METHOD AND DEVICE FOR POSITIONING PLANTS IN A PLOT OF LAND

Abstract

According to some embodiments of the invention, there is provided a method for determining the spatial positions of plants in a plot, the plot being arranged in lines and rows, after obtaining a set of reference locations, each reference location belonging to at least one line or row of the plot, and being positioned in a global coordinate system, the method comprises determining, iteratively, a transformation between the global coordinate system and a line-row coordinate system associated with the plot, at least one iteration using a plurality of primary reference locations, a primary reference location being a location of the set for which the at least one corresponding line or row is known relative to at least one other primary reference location. [Fig. 2]

Description

Title of the invention: METHOD AND DEVICE FOR POSITIONING PLANTS IN A PLOT OF LAND

[0001] The present invention relates generally to precision agriculture. More specifically, the present invention relates to a method and device for determining the geographic coordinates of individual plants in a plot of land.

[0002] Precision agriculture has become essential to improve the sustainability, productivity and environmental footprint of agricultural production. Precision agriculture involves the use of multiple advanced technologies to optimize various aspects of agriculture, including crop profitability, resource utilization and overall productivity.These advanced technologies may include 1- the collection and analysis of data on relevant factors, such as soil conditions, weather conditions, crop health, using various means, including satellite imagery, sensors and drones, 2- global navigation satellite systems (GNSS) that allow receivers to be geolocated almost anywhere on a plot of land, 3- automated machinery, such as autonomous tractors and drones, which can perform tasks with a high degree of precision, including planting, harvesting and the application of inputs in a targeted manner, and 4- internet of things (IoT) sensors that can be used to monitor conditions such as soil and / or air moisture, temperature and crop health in real time.To optimize the overall efficiency and sustainability of farming operations, it is imperative that the various advanced technologies employed in precision agriculture collaborate seamlessly within a single plot of land. In one scenario, a plant geolocated by a GNSS receiver must be unambiguously identified by a tractor or autonomous drone operating on the plot. In another scenario, a plant flagged by an IoT sensor must be accurately identified in an image captured by an imaging satellite. In other words, a plantation map representing the spatial distribution of plants in the plot, in a local coordinate system and in a GNSS coordinate system, is required to ensure seamless collaboration between the advanced technologies employed.

[0003] A known approach to creating a planting map is to record, using a GNSS receiver, the global coordinates of each plant in the plot. Although remarkably accurate, this approach is very time consuming in due to the need to carry out measurements equivalent in number to the number of plants in the plot. This approach is particularly unsuitable for intensive agricultural scenarios characterized by a high density of plants.

[0004] Another known method for creating a planting map of a plot of land involves deploying a drone equipped with a GNSS receiver and a high-resolution camera. The drone captures calibrated images as it flies over the plot, so that each pixel in the captured image can be associated with GNSS coordinates. Software analysis is then performed to identify the pixels corresponding to each plant in the plot. However, this approach is limited to plots where plant vegetation, including branches and leaves, does not overlap. This limitation poses problems for modern plots, typically arranged in lines and rows, where plant overlap increases with their age. In addition, adverse weather conditions can degrade image quality, which is another disadvantage of this approach.

[0005] Thus, there is a need for an improved method for determining the geographic coordinates of individual plants within a plot of land. Disclosure of the Invention

[0006] According to a first aspect of the invention, there is provided a method for determining the spatial positions of plants in a plot, the plot being arranged in lines and rows, my method comprising: • obtain a set of reference locations, each reference location belonging to at least one line or row of the parcel and being positioned in a global coordinate system; • iteratively determining a transformation between the global coordinate system and a line-row coordinate system associated with the parcel, at least one iteration of the determination using a plurality of primary reference locations, a primary reference location being a reference location of the set for which the at least one corresponding line or row is known relative to at least one other primary reference location.

[0007] Accordingly, the method for determining the spatial positions of plants in a plot allows for precise positioning of the plants in the plot in the global and local coordinate systems, based on a set of reference locations taken from the plot, the number of reference locations used being, for example, several orders of magnitude smaller than the number of plants present in the plot. This results in a considerable saving of time and resources when determining a planting map of a plot of land involving global and local coordinate systems, making this invention particularly compatible with large plots of land with high plant density.

[0008] In some embodiments, the at least one iteration of the determination using a plurality of primary reference locations may comprise the first iteration of the determination, wherein the first iteration of the determination uses only primary reference locations.

[0009] In some embodiments, the primary reference locations used may be neighboring primary reference locations occupying a narrower spatial extent than a spatial extent associated with the parcel.

[0010] In some embodiments, the number of primary reference locations used may be such that two primary reference locations used belong to two different lines of the plot, and two primary reference locations used belong to two different rows of the plot.

[0011] In some embodiments, a current iteration of the determination, following a previous iteration of the determination, may use at least one new reference location of the set, a new reference location of the set being a reference location that has not been used in any previous iteration of the determination, the current iteration of the determination providing for each of the at least one new reference location an estimate of the at least one corresponding line or row in the line-row coordinate system, thereby becoming a secondary reference location.

[0012] In some embodiments, a current iteration of the determination may comprise: • estimate, for each of the at least one new reference location, the at least one corresponding line or row in the line-row coordinate system, using the transformation determined in the previous iteration of the determination; • solving an optimization problem using a transformation type and a subset of primary and / or secondary reference locations including the at least one new reference location, thereby providing a new transformation; • evaluate a reprojection error associated with the new transformation.

[0013] In some embodiments, estimating, for a new reference location, the at least one corresponding line or row may comprise: • if the new reference location belongs to both a line and a row of the plot, estimate both the corresponding line and row; • if the new reference location belongs to a line and not to a row, respectively to a row and not to a line, of the plot, estimate only the corresponding line, respectively only the corresponding row.

[0014] In some embodiments, the method may further comprise: • reiterate, for a variety of transformation types, the steps of resolution and evaluation; • select an optimal transformation based on the evaluated reprojection errors.

[0015] In some embodiments, the subset used may further include at least one reference location that became "secondary" in the previous iteration of the determination.

[0016] In some embodiments, the at least one new reference location may be selected, from among available new reference locations, based on at least one selection criterion.

[0017] In some embodiments, the at least one selection criterion may include a selection criterion based on geographic proximity to reference locations already used in previous iterations of the determination.

[0018] In some embodiments, the at least one selection criterion may include a selection criterion based on an amount of affiliation information, where a reference location belonging to a line and a row of the parcel is preferred over a reference location belonging to only a line or only a row of the parcel.

