorized trolley, shown empty.
  • Figure 2B is a top view of a motorized trolley, shown empty.
  • Figure 2C is a perspective view of a motorized trolley, shown empty.
  • Figure 2D is a front view of a motorized trolley, shown empty.
  • Figure 2E is a side view of a motorized trolley, shown empty.
  • Figure 2F is a back view of a motorized trolley, shown empty.
  • Figure 2G is a perspective view of a motorized trolley, shown empty.
  • Figure 2H is a bottom view of a motorized trolley, shown empty.
  • Figure 21 is a perspective view of a motorized trolley, shown empty.
  • This physical effort required to move a crop (e.g fruit) storage trolley motivates the first feature of one implementation of the invention described here, which is a motorized trolley that can move under its own power. Since fruit pickers generally need both hands for picking, the trolley can be moved hands-free whist the operator is picking, positioning itself just in front of the picker and steering along the crop row.
  • growers can benefit from a detailed understanding of the productivity of individual operators/workers, e.g. so as to be able to give more meaningful feedback to workers concerning their productivity.
  • a related feature is to instrument trolleys or robots with sensors capable of gathering data about crop condition for yield monitoring and/or forecasting applications.
  • the autonomous picking robot is configured to navigate autonomously using a computer vision guidance system and may be instrumented with one or more sensors or one or more actuators, as defined below.
  • picking trolley in the description below may also refer to a picking robot.
  • Figures 1A to II show different views of a motorized trolley 10, equipped with multiple trays for storing picked fruit.
  • Figures 2A to 21 show different views of a motorized trolley 10, empty without trays. In the filled configuration shown, the trolley securely holds multiple stacked trays that are organised in two vertical rows. The trays can be easily loaded onto or removed from the trolley, making it adaptable for various applications beyond fruit transportation.
  • the trolley may be equipped with an electrical motor and a rechargeable battery, providing efficient movement.
  • the electrical motor and power source may be integrated into the trolley’s frame or may be integrated inside an easily removable electronics module 12.
  • the motorized trolley is equipped with a padded lever 13, allowing for hip- controlled operation by the user.
  • the padded lever 13 may be located at the front of the trolley.
  • Another lever may also be located at the rear of the trolley, enabling effortless movement in any direction without the need to turn the trolley around, thus minimizing any unnecessary effort.
  • the user can start, stop and/or control the speed of the trolley’s movement. This significantly reduces user fatigue and strain.
  • the trolley may also feature wheels and a suspension system for smooth movement over uneven terrain.
  • the motorized trolley can remain at a desired distance from the operator as both move along the row. If the trolley is too close, then it is in the way; if it is too far then it may be out of reach.
  • Various control systems may be employed to ensure that the trolley remains at the ideal distance from the picker.
  • One simple approach is to have the trolley move at (possibly user configurable) constant speed.
  • Another is to equip the trolley with a switch or pressure sensor designed to be actuated by some part of the body, e.g. the hip of a fruit picking worker using his or her hands for picking. This switch or sensor can provide a control signal to the motors so that the worker can cause the trolley to advance along the crop row by simply moving closer to it.
  • Another approach is for the trolley to be attached to the picker via an elastic cord. When the tension in the cord exceeds some threshold, the trolley could be made to move further along the row until the tension is less than the threshold.
  • Contactless methods of measuring distance such as ultrasonic proximity sensors can also be used as the basis of a suitable control system.
  • a feature is to allow the user to select or configure a desired velocity profile.
  • this velocity profile would typically define how quickly the trolley accelerates when the switch is actuated, its maximum velocity, and how quickly the trailer decelerates when the switch is no longer actuated.
  • the acceleration and deceleration might be computed as functions of the variable signal.
  • the motorized trolley should be capable of steering (or being easily steered) along the row so as to avoid collision, e.g. with the crop or the infrastructure that supports it (tabletops, trellises, raised beds, etc). This might be achieved by having the picker control the direction of travel of the motorized trolley by applying force directly, e.g. using the hands, or using the hips so as to leave hands free for picking. In practice however, such solutions may be cumbersome, particularly if the need for frequent steering inputs reduces the worker’s efficiency, e.g. by breaking the worker’s concentration. A more sophisticated solution might allow the trolley to steer autonomously or quasi-autonomously along the row (while still maintaining a desired distance from the operator).
  • Autonomous or quasi-autonomous steering is beneficial because it allows the worker to concentrate on the task at hand (such as fruit picking) instead of on steering.
  • Autonomous steering might be achieved e.g. by having the motorized trolley do one or more of the following: i) follow a nylon guide rope; ii) follow a buried metallic cable, possibly carrying an electric current; iii) use a computer vision system to determine its orientation and/or position with respect to the row; iv) using a computer vision system to determine its orientation and/or position with respect to easily visually identifiable targets; v) using active beacons emitting EM radiation (e.g. light) or sound to determines it’s position by triangulation or otherwise; vi) using another type of sensor to determine position and/or heading, e.g.
  • LIDAR LIDAR
  • IMU intial measurement unit
  • the IMU may also be use for one or more of the following:
  • the powered trolley can also be controlled manually, i.e. that its heading and possibly its speed can be controlled entirely by the user. This may be facilitated by a simple mode selection switch that allows the trolley to be steered either automatically (along a crop row) or manually.
  • Another feature is to allow the trolley to return automatically to the end of the crop row or other location for loading or unloading. This idea may be useful if the worker can continue to perform useful work in the absence of the trolley.
  • the motorized trolley might usefully be equipped with an emergency stop button, designed to ensure that any movement of the trolley can be arrested even in the event of a fault.
  • the emergency stop button might be designed to be actuated via an elastic lanyard attached to some part of the operator, e.g. the belt. By this means it is possible to guarantee that a faulty trolley cannot drive far away from the operator, e.g. into a road.
  • the motorized trolley might be powered in various ways, e.g. using Lilon batteries.
  • a related idea is to ensure that the valuable parts of the motorized trolley (e.g. the electronic control system) are also included within the removeable module so that only less valuable parts of the system need to be left outside - and so that a stolen trolley without a battery module cannot be made operable by simply replacing a battery.
  • the valuable parts of the motorized trolley e.g. the electronic control system
  • a feature is to equip trolleys with sensors that can be used to record information about the status of the crop or farm infrastructure, typically in the course of normal use of the trolley for a different purpose such as carrying picked fruit.
  • sensors can provide useful information about the condition of the crop or the growing environment or farm infrastructure or the productivity of the trolley operator.
  • cameras including stereo cameras and depth sensing cameras
  • sensors can be used to determine the number or proportion of fruits at different stages of maturity (e.g., developing strawberries are sometimes classified in order of increasing maturity as small green, large green, large white, pink, ripe), or the amount of foliage
  • temperature and humidity sensors might be used to record information about growing conditions.
  • GPS position information might be used to measure the rate of progress along a row or to geo-reference other sensor readings to a particular part of the farm, e.g. a particular part of a particular crop row.
  • Another way of geo- referencing sensor readings to a specific crop row or part of crop row might be to instrument the trolley with a sensor capable of reading an identifier attached to the row infrastructure, e.g. an RFID tag or a QR code.
  • camera calibration devices such as grey cards or checkerboards may be mounted on the sensor module as to be visible by the camera modules. This would allow continuous calibration of the cameras post-manufacture.
  • actuators for example actuators capable of applying pesticide where needed, or releasing insects for pollination or pest control.
  • One scheme for mounting the sensor module to the trolley is to mount a bracket designed to support the sensor module permanently to the trolley and to allow the sensor model conveniently to be attached to or detached from the bracket as necessary, e.g. via a quick release handle.
  • the bracket might also incorporate a mechanism for adjusting the height of the sensor module when attached to the trolley. This allows for application to a wide variety of physical contexts and the ability for fine-tuning post-deployment such as to align cameras for a variety of crop heights.
  • the trolley is a motorized trolley, it may be beneficial to make provision for the sensors to be powered via the same power source as the trolley, e.g. Lilon batteries.
  • Suitable software running on a central server might recover data from the trolley’s computer whenever it is visible on a local network, freeing up storage for future data. It may be convenient to progress data offload when the trolley’s batteries are being charged, particularly if batteries are charged in a central location.