[0019] In some embodiments, the at least one selection criterion may comprise a selection criterion based on a relevance of affiliation information, where a reference location belonging to at least one row of the parcel is preferred to a reference location belonging to at least one line of the parcel.

[0020] In some embodiments, the determining step may be iterated until a completion condition is met, and the completion condition may be met if one or more of the following basic conditions are met: - the number of iterations reaches a threshold number of iterations; - the number of reference locations used belonging to both a line and a row of the plot reaches a first threshold; - the number of reference locations used belonging to a line and not to a row, respectively to a row and not to a line, of the plot reaches a second threshold; - the reprojection error associated with the last determined transformation is less than a reprojection error threshold; - all reference locations belonging to at least one row of the plot were used.

[0021] In some embodiments, the obtaining and determining steps may be performed in parallel, such that an iteration of the determination is triggered each time a reference location of the set is obtained.

[0022] In some embodiments, the obtaining and determining steps may be performed serially, such that iterations of the determination are triggered after the complete set is obtained.

[0023] In some embodiments, the obtaining and determining steps may be performed in a hybrid manner, such that an iteration of the determination is triggered each time a predefined number of reference locations of the set is obtained.

[0024] In some embodiments, each reference location in the set may belong to, at most, one line and one row of the parcel.

[0025] In some embodiments, a reference location of the assembly may represent one of the following: - the location of a plant in the plot; - the location of an element of the plot within it, associated with one or more plants; - an unoccupied location within the plot associated with one or more plants.

[0026] In some embodiments, the method may further comprise at least: • calculate, using the last determined transformation, the spatial positions of one or more plants in the plot, in the global and line-row coordinate systems; • calculate, using the last determined transformation, the spatial positions of one or more rows of plants in the plot, in the global and row-row coordinate systems.

[0027] In some embodiments, at least the lines or rows of the plot may be curved.

[0028] According to a second aspect of the invention, there is provided a processing device configured to perform the method according to any aspect or embodiment described above.

[0029] Any feature of one aspect of the invention may be applied to other aspects of the invention, in any appropriate combination. In particular, aspects related to methods can be applied to aspects related to devices / units, and vice versa.

[0030] Furthermore, features implemented in hardware may be implemented in software, and vice versa. Any reference to software and hardware features herein should be construed accordingly. For example, according to other aspects of the invention, there is provided a computer program comprising instructions that, when the program is executed by a processing unit, cause the processing unit to perform the method of any aspect or example described above and a computer-readable storage medium containing the computer program. Brief description of the drawings

[0031] Embodiments of the invention will now be described, by way of example only, and with reference to the following drawings in which: Fig.l

[0032] [Fig.l] is a diagram of an agricultural system associated with a plot of land, in which the present invention may be implemented, according to one or more embodiments; Fig. 2

[0033] [Fig.2] is a flowchart of a method for determining spatial positions individual plants in a plot of land, according to certain embodiments of the invention; Fig. 3

[0034] [Fig.3] is a flowchart of an iterative process that can be used to put implementing the determination process of the method of [Fig.2], according to embodiments of the invention; Fig. 4

[0035] [Fig.4] is a flowchart of another iterative process that can be used to implementing the determination process of the method of [Fig.2], according to embodiments of the invention; Fig. 5

[0036] [Fig.5] is a flowchart of a method for determining spatial positions individual plants in a plot of land, according to other embodiments of the invention; Fig. 6

[0037] [Fig.6] is a schematic diagram of a plot of land, compatible with the embodiments of the invention; Fig. 7

[0038] [Fig.7] illustrates schematic diagrams of various plots including stakes, compatible with embodiments of the invention; Fig. 8

[0039] [Fig.8] is a schematic diagram of a processing device in which the methods according to embodiments of the present invention can be implemented. Detailed description

[0040] Embodiments of the present invention provide methods, devices, and computer program products for determining the spatial positions of individual plants in a plot of land, based on a set of reference locations taken from the plot of land, the number of reference locations used being, for example, several orders of magnitude smaller than the number of plants present in the plot. More specifically, embodiments of the present invention enable dual positioning of each plant in the plot, including a position in a global coordinate system, such as a GNSS coordinate system, and a position in a local coordinate system associated with the plot, such as a line-row coordinate system, leading to the creation of a detailed planting map of the plot.The creation of such a detailed planting map is performed, according to embodiments of the present invention, progressively using an iterative approach, each new iteration using at least one new reference location from the set of reference locations.

[0041] The methods, devices, and computer program products according to the various embodiments of the present invention can be implemented in any agricultural system associated with a plot of land and requiring a detailed planting map of the plot to efficiently perform one or more tasks associated with its management, including land preparation, crop planting, irrigation, fertilization and soil management, disease detection and management, and harvesting.

[0042] The agricultural system may be part of a plot monitoring system using both aerial means, such as satellites or drones, and ground means, such as sensors directly attached to the plants in the plot and providing detection data as well as the local locations of the respective plants. In such a plot monitoring system, the various deployed means must interact to effectively monitor the plot. This ensures that measurements made by the ground means can be unambiguously correlated with data provided by the aerial means.

[0043] The agricultural system may be part of an autonomous tractor equipped with advanced technologies, such as sensors capable of detecting plants in the plot, and artificial intelligence (AI), in order to autonomously perform various tasks on the plot, such as plowing, planting, harvesting or applying inputs. In this scenario, the autonomous tractor may be configured to perform a task defined by GNSS coordinates. This task may consist of applying inputs to the plants in the plot, such that the amount applied is determined based on the GNSS coordinates.

[0044] The agricultural system may be part of a navigation device, which may be either portable or seamlessly integrated into a vehicle, configured to convert local coordinates of any plant in the field to global coordinates, or vice versa. This type of navigation device is of interest for precision agriculture, particularly when performing tasks in the field, such as installing stakes, or when determining global coordinates in an offline area lacking connectivity or mapping.

[0045] For illustration purposes only, the following description will be made with reference to a plot of land comprising a two-dimensional arrangement of plants obeying, at least approximately, a certain regularity. However, those skilled in the art will readily understand that the various embodiments of the invention apply to other types of plots, in particular to plots having a regular distribution of plants in the lines and not in the rows, or vice versa, or to plots where the lines of plants and / or the rows of plants are curved. Generally speaking, the invention can be integrated into any agricultural system requiring a planting map involving two coordinate systems.

[0046] In the following description, the expressions "plot of plants", "plot of land" and "plot" are used interchangeably to designate a plot of plants, where the plants are arranged in lines and rows. Without loss of generality, it is assumed that the lines of plants in the plot are not parallel, are not straight and do not intersect, and that the same is true for the rows of plants in the plot. For example, the rows and lines of plants are such that the distance between two adjacent plants in a row is less than the distance separating two adjacent plants in a line. As another example, the rows of plants may be curved, the distance between the plants in each row being approximately constant and identical from one row to the next.Furthermore, the term "plant" is used to refer to a plant location, i.e., a location intended for the planting of a plant, regardless of whether an actual plant is in that particular location at any given time in the life of the plot.