  • Another feature is to associate each trolley and/or operator with a unique ID so that captured data can be associated with that specific trolley, operator, etc.
  • This association may be recorded by the operator interacting with the unit’s sensors such as a camera taking a timestamped image of an ID, RFID reader recording from an ID card, or by using a touchscreen, keyboard, or keypad to enter the ID.
  • the trolley may be equipped with position sensor (optical encoder or GPS receiver) to allow tracking of distance travelled vs. time.
  • This sensor might be used both to log information about the movement of the trolley or to give real time feedback to the worker.
  • It might also be equipped with scales to keep track of how much fruit it contains, either as a whole or per tray or per individual punnet.
  • the scales might be used to give an indication to the worker of the total weight of fruit or the weight of fruit in each punnet as an aid to filling punnets to a requisite minimum weight.
  • the data produced by such a combination of sensors could be displayed in software user interface intended to allow the farm manager to gain insight into the factors that affect the productivity of the workforce, e.g. : to compare the productivity of one worker with another; to compare workers on the basis of speed of travel (e.g. meters per second) and extraction rate (how much fruit is harvested per meter); to compare workers on the basis of how much waste fruit they are picking; to see how a workers’ productivity decreases during the course of the day.
  • a useful idea is to associate the operator with a code that uniquely identifies him or her and to ensure that the productivity data captured by the trolley is associated with the code. This might be achieved by having the operator or supervisor enter a code via a key pad or by presenting an RFID tag or printed bar code to a suitable reader.
  • a further feature arising from the general idea of instrumenting trolleys with sensors is the possibility of using trolleys as the basis of a crop yield monitoring and/or forecasting solution.
  • Growers are interested to monitor crop yield across the farm because a reduction in yield can indicate a problem with the health of the crop - such as mildew or SWD infestation - and earlier indication of such problems allows more effective interventions, such as the application of fungicides or pesticides, or change to environmental conditions.
  • growers must often try to predict future crop yields a week or two in advance in order to understand their scope for making supply contracts with their customers, e.g. supermarkets.
  • Predicting crop yields also helps growers to make more efficient use of their human workforce, e.g. to employ the right number of workers to harvest the crop or to inform pickers about how conservative they should be with respect to the judgement of ripeness.
  • the yield for a particular area of the farm is estimated by measuring the total weight of fruit picked from that area.
  • a nuanced picture may be obtained by an instrumented trolley that is capable of real time measurement of both the weight of fruit picked so far and its position (e.g. crop row and distance travelled along the row). This would allow yield to be measured for a particular section of a particular crop row as the difference between the weight of fruit that had been picked when the picker reached the start of that section and the weight of fruit that had been picked when the picker reached the end of that section.
  • the benefit is that this allows growers to monitor yield as a function of location with much higher locational precision. For some crops, it is important that pickers should pick damaged fruit (e.g.
  • the trolley might usefully be equipped with separate scales for weighing damaged and saleable fruit picked. By this means, wastage as well as yield can be estimated.
  • One way predicting future yield is to count berries at different stages of maturity (e.g. developing strawberries are sometimes classified in order of increasing maturity as small green, large green, large white, pink, ripe).
  • future yields for desired timescales may be estimated from (i) counts of the number of berries at each stage of maturity and (ii) a weather forecast.
  • model inputs include one or more of the following:
  • model outputs include one or more of the following:
  • the models have a range of ‘observation_windows’ and Tead_windows’ and the forecasting script chooses based on what's currently available in the dataset to allow for the maximum ‘observation_window’ and/or minimum Tead_window’ for forecasting a particular day.
  • Such berry counts can be obtained by recording information manually about the condition of a selected sample of individual plants, but manual counting is time consuming and error prone. This gives rise to the idea that picking trolleys might be instrumented with computer vision systems capable of identifying and counting crops (e.g. berries) at various stages of maturity.
  • crops e.g. berries
  • Such a computer vision system might contain one or more cameras (including stereo cameras), a means of determining position (such as a GPS- or IMU-based system), a means of capturing images at controlled intervals in space or time, and a means of storing captured images for later analysis (or for analysing captured images in situ).
  • a suitable solution to the problem of detecting berries in images is to use e.g. a neural network trained to perform berry instance segmentation using training data labelled by human judges.
  • a suitable classifier may also be used to judge ripeness.
  • One approach is to allow the grower to provided training examples of ripe and unripe berries so as to capture subjective preferences.
  • a multi-view stereo camera system For the purpose of determining berry mass, it may be desirable to use a multi-view stereo camera system to facilitated the determination of the 3D position of berries in a convenient coordinate frame. Under the assumption that berry shape is approximately described by a volume of revolution (or similar), knowledge of the depth of a berry in the camera frame allows trivial estimation of its volume from its projected image area. Furthermore, knowledge of berry density can then be used to estimate its mass from its volume.
  • An alternative method is to normalise berry counts by the width (or area) of the crop row observed during each image capture (to generate a berry count per meter or per meter squared).
  • the observed width or area can be determined from stereo images by trigonometry using e.g. median or mean berry distance (from berries observed in the target row during the individual image capture) and the camera field of view.
  • a difficulty in obtaining accurate berry observations from computer vision is ensuring the images used are of the target row and not an adjacent row or other part of the farm.
  • a suitable trained classifier may be used to judge whether a particular image was captured by a trolley located in an aisle.
  • the distance of berries calculated by stereo cameras can be used to filter for the known range of acceptable distances from the farm’s layout.
  • Another difficulty inherent in the use of computer vision for berry counting is the problem of undercounting berries due to occlusion.
  • berries are occluded e.g. by other berries or foliage or because they are outside of the camera field of view.
  • the proportion of fruit that is occluded may vary from day to day, e.g. because ripe berries are picked off during the course of the harvest window or because foliage is thinned as part of crop husbandry.
  • Occlusion may mean that berries are systematically undercounted by a computer vision system.
  • one benefit in deploying the computer vision system is that an estimate of the fraction of berries that have been missed can be estimated using the mass of fruit actually picked by the trolley operator.
  • the computer vision system therefore monitors crop condition before the crop is harvested, whether by robot or human.
  • the computer vision system may be deployed either on a trolley or on a robot being actively used for picking.
  • the computer vision system may not necessarily be mounted to the same vehicle (such as trolley or robot) used to store picked fruit as long as: (i) the computer vision system records the amount of ripe fruit before picking and (ii) the amount of picked fruit is used to normalize the output of the computer vision system.
  • a worker will be required to perform certain actions sporadically as he or she moves around the crop. For example, the worker may be required:
  • a picking trolley is equipped with a means of determining its position in real time (such as GPS), it might usefully give indication to the user of the need for a particular action when it reaches a particular position, e.g. via visible or audible notification.
  • Another feature is to warn the user when the oldest picked fruit in a trolley is older than some configurable time limit - since often fruit must be removed from the field e.g. for refrigeration within some time limit in order to preserve shelf life.
  • the age of the oldest fruit in the trolley might be determine by starting a timer when the weight of fruit in the trolley starts to increase, or in response to a user input such as pressing a button.
  • Another feature is to equip the trolley with indicators designed to give the worker real time feedback about his or her productivity, e.g. in meters travelled per minute (as determined by on-board GPS or other sensor) or mass of fruit harvested per hour (as determined by built-in scales).
  • One approach is to show the worker how his or her productivity compares with a target value (for example the average productivity measured over a team of workers picking in similar crop conditions) for example by illuminating a color-coded indicator according to whether the worker is on target, ahead of target, or behind target. This approach can help workers understand and improve their productivity.
  • Indicators can also be used to show when a break might be desirable, for example when the worker’s productivity has recently decreased compared to the day’s average.