[0047] In the following description, the term "plot line" is not limited to a row of plants in the plot, and may be defined based on another element of the plot or may exist solely as an imaginary line. Similarly, the term "plot row" is not limited to a row of plants in the plot and may be defined based on another element of the plot or may exist solely as an imaginary row. With respect to the lines and rows of plants, it is assumed, without loss of generality, that the plot lines are not parallel, are not straight, and do not intersect, and that the same is true for the rows in the plot. Furthermore, the plot lines and rows are used to define reference points within the plot.

[0048] In the following description, the term "global coordinate system" is used to refer to any standardized framework that can be used to specify locations on the Earth's surface, particularly anywhere on any parcel of land. The global coordinate system may be the geographic coordinate system (GCS), which uses latitude and longitude to represent points on the Earth. An example of a GCS is the World Geodetic System 1984 (WGS 84), which is widely used as a standard coordinate system for GPS and mapping applications. The term "local coordinate system" is also used to refer to a coordinate system associated with a parcel of land, i.e., a reference frame used to specify locations within the parcel. The main characteristics of a local coordinate system are an origin and two non-collinear coordinate axes.A row-line coordinate system is an example of a local coordinate system, in which the two coordinate axes correspond to the rows and rows of the plot, respectively, and the origin is any intersection between a row and a row of the plot. For simplicity, the row-line coordinate system is used throughout this description as the local coordinate system, with integers being assigned to the rows of the plot and half-integers being assigned to the lines of the plot.

[0049] In the following description, the term "reference location" is used to refer to a location within the parcel that is positioned in the global coordinate system, i.e., its coordinates in the global coordinate system are known with high accuracy and precision. Furthermore, the reference location is a location that belongs to at least one line or row of the parcel. In other words, the reference location may belong to a line and not to a row, belong to a row and not to a line, or belong to both a line and a row of the parcel. Furthermore, it is assumed throughout this description that a location of reference belongs, at most, to a single line and a single row of the plot. Furthermore, the choice of a reference location within the plot is not arbitrary and is made depending on the plants and / or any other element of the plot. In general, the choice of a reference location within the plot is such that the position of a neighboring plant (relative to the reference location) can be derived, at least approximately, using the position of the reference location. Of course, the line and / or row of the plot to which a reference location belongs may be unknown in the local coordinate system, before triggering the iterative approach for creating the planting map of the plot according to the embodiments of the invention.

[0050] In the following description, the term "primary reference location" is used to designate a reference location for which the at least one corresponding line or row is known relative to at least one other primary reference location. Similarly, the term "secondary reference location" is used to designate a reference location for which the at least one corresponding line or row is estimated in the local coordinate system, after triggering the iterative approach for creating the planting map of the land plot according to the embodiments of the invention.

[0051] In the remainder of the description, the expression "transformation" is used to designate a mathematical transformation for converting the coordinates of a point from the global coordinate system to the local coordinate system, or vice versa. More specifically, the transformation comprises an elementary LocalToGlobal transformation for converting the coordinates of a point from the local coordinate system to the global coordinate system, and an elementary GlobalToLocal transformation for converting the coordinates of a point from the global coordinate system to the local coordinate system. The transformation is characterized by a transformation type and a set of parameters depending on the transformation type. Examples of transformation types include, but are not limited to, Helmert transformations, affine transformations, and polynomial transformations.

[0052] [Fig. 1] illustrates an agricultural system 10 associated with a plot of land 100, in which embodiments of the present invention may be implemented. The plot of land 100 includes a plurality of plants 110, including regular plants, arranged in lines and rows, and irregular plants that belong to a row of plants but not to a line of plants in the plot. The plot 100 also includes a plurality of stakes 120, each stake being strategically placed between or next to adjacent plants in a row. This placement is intentional and serves to support and guide the growth of the plants. By For example, wires 130 connect the stakes in a row, creating a framework that promotes upward growth and overall stability of the plants. As with the plants in the plot, a stake can be classified as regular or irregular. This classification depends on whether the stake is placed at the intersection of a row and a row of stakes, or whether the stake belongs solely to a row of stakes without being part of a row of stakes. The plot 100 can be divided into two distinct areas: a regular area characterized by a constant number of regular plants separating two rows of stakes and an irregular area where the number of regular plants separating two rows of stakes varies.In the following description, it is assumed that the rows of stakes are collinear with their respective rows of plants, while a line of regular stakes is, at least approximately, parallel to the two adjacent lines of regular plants. It is also assumed that the spatial distribution of the regular stakes can be derived directly and unambiguously from the spatial distribution of the enclosed regular plants, and vice versa.

[0053] The plot 100 is associated with a two-dimensional local coordinate system 160 defined on the basis of the regular stakes, such that the two coordinate axes are aligned respectively along the rows and lines of the stakes, and its origin is chosen as the position of a regular stake of the plot. Furthermore, the divisions of each coordinate axis are defined according to the number of regular plants separating two adjacent regular stakes. Of course, the local coordinate system can be defined on the basis of other elements of the plot, such as the plants of the plot. The plot 100 is further associated with a global coordinate system 150, defined independently of the plot 100, which can be two-dimensional or three-dimensional.While using a two-dimensional global coordinate system is suitable for flat plots, using a three-dimensional global coordinate system is advantageous when the plot has multiple elevation changes.

[0054] The agricultural system 10 associated with the land plot 100 comprises a positioning device 20 that can be handheld or integrated into a vehicle capable of navigating within the land plot 100 and detecting the reference locations 140. The positioning device 20 is configured, when activated at a reference location 140, to record its global coordinates, i.e., the global coordinates of the corresponding reference location 140, provided by a GNSS 40, as well as information indicating whether the reference location 140 is associated with a line of the plot, a row of the plot, or both. Furthermore, the positioning device 20 is configured, when activated at a primary reference location, to also record the corresponding line and / or row relative to at least one other primary reference location. The positioning device 20 is further configured to transmit the recorded measurement data, preferably by wireless means.

[0055] The agricultural system 10 further comprises a processing device 30 which may be located within the plot of land, at a remote location, or integrated into a cloud infrastructure. The processing device 30 is configured to receive the measurement data sent by the positioning device 20 and to implement a method for determining the spatial positions of the individual plants according to embodiments of the present invention, in order to provide a detailed planting map of the plot.