  • the targets for individual workers might be set by the supervisor or farm manager via a suitable user interface.
  • the worker’s supervisor might be equipped with a tablet capable of reading displaying productivity data captured by the trolley (perhaps accessed via WiFi or Blutooth network connection to the trolley).
  • a fruit, or other crop, picking trolley configured to include trays or other containers for holding picked fruit or crops, in which the trolley includes an electric traction motor system and/or is configured to be moved manually by an operator; and in which the trolley is instrumented with one or more sensors.
  • the trolley can include a sensor mounting system to which different sensors, or multiple combinations of sensors, can be attached.
  • the sensor mounting system enables different sensors to be connected to common resources, such as a power supply, data storage and data ports.
  • a fruit or other crop picking trolley configured to include trays or other containers for holding picked fruit or crops, in which the trolley includes an electric traction motor system configured to enable hands-free movement of the trolley and is also configured to be manually moved by an operator.
  • a fruit, or other crop, autonomous picking robot the robot being configured to navigate autonomously using a computer vision guidance system, in which the robot is instrumented with one or more sensors capable of gathering data about crop status or conditions.
  • Optional features may include one or more of the following: Sensor features
  • the picking trolley is instrumented with one or more sensors capable of gathering data about crop status or condition.
  • the picking trolley is equipped with sensors that provide data to a crop yield monitoring and/or forecasting system.
  • the picking trolley is instrumented with one or more cameras (e.g. stereo cameras and depth sensing cameras) capable of determining the number or proportion of fruits at different stages of maturity.
  • cameras e.g. stereo cameras and depth sensing cameras
  • the picking trolley is instrumented with one or more sensors capable of recording information about the status of crop status or condition in the course of normal use of the trolley for a different purpose such as carrying picked fruit.
  • the picking trolley is instrumented with one or more sensors capable of providing useful information about the growing environment, such as temperature, humidity, incident light levels.
  • the picking trolley is instrumented with one or more cameras (e.g. stereo cameras and depth sensing cameras) capable of determining the amount of foliage.
  • cameras e.g. stereo cameras and depth sensing cameras
  • the picking trolley is equipped with weight sensors or scales to keep track of how much fruit the trolley contains, either as a whole or per tray or per individual punnet.
  • the picking trolley is equipped with weight sensors or scales and a feedback system configured to give an indication to the operator of the total weight of fruit or the weight of fruit in each punnet, as an aid to filling punnets to a requisite minimum weight.
  • the picking trolley is instrumented with sensors so that it is capable of real time measurement of both the weight of fruit picked and its position, e.g. for a particular section of a particular crop row.
  • the picking trolley is equipped with separate weight sensors or scales for weighing damaged and saleable fruit picked, so that wastage as well as yield can be determined or estimated.
  • the picking trolley is instrumented with one or more sensors capable of recording information about the status of farm infrastructure.
  • the picking trolley is instrumented with one or more sensors that generate data is geo-referenced.
  • the picking trolley is equipped with a position sensor for determining its position in real time (such as GPS), to give an indication to the user of the need for a particular action when it reaches a particular position and/or to relate other sensor measurements to particular locations on a farm (such as crop rows).
  • the picking trolley is equipped with a position sensor (e.g. optical encoder or GPS receiver) to allow tracking of distance travelled vs. time, e.g. to log information about the movement of the trolley, to give real time feedback to the operator, and/or to allow geo-referencing of data captured by other sensors.
  • a position sensor e.g. optical encoder or GPS receiver
  • the picking trolley is instrumented with one or more sensors that are capable of real-time measurement of the distance between the trolley and the operator and/or between the trolley and any surrounding objects.
  • the picking trolley is instrumented with one or more sensors that generate data is time-referenced.
  • the picking trolley is instrumented with a reader that is capable of reading an identifier attached to a row infrastructure, such as an RFID tag or QR code.
  • the picking trolley is configured to warn the user when the oldest picked fruit in the trolley is older than a configurable time limit.
  • the picking trolley and/or operator is associated with a unique ID so that captured data can be associated with that specific trolley and/or operator.