[0056] [Fig. 2] is a flowchart of a method 200 for determining the spatial positions of individual plants in a plot of land, according to embodiments of the invention. The plot comprises several plants, some of which are regular because they are at the intersection of a line and a row of plants, while other plants are irregular because they belong to a row of plants without being part of a line of plants. The plot may also comprise other elements, such as stakes or markers, which may also be classified as regular or irregular, as described with reference to [Fig. 1]. The plot is further associated with a global coordinate system and a local coordinate system.Without loss of generality, the global coordinate system is a geographic coordinate system (GCS), while the local coordinate system is a line-row coordinate system whose two coordinate axes correspond, respectively, to the lines and rows of the plot. The method can be implemented on any processing device, such as the processing device 30 described with reference to [Fig.l]. .

[0057] In step S210, a set of reference locations in the plot is obtained. A reference location of the set may correspond to an object of the plot or be unoccupied, devoid of any perceptible element. The reference locations of the set may be homogeneous or heterogeneous depending on whether they correspond to the same category, such as plants, stakes or unoccupied locations, or whether they correspond to a combination of all or part of these categories. If the reference locations of the set are homogeneous and correspond to plant locations, respectively to stake locations, the lines and rows of the plot are preferably those which are associated with the plants, respectively to the stakes.If the reference locations in the set are heterogeneous and correspond to plant and stake locations, for example, the lines and rows of the plot can be associated with both plants and stakes. Each reference location is . positioned in the global coordinate system, that is, its coordinates in the global coordinate system are known with high accuracy and precision. In addition, each reference location in the set is associated with affiliation information indicating whether the reference location belongs to a line and not to a row, whether it belongs to a row and not to a line, or whether it belongs to both a line and a row of the parcel. One or more reference locations in the set are primary reference locations. This implies that the line or row to which the primary reference location belongs is known relative to at least one other primary reference location. No additional information about the position of the reference locations, including the primary reference locations, in the line-row coordinate system is assumed to be known.Preferably, the number of reference locations in the set is several orders of magnitude smaller than the total number of plants in the plot.

[0058] In a first variant of step S210, the set of reference locations is obtained after being measured so that each row of the plot is represented. This involves measuring a reference location in each row of the plot. A measured reference location may belong to an arbitrary line or to none of the lines of the plot. This first variant of step S210 is advantageous in the case of a plot having straight rows that may not be parallel.

[0059] In a second variant of step S210, the set of reference locations is obtained after being measured, so that the represented rows are selected by applying sampling, according to a predefined sampling factor, to the rows of the plot. The sampling factor can be defined as a compromise between accuracy and complexity of processing. Of course, the applied sampling can be approximate in the sense that it is permissible to consider a reference location neighboring or close to that provided by the sampling. A measured reference location can belong to an arbitrary line or to none of the lines of the plot. This second variant of step S210 is advantageous in the case of a plot having straight rows that gradually slope.

[0060] In a third variant of step S210, the set of reference locations is obtained after being measured, so that the represented rows are arbitrarily selected. A measured reference location may belong to an arbitrary line or to none of the lines of the plot. This third variant of step S210 is advantageous in the case of a plot having straight rows that are almost parallel or arbitrarily oriented.

[0061] In some embodiments compatible with the aforementioned variants of step S210, the set of reference locations is obtained after being measured so that the represented lines are selected to approximately cover the entire plot. This involves measuring reference locations belonging to several respective lines of the plot. The selection may be performed arbitrarily or according to a predefined selection method.

[0062] In other embodiments also compatible with the aforementioned variants of step S210, the set of reference locations is obtained after being measured so as to include several reference locations corresponding to the row-line intersections and to include several reference locations corresponding to the ends of the rows.

[0063] Advantageously, the primary reference locations of the set are such that a basis for the row-line coordinate system can be established unambiguously. To achieve this, two diagonally opposite primary reference locations, i.e. belonging to different rows and rows, can be used, the row, respectively, of the first primary reference location being known relative to the row, respectively, of the second primary reference location. Furthermore, the origin of the row-line coordinate system can correspond to one of the two primary reference locations.

[0064] In step S220, a transformation between the global coordinate system and the line-row coordinate system associated with the parcel is iteratively determined by performing multiple iterations, using the obtained set of reference locations. At least one iteration of the determination process, represented by step S220, may use one or more primary reference locations from the obtained set. More generally, a current iteration of the determination process, following a previous iteration of the determination process, may use at least one new reference location from the obtained set, a new reference location being a reference location that was not used in a previous iteration of the determination process.The new reference location to be used in a current iteration of the determination process may be selected, from among the available new reference locations, either arbitrarily or on the basis of at least one selection criterion. Furthermore, the determination process is iterated until a completion condition, which may involve one or more elementary conditions, is met. The completion condition may be, for example, the processing of all the reference locations of the obtained set. The main result of step S220, and consequently of the overall method, is the last . transformation determined between the two coordinate systems. This result can be used to deduce, precisely or approximately, the spatial position of any plant in the plot. This result can also be used to create a map of the plot's plantations, providing the corresponding local and global coordinates for each plant.

[0065] In a first variant of step S220, at least one iteration of the determination process, such as the first iteration, preferably uses only primary reference locations of the obtained set, for which the global coordinates are precisely known and the local coordinates are such that an origin and a base for the line-row coordinate system can be established unambiguously. After establishing an origin and a base for the line-row coordinate system, each primary reference location of the obtained set can be positioned either at a point (first type) or along a line segment (second type), depending on whether the primary reference location belongs to both a line and a row of the parcel, or exclusively to a line or a row of the parcel.Preferably, the primary reference locations used are placed at points belonging to different lines and rows of the plot. For example, the primary reference locations may comprise two diagonally opposite primary reference locations, i.e., belonging to different lines and rows. Based on the primary reference locations, and more precisely on their respective coordinates in the global and local coordinate systems, a transformation between the two coordinate systems is determined. This may involve determining a transformation type and a set of parameters depending on the transformation type.For example, the transformation type to be used in the first iteration of the determination may be predefined as linear, and the associated parameter set may include the elements of a two-by-two projection matrix and the two elements of a translation vector, assuming that each reference location is described by two coordinates in each coordinate system. In this case, three primary reference locations of the first type, whose coordinates in the row-line coordinate system are precisely known, may be used to estimate the four elements of the projection matrix and the two elements of the translation vector. Of course, primary reference locations of the second type, or a combination of primary reference locations of the first type and of the second type, may be used to estimate the four elements of the projection matrix and the two elements of the translation vector.More generally, the transformation type can be of any other type, and the number of associated parameters can be greater or less. lower than that of a linear transformation type. Generally, the greater the number of parameters to be estimated, the greater the number of primary reference locations. In the case where the primary reference locations are numerous enough to provide multiple estimates of the transformation parameters, an average calculation can be performed on the resulting transformation parameters. Furthermore, only the nearest or furthest primary reference locations capable of estimating the parameters of the transformation type can be used to estimate the transformation parameters.Furthermore, a reprojection error associated with the transformation provided by the first iteration of the determination process can be calculated by measuring, for example, the root mean square error (RMSE) between the actual positions of the primary reference locations and their respective positions calculated using the provided transformation.

[0066] In some embodiments compatible with the first variant of step S220, the primary reference locations used in the first iteration of the determination process comprise only first-type primary reference locations whose respective positions in the row-line coordinate system are precisely known. The first-type primary reference locations used may be such that they form neighboring reference locations by occupying a spatial extent narrower than a predefined spatial extent threshold. More generally, the primary reference locations used are preferably neighboring primary reference locations occupying a spatial extent narrower than the spatial extent associated with the parcel.Further, the number of primary reference locations used is preferably such that two primary reference locations used belong to two different lines of the plot and two primary reference locations used belong to two different rows of the plot.

[0067] In other embodiments also compatible with the first variant of step S220, the number of primary reference locations used in the first iteration of the determination is equal to or greater than the minimum number of primary reference locations making it possible to obtain, at least once, an estimate of each parameter of the set of parameters associated with the type of transformation used.

[0068] In a second variant of step S220, a current iteration of the determination process, following a previous iteration of the determination process, uses at least one new reference location of the obtained set, which has preferably not been used in a previous iteration of the determination process. Generally, the new reference location is assumed not be a primary reference location, which implies that only its affiliation information and its coordinates in the global coordinate system are known. It should be remembered that the affiliation information associated with a reference location indicates whether it belongs to a line and not to a row, whether it belongs to a row and not to a line, or whether it belongs to both a line and a row of the parcel. The current iteration of the determination process can use the previously determined transformation between the two coordinate systems and a subset including, in addition to the new reference location, reference locations already positioned in the line-row coordinate system, in order to provide an estimate of the position of the new reference location in the line-row coordinate system, as well as a new transformation between the two coordinate systems.This subset may include primary reference locations and / or reference locations already processed in previous iterations of the determination process, also called secondary reference locations. For example, the reference locations of the subset may be chosen based on their respective position in the global coordinate system, whether they are primary or secondary reference locations, and / or based on their respective affiliation information. The number of reference locations of the subset is preferably such that the reference locations are not collinear.

[0069] In some embodiments, steps S210 and S220 are performed in series, such that iterations of the determination process are triggered after obtaining the set of reference locations. In other embodiments, steps S210 and S220 are performed in parallel, such that an iteration of the determination process is triggered each time a new reference location is available in the set of reference locations. Of course, steps S210 and S220 may also be performed in a hybrid manner, such that an iteration of the determination process is triggered each time a predefined number of reference locations is available in the set of reference locations.

[0070] [Fig. 3] is a flowchart illustrating an iterative process 300, represented by multiple steps that can be used to perform an iteration of the determination process described with reference to [Fig. 2], according to embodiments of the invention. The iterative process 300 receives a set of reference locations, such as the set of reference locations obtained in step S210, and is iterated multiple times until a completion condition is met. The multiple steps of the iterative process 300 are described below, considering an iteration current iteration of loop L301 that follows one or more previous iterations of loop L301.

[0071] In step S310, a new reference location is selected from the set of obtained reference locations. Without loss of generality, the new reference location is assumed not to be a primary reference location, meaning that only its affiliation information and its coordinates in the global coordinate system are known. The selection of a new reference location, from among the available new reference locations, can be done either arbitrarily or according to a first selection policy. The first selection policy can comprise one or more selection criteria, to be considered either in a preferential order or in an indifferent order.In the case of a preferential order, the selected reference location is the one that satisfies the most preferred criterion, while in the case of an indifferent order, the selection is based on the reference location that satisfies the greatest number of criteria. A first example of a selection criterion is based on the geographical proximity, using global coordinates, to the reference locations already processed, i.e., used, in the previous iterations of the L301 loop. According to this selection criterion, the selected reference location is the closest, in terms of distance, to the reference locations already processed. A second example of a selection criterion is based on the amount of affiliation information, with a reference location belonging to a line and a row of the plot being preferred to a reference location belonging only to a line or a row of the plot.A third example of a selection criterion is based on the relevance of affiliation information, with a reference location belonging to at least one row of the plot being preferred to a reference location belonging to at least one line of the plot.

[0072] In step S320, the position of the new reference location in the line-row coordinate system is estimated using a previously determined, preferably the most recent, transformation between the global coordinate system and the local coordinate system. Based on the associated affiliation information, the estimation of the position of the new reference location preferably comprises the estimation of the line and the row if the new reference location belongs to both a line and a row of the parcel, and the estimation of only the line, respectively only the row, if the new reference location belongs to only a line, respectively only a row, of the parcel.To do this, the following two operations can be performed: 1- apply the global coordinates of the new reference location to the transformation used, thus providing an intermediate estimate of the line and / or . of the corresponding row, 2- estimate the line and / or row of the new reference location, as being the closest to the intermediate estimate of the line and / or row. If we consider the local coordinate system in which integers are assigned to rows and half-integers to lines, the first operation can provide, for a reference location belonging to both a line and a row of the plot, real values ​​as an intermediate estimate of the corresponding line and row. Then, the second operation consists of rounding the intermediate estimate of the row to the nearest integer, and rounding the intermediate estimate of the line to the nearest half-integer. Of course, displacement errors, for the line and the row, can be observed and measured between the corresponding intermediate estimate (operation 1) and final estimate (operation 2).Once a final estimate of the corresponding line and / or row is provided, the new reference location becomes a secondary reference location. For example, the new reference location can be allowed to become a secondary reference location only if the corresponding displacement errors are less than predefined displacement error thresholds. This helps reduce any risk of error propagation during the iteration of loop L301.