  • the picking trolley is instrumented with a data capture module capable of recording information about the productivity of individual users.
  • the picking trolley is instrumented with one or more sensors, e.g. cameras, that are packaged for attachment to the trolley within a removable sensor module.
  • sensors e.g. cameras
  • the picking trolley configured with removable actuators, for example actuators capable of applying pesticide where needed, or releasing insects for pollination or pest control.
  • the picking trolley is equipped with a padded lever in the handle to control the speed of the traction motor.
  • the picking trolley is configured to move at a pre-set speed or with a pre-set distance from the operator.
  • the picking trolley is configured to move automatically along a row or path to avoid colliding with objects at the side of the path (e.g. plants).
  • the picking trolley includes an actuator system that initiates movement of the trolley, in which the actuator system includes one or more of the following: a button, a touch sensitive interface or voice command system.
  • actuator system is pressure sensitive, and the movement of the picking trolley responds to the applied pressure.
  • the picking trolley is attached to the operator using a flexible connecting element, such as an elastic cord, and the trolley is configured to move or stop in response to the tension in the flexible connecting element.
  • a flexible connecting element such as an elastic cord
  • the picking trolley is configured to move if the tension in the flexible connecting element exceeds a predefined threshold and to stop when the tension falls below the predefined threshold.
  • a computer vision system is used to classify fruits at different stages of maturity, based on a model of how fruits vary as a function of time and/or environment conditions.
  • a neural network or other machine learning system trained from a database of images provided by a grower, is used.
  • Computer vision system estimates the fruit’s volume and/or weight using a projected image area and the depth of the item in the camera frame.
  • the computer vision system samples non-overlapping images (or stereo images) of a scene and apply a scale factor to the resulting fruit counts reflecting the inverse of the fraction of a row that is covered by the images.
  • the picking trolley is instrumented with sensors and a feedback system configured to give an indication to the operator of number of fruits that have been missed by the operator.
  • the picking trolley includes a computer vision system that outputs a probability of fruit that have been missed by the operator, in which the probability of fruit that have been missed by the operator Pmiss (due to occlusion or otherwise) is calculated as the function of the estimated weight of the fruits observed by the one or more cameras and the weight of the picked fruits placed inside the trolley.
  • feedback system includes real-time recommendations on several functions to be performed at specific crop location, including one or more of the following: applying suitable herbicides or pesticides, releasing insects for pollination or pest control, performing crop husbandry tasks, where needed.
  • real-time recommendations include visible and/or audible notifications.
  • the picking trolley includes or is connected to a user interface that displays productivity per user or per trolley including one or more of the following: meters travelled per minute, extraction rate such as weight of fruits harvested per hour or per meter, quantity or weight of waste fruits, number or probability of fruits that have been missed.
  • user interface outputs a visual indicator that changes based on the current productivity in relation to a pre-defined target.
  • user interface outputs a visual indicator that indicates when a break is desirable when the user’s productivity is below a certain threshold or when the user’s productivity has decreased compared to the day’s average.
  • the picking trolley includes a communication module for wirelessly transmitting picking data to and receiving data from a central control unit or other trolleys, thereby facilitating coordination with other trolleys.
  • the communication module includes a network adaptor, such as WIH or Ethernet, to enable download of picking data on a local network.
  • a network adaptor such as WIH or Ethernet
  • the picking trolley includes a plug-in memory card that is removable for easy data transfer.
  • the picking trolley is configured to return automatically to a pre-defined location for loading and/or unloading.
  • the picking trolley includes an emergency stop button that is activated via an elastic lanyard attached to the operator.
  • the picking trolley includes a removable battery module.
  • the picking trolley is configured to steer autonomously or quasi-autonomously along a crop row.
  • the picking trolley navigates amongst fruit producing plants, such as along rows of apple trees or any other fruit trees or plants or strawberry plants, including table grown strawberry plants, or raspberry plants.