[0073] In step S330, a new transformation between the two coordinate systems is determined using a subset of primary and / or secondary reference locations, including the new reference location that became "secondary" in step S320 of the current iteration of loop L301. The new transformation is preferably determined by solving an optimization problem using a transformation type and the subset of reference locations. An already determined, preferably the most recent, transformation between the two coordinate systems may also be used for initialization purposes, for example. Several techniques may be used to solve such an optimization problem, such as gradient descent, linear programming, and genetic algorithms.With the exception of the new reference location, the other reference locations of the subset may be chosen from the available primary and secondary reference locations, either arbitrarily or according to a second selection policy. As for the first selection policy described in step S310, the second selection policy may comprise one or more selection criteria, to be considered either in a preferential order or in an indifferent order. A first example of a selection criterion associated with the second selection policy is based on geographic proximity, using either global coordinates or local coordinates, relative to the new reference location. According to this selection criterion, the locations of . reference location of the subset are closest, in terms of distance, to the new reference location. A second example of a selection criterion associated with the second selection policy is based on the amount of affiliation information, with a reference location belonging to a row and a line of the plot being preferred to a reference location belonging only to a row or a line of the plot. A third example of a selection criterion associated with the second selection policy is based on the relevance of the affiliation information, with a reference location belonging to the same row, respectively the same line, of the plot as the new reference location being preferred to a reference location belonging to a different row, respectively a different line, of the plot.A fourth example of a selection criterion associated with the second selection policy is based on the recency of the primary and / or secondary reference locations, with a secondary reference location being preferred over a primary reference location and, among two secondary reference locations, the reference location that became "secondary" in the most recent iteration of the L301 loop being preferred over the other. Furthermore, the new transformation between the two coordinate systems may be determined using the same transformation type as the previous iteration of the L301 loop, meaning that the same transformation type is used for all iterations of the L301 loop. Examples of transformation types include, but are not limited to, linear, polynomial, affine, orthogonal, and trigonometric transformations.Of course, the transformation type can be a combination of several elementary transformation types, such as the transformation types listed above. Alternatively, the new transformation between the two coordinate systems can be determined to use a different transformation type than the one in the previous iteration of the L301 loop.

[0074] In step S340, a reprojection error associated with the new transformation is evaluated, based, for example, on the used subset of primary and / or secondary reference locations. Of course, other primary and / or secondary reference locations not belonging to the subset may be used to evaluate the reprojection error. More generally, the reprojection error may be evaluated as the root mean square error (RMSE) between the actual positions of the primary and / or secondary reference locations in the local coordinate system and their respective positions calculated using the new transformation applied to their positions in the global coordinate system.

[0075] In step S350, a completion condition is evaluated to determine whether it is met. If the completion condition is met, loop L301 terminates by providing, as the main result, the last determined transformation. Otherwise, a new iteration of loop L301 is triggered with respect to a new reference location. The completion condition may include a single elementary condition, such as processing all reference locations in the resulting set. More generally, the completion condition may include several elementary conditions. In this case, the completion condition is considered fulfilled when one, several, or all of the elementary conditions are satisfied.Examples of elementary conditions include, but are not limited to, the number of iterations of loop L301 reaches an iteration number threshold, the number of processed reference locations belonging to both a line and a row of the plot reaches a first threshold, the number of processed reference locations belonging to a line and not to a row, respectively to a row and not to a line, of the plot reaches a second threshold, the reprojection error associated with the last determined transformation is less than a reprojection error threshold, all reference locations belonging to at least one row of the plot have been processed.

[0076] [Fig. 4] is a flowchart illustrating an iterative process 400, represented by several steps and usable for performing step S330 of the process 300 described with reference to [Fig. 3], according to embodiments of the invention, which are particularly compatible with a scenario in which the type of transformation can be modified. For example, the process 300 described in [Fig. 3] may incorporate the iterative process 400 illustrated in [Fig. 4] in the form of a nested loop, the nested loop being iterated several times during each iteration of the main loop of the process 300 of [Fig. 3], until a nested completion condition is met. For simplicity, the iterative process 400 of [Fig.4] is called a "nested loop" (even though it is described as a main loop), receives a subset of primary and / or secondary reference locations, is iterated each time with respect to a different transformation type chosen from a predefined list of transformation types, and provides, as its main result, an optimal transformation between the global and local coordinate systems, as well as an associated reprojection error. The multiple steps of the iterative process 400 are described below, considering a current iteration of the nested loop L401 that follows one or more previous iterations of the nested loop L401.

[0077] In step S410, the subset of primary and / or secondary reference locations is received and a transformation type is selected from a predefined list of transformation types that includes several transformation types, such as linear, polynomial, affine, orthogonal, and trigonometric transformations. The predefined list of transformation types may also include combinations of basic transformation types, such as those listed above. In some embodiments, the transformation type is chosen based on the number of associated parameters, such that the chosen transformation type has the smallest number of associated parameters among the remaining available transformation types. In other embodiments, the transformation type is chosen arbitrarily from among the available transformation types.

[0078] In step S420, a transformation between the global and local coordinate systems is determined by solving an optimization problem using the chosen transformation type and the subset of reference locations. Preferably, the last determined transformation between the two coordinate systems is used for initialization purposes, for example. Furthermore, in step S420, the reprojection error associated with the determined transformation is evaluated, based, for example, on the subset of reference locations used.

[0079] In step S430, a nested completion condition is evaluated to determine whether it is met. If the nested completion condition is met, the nested loop L401 terminates by providing, as its main result, the transformation associated with the lowest reprojection error, also called the "optimal transformation". Otherwise, a new iteration of the nested loop L401 is triggered for a new transformation type that has not been tested. An example of a nested completion condition is that the reprojection error is less than a certain reprojection error threshold. Another example of a nested completion condition is that there are no other transformation types remaining in the predefined list of transformation types.

[0080] [Fig.5] is a flowchart of a method 500 for determining the spatial positions of individual plants in a plot of land, according to other embodiments of the invention. The plot of land comprises several plants, some of which are regular because they are at the intersection of a line and a row of plants, while other plants are irregular because they belong to a row of plants without being part of a line of plants. The plot also comprises stakes which are also classified as regular or irregular, depending on whether a stake is at the intersection of a line and a row of stakes. The lines and rows of the plot are defined by those of the stakes, which are also used to determine the reference locations within the plot. In particular, the line-row coordinate system associated with the plot has its two coordinate axes corresponding, respectively, to the lines and rows of stakes.The plot is further associated with a global coordinate system which is a geographic coordinate system (GCS). The method can be implemented on any processing device, such as the processing device 30 described with reference to [Fig.l]. Of course, it is . assumed that the spatial distribution of regular stakes can be derived directly and unambiguously from the spatial distribution of surrounded regular plants, and vice versa.

[0081] [Fig. 6] schematically illustrates an exemplary plot of land 100 including stakes 120 for supporting and guiding plant growth, for example, and consistent with the method 500 described with reference to [Fig. 5], the stakes 120 being used to determine the reference locations 140. The stakes shown may be classified as regular or irregular, as stakes belonging to a regular area or belonging to an irregular area (as defined with reference to [Fig. 1]), or as boundary stakes located approximately at the perimeter of the plot or as interior stakes located within the plot.Based on these different classifications, a particular stake in the plot can be classified as regular, associated with an irregular area and serving as a boundary stake, while another particular stake in the plot can be classified as irregular, associated with a regular area and serving as an interior stake. In [Fig.6], the regular area is separated from the irregular area by a solid line.

[0082] Referring to [Fig.5], in step S510, a plurality of reference locations associated with stakes in the plot are measured, recording their respective global coordinates and their respective affiliation information, thereby obtaining a set of reference locations. The set of reference locations includes, in particular, a plurality of primary reference locations for establishing an origin and a base for the local coordinate system. For example, the primary reference locations may correspond to regular stakes of which the number of intermediate regular stakes (or regular plants) is precisely known, and may be four neighboring regular stakes belonging to two adjacent stake rows and two adjacent stake lines.Furthermore, the set of reference locations preferably comprises reference locations representing stakes of different categories, including regular and irregular stakes, stakes belonging to regular areas and stakes belonging to irregular areas, boundary stakes and interior stakes. For example, as many boundary stakes as necessary to determine the perimeter of the plot may be measured, either precisely or approximately, to obtain their respective reference locations. This may involve making measurements on the boundary stakes of each row of the plot, or only on specific boundary stakes specified based on a sampling approach. Preferably, the set of reference locations comprises reference locations representing . regular stakes belonging to regular areas and other regular stakes belonging to irregular areas. In general, the number and location of regular and / or irregular stakes measured can be strategically chosen to approximately represent the entire plot.

[0083] Advantageously, the local coordinate system, in particular its associated basis, may be defined, in part, by considering the number of regular plants between directly adjacent regular stakes along the rows of the plot and the number of regular plants between directly adjacent regular stakes along the lines of the plot. Without loss of generality, assuming that there is no regular plant between two adjacent regular stakes along a line of stakes, while there are four regular plants between two adjacent regular stakes along a row of stakes, in this case, the vertical coordinate axis of the local coordinate system, corresponding to the lines of stakes (and possibly parallel to the rows of stakes), may be scaled so that two adjacent lines of regular stakes are separated by four units.On the other hand, the horizontal coordinate axis of the local coordinate system, corresponding to the rows of stakes (and possibly parallel to the lines of stakes), can be graduated so that two adjacent regular rows of stakes are separated by one unit. Furthermore, the rows of the vertical axis of the local coordinate system can be numbered using half-integers, while the rows of the horizontal axis of the local coordinate system can be numbered using whole numbers, or vice versa.

[0084] Figures 7A-7E show examples of plots compatible with various embodiments of the invention, where only the regular stakes 120 of a regular area are shown, the regular stakes associated with the primary reference locations being highlighted using white circular discs, the stake lines being horizontal and the stake rows being vertical. In Figure 7A, the stake lines are not straight or parallel, as they have arbitrary curvilinear shapes, while comprising the same number of stakes, and the same is true for the rows of the plot. In Figure 7B, the stake rows are parallel, while the stake lines have a gradual incline with an increasing slope. In Figure 7C, both the stake rows and the stake lines have a gradual incline with an increasing slope.In Figure 7E, two sub-areas of stakes can be distinguished with parallel rows and gradually inclined lines, each of the two sub-areas being characterized by a different slope. In Figure 7E, the rows of stakes are not straight (curved), as indicated by the solid lines, and are parallel, while the rows of stakes are straight and parallel.

[0085] With reference to [Fig. 5], in step S520, a transformation between the local and global coordinate systems is iteratively determined using the set of reference locations measured in step S510, as described in the embodiments of the invention, thereby providing a final transformation between the local and global coordinate systems. Preferably, the type of transformation is allowed to change from one iteration to another, considering a predefined list of transformation types. For example, a first set of iterations of the determination process is triggered with respect to the primary reference locations. Then, a second set of iterations of the determination process is triggered with respect to the first type reference locations, each of which belongs to both a line and a row of the parcel.Finally, a third series of iterations of the determination process is triggered with respect to the reference locations of the second type, each of them belonging only to one line or row of the plot.

[0086] In step S530, firstly, the local and global coordinates of the regular stakes within the plot are calculated. To do this, the local coordinates of the regular stakes can be obtained by applying an interpolation, in the local coordinate system, to the reference locations of the first type of the measured set (which are primary or become secondary in step S520), preferably taking into account the boundary stakes defining the perimeter of the plot. Then, the global coordinates of the regular stakes can be obtained by applying their respective local coordinates to the final transformation provided in step S520 (more precisely to the elementary transformation LocalToGlobal of the final transformation). Secondly, the local and global coordinates of the irregular stakes within the plot are calculated.To do this, the local coordinates of the irregular stakes may be obtained by considering the second-type reference locations of the measured set (which are primary or become secondary in step S520) and applying their respective global coordinates to the final transformation provided in step S520 (more precisely to the elementary transformation GlobalToLocal of the final transformation). The global coordinates of the irregular stakes may preferably be those directly measured in step S510, or may be those obtained by applying their respective local coordinates to the final transformation provided in step S520 (more precisely to the elementary transformation LocalToGlobal of the final transformation).Of course, if one has prior knowledge about other irregular stakes (not associated with reference locations), for example that an irregular stake is located at each end of a row, this prior knowledge can be used to . determining, for these other irregular stakes, their respective local and global coordinates, based on the final transformation provided in step S520.

[0087] In step S540, firstly, the local and global coordinates of the regular plants within the plot are calculated. For example, the local coordinates of the regular plants can be obtained by applying an interpolation to the local coordinates of the regular stakes, taking into account a direct transformation between the spatial distribution of the regular stakes and the spatial distribution of the surrounded regular plants. Then, the elementary transformation LocalToGlobal of the final transformation provided in step S520 can be used to derive the global coordinates of the regular plants within the plot. Secondly, the local coordinates of the irregular plants can be obtained by taking into account information including, but not limited to, the stakes defining the perimeter of the plot, the regular and irregular stakes, the distances separating the plants and the stakes.Then, the elementary transformation LocalToGlobal of the final transformation provided in step S520 can be used to derive the global coordinates of the irregular plants within the plot.

[0088] Of course, when the GlobalToLocal elementary transformation is used to derive the local coordinates of a plant (or stake) in the plot, rounding may be performed depending on how the rows and lines of plants (or stakes) are numbered in the local coordinate system. For example, if the rows and lines of plants are numbered using integers, rounding to the nearest integer may be performed when deriving the position of a regular plant in the local coordinate system. This rounding may be limited to the rows for an irregular plant belonging to a row of plants, but not to a line of plants.