Landscapes

  • Life Sciences & Earth Sciences (AREA)
  • Environmental Sciences (AREA)
  • Engineering & Computer Science (AREA)
  • Robotics (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

Un chariot de cueillette de fruits est conçu de sorte à comprendre des plateaux ou d'autres récipients pour contenir des fruits ou des cultures prélevés, et comprend également un système de moteur de traction électrique et/ou est conçu de sorte à être déplacé manuellement par un opérateur ; le chariot est également muni de capteurs. Une large gamme de capteurs différents peuvent être déployés, tels que des capteurs aptes à collecter des données concernant une condition de récolte pour des applications de surveillance et/ou de prévision de rendement ; des capteurs pour fournir des informations utiles concernant l'environnement de culture, telles que la température, l'humidité ; des capteurs aptes à déterminer le nombre ou la proportion de fruits à différents stades de maturité. Les données de capteur peuvent être géoréférencées.
PCT/GB2023/053011 2022-11-16 2023-11-16 Chariot de cueillette de fruits WO2024105405A1 (fr)

Applications Claiming Priority (4)

Application Number Priority Date Filing Date Title
GBGB2217136.7A GB202217136D0 (en) 2022-11-16 2022-11-16 Trolley
GB2217136.7 2022-11-16
GBGB2301007.7A GB202301007D0 (en) 2023-01-24 2023-01-24 Trolley II
GB2301007.7 2023-01-24

Publications (1)

Publication Number Publication Date
WO2024105405A1 true WO2024105405A1 (fr) 2024-05-23

Family

ID=89222010

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/GB2023/053011 WO2024105405A1 (fr) 2022-11-16 2023-11-16 Chariot de cueillette de fruits

Country Status (1)

Country Link
WO (1) WO2024105405A1 (fr)

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
AU2017357645A1 (en) * 2016-11-08 2019-05-23 Dogtooth Technologies Limited A robotic fruit picking system
US10459443B2 (en) * 2017-09-22 2019-10-29 Erik Jertberg Semi-autonomous farm transport vehicle for picked produce
AU2020251751A1 (en) * 2019-04-04 2021-10-28 Fieldwork Robotics Limited Robotic handling of picked fruit or vegetables
DE102021127064A1 (de) * 2020-11-17 2022-05-19 Deere & Company Intelligenter obstplantagen-erntewagen mit analyse
WO2022234573A2 (fr) * 2021-05-02 2022-11-10 Metomotion Ltd. Système autonome de cueillette et de mise en boîte automatique de fruits et son procédé de manœuvre

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
AU2017357645A1 (en) * 2016-11-08 2019-05-23 Dogtooth Technologies Limited A robotic fruit picking system
US10459443B2 (en) * 2017-09-22 2019-10-29 Erik Jertberg Semi-autonomous farm transport vehicle for picked produce
AU2020251751A1 (en) * 2019-04-04 2021-10-28 Fieldwork Robotics Limited Robotic handling of picked fruit or vegetables
DE102021127064A1 (de) * 2020-11-17 2022-05-19 Deere & Company Intelligenter obstplantagen-erntewagen mit analyse
WO2022234573A2 (fr) * 2021-05-02 2022-11-10 Metomotion Ltd. Système autonome de cueillette et de mise en boîte automatique de fruits et son procédé de manœuvre

Similar Documents

Publication Publication Date Title
CN110139552B (zh) 机器人水果采摘系统
EP3000305B1 (fr) Estimation de rendement
US20180047177A1 (en) Systems, devices, and methods for monitoring and assessing characteristics of harvested specialty crops
JP7026113B2 (ja) 果樹園、農園、及び温室のために、空中ドローンを用いて、収穫及び希薄化（ｄｉｌｕｔｉｎｇ）する（間引く）ためのデバイス、システム、及び方法
US10188037B2 (en) Yield estimation
US10796275B1 (en) Systems and methods for inventory control and delivery using unmanned aerial vehicles
US20160223511A1 (en) Soil quality measurement device
US11830191B2 (en) Normalizing counts of plant-parts-of-interest
US20220156670A1 (en) Smart orchard harvesting cart with analytics
US20210360848A1 (en) Produce harvesting apparatus and precision farming system
US20230030848A1 (en) Orchard vehicle and system
Ampatzidis et al. Portable weighing system for monitoring picker efficiency during manual harvest of sweet cherry
WO2024105405A1 (fr) Chariot de cueillette de fruits
JP7365984B2 (ja) 情報管理システム
JP7317779B2 (ja) 情報管理システム
US20240206456A1 (en) Weed control system
JP2022047500A (ja) 情報管理システム
Salah A hybrid control multi-agent system to automate fruit harvest and yield management.
KR20230097457A (ko) 농작물 성장 및 병해충 상태를 진단할 수 있는 다목적 이송장치
GB2616879A (en) An autonomous mobile robot and automated produce transportation system
WO2023222855A1 (fr) Système et procédé d'assistance robotique autonome dans des tâches agricoles
NZ794063A (en) A robotic fruit picking system

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 23822440

Country of ref document: EP

Kind code of ref document: A1