[0089] Further, a spatial position may be determined and associated with each row or line of plants, based, for example, on the spatial positions of the plants and / or stakes belonging to that row or line of plants.

[0090] The various embodiments of the invention make it possible to determine the global and local coordinates of plants and any other elements, such as stakes, within a plot, by taking measurements at strategically chosen reference locations within the plot, the number of chosen reference locations being several orders of magnitude smaller than the total number of plants. This results in considerable time and resource savings when establishing a planting map of a plot involving global and local coordinate systems, which makes this invention particularly compatible with large plots with a high density of plants.

[0091] The various embodiments of the invention are particularly compatible with plots where the lines and rows of plants are not parallel and / or not straight, since the transformation between the two coordinate systems is determined iteratively using, at each iteration of the determination process, at least one new reference location and one type of transformation, the type of transformation being able to be different compared to a previous iteration of the determination process. Furthermore, the subset of reference locations used at each iteration of the determination process can also be chosen strategically to avoid any type of error propagation.For example, when the new reference location is close to a corner of the parcel, primary and / or secondary reference locations that are closer to the diagonally opposite corner of the parcel may be excluded from the subset used.

[0092] [Fig. 8] is a diagram illustrating exemplary hardware of a processing device 30 implementing a method for determining the spatial positions of individual plants in a plot of land, according to embodiments of the invention. The processing device 30 may be coupled to a positioning device 20 configured to measure reference locations within the plot. The processing device may also be coupled to a transmitting unit 50 configured to generate and transmit information generated by the processing device 30, such as the spatial positions of individual plants in the plot of land. The processing device 30 may be implemented with a bus architecture connecting various circuits, including, but not limited to, a processor 31, a computer-readable memory 32, and multiple components 33-35.Each of the multiple components may be coupled to both the processor and the computer-readable memory. A first component is a reference location identification unit 33 configured to identify reference locations, particularly primary reference locations, and their respective affiliation information, from externally provided data. The second component is a transformation unit 34 configured to use a obtained set of reference locations, including primary reference locations, to iteratively determine a transformation between a local coordinate system and a global coordinate system. A third component is a spatial positioning unit 35 configured to determine spatial positions, in both coordinate systems, for one or more individual plants in the plot of land.

[0093] Although the present invention has been described with reference to embodiments, it is understood that the invention is not limited to the disclosed embodiments. Persons Those skilled in the art will appreciate that various changes and modifications may be made without departing from the scope of the invention as defined in the appended claims. Any features disclosed in this specification (including the claims, the accompanying abstract and drawings), and / or any steps of a disclosed method or process, may be combined in any combination except for combinations in which at least some of these features and / or steps are mutually exclusive. Each feature disclosed in this specification (including the claims, the accompanying abstract and drawings) may be replaced by other features having the same, equivalent or similar purpose, unless expressly indicated otherwise.Thus, unless expressly stated otherwise, each disclosed feature is only an example of a generic series of equivalent or similar features.

[0094] In the claims, the word "comprising" does not exclude other elements or steps, and the indefinite article "a" or "an" does not exclude a plurality. The mere fact that different features are recited in mutually different dependent claims does not indicate that a combination of those features cannot be used advantageously.

[0095] In the preceding embodiments, the described functions may be implemented in hardware, software, firmware, or any combination thereof. If implemented in software, the functions may be stored or transmitted on a computer-readable medium, in the form of one or more instructions or code, and executed by a hardware processing unit.

[0096] Computer-readable media may include computer-readable storage media, which is a tangible medium such as a data storage medium, or communication media, which is any medium that facilitates the transfer of a computer program from one location to another, for example, in accordance with a communication protocol. Thus, a computer-readable medium may generally be (1) a tangible computer-readable storage medium that is not transient or (2) a communication medium such as a signal or carrier wave. Data storage media may be any available medium that can be accessed by one or more computers or one or more processors to retrieve the instructions, code, and / or data structures necessary to implement the techniques described herein.A computer program product may include a computer-readable medium.

[0097] By way of example and in a non-limiting manner, such a computer-readable storage medium may comprise RAM, ROM, EEPROM, CD- ROM or other optical disks, magnetic disks or other magnetic storage devices, flash memory, or any other medium that can be used to store desired program code in the form of instructions or data structures, and that can be accessed by a computer. Similarly, any connection is properly referred to as a computer-readable medium. For example, if instructions are transmitted from a website, server, or other remote source using coaxial cable, fiber optic cable, twisted pair, digital subscriber line (DSL), or wireless technologies such as infrared, radio, and microwave, coaxial cable, fiber optic cable, twisted pair, DSL, or wireless technologies such as infrared, radio, and microwave are included in the definition of the term "medium."It should be understood, however, that computer-readable storage media and data storage media do not include connections, carrier waves, signals, or other transient media, but rather are tangible, non-transitory storage media. The terms "disc" and "disk," as used herein, include compact disc (CD), laser disc, optical disc, digital versatile disc (DVD), floppy disk, and Blu-ray Disc, with discs generally reproducing data magnetically, while disks reproduce data optically using lasers. Combinations of the above should also be included within the scope of computer-readable media.

Claims

Claims

1. Method for determining the spatial positions of plants in a plot (100), the plot (100) being arranged in lines and rows, characterized in that the method comprises: • obtaining a set of reference locations, each reference location (140) belonging to at least one line or row of the plot (100) and being positioned in a global coordinate system (150); • determining, iteratively, a transformation between the global coordinate system (150) and a line-row coordinate system (160) associated with the plot (100), at least one iteration of the determination using a plurality of primary reference locations, a primary reference location (140) being a reference location (140) of the set for which the at least one corresponding line or row is known relative to at least one other primary reference location (140).

2. The method of claim 1, wherein the at least one iteration of the determination using a plurality of primary reference locations comprises the first iteration of the determination, the first iteration of the determination using only primary reference locations.

3. A method according to claim 1 or 2, wherein the primary reference locations used are neighboring primary reference locations occupying a narrower spatial extent than a spatial extent associated with the parcel (100).

4. A method according to claim 3, wherein the number of primary reference locations used is such that two primary reference locations used belong to two different lines of the plot (100), and two primary reference locations used belong to two different rows of the plot (100).

5. Method according to one of the preceding claims, in which a current iteration of the determination, following a previous iteration of the determination, uses at least one new reference location (140) of the set, a new reference location (140) of the set being a reference location (140) which has not been used in any previous iteration of the determination, the current iteration of the determination providing for each of said at least one new reference location (140) an estimate of the at least one corresponding line or row in the line-row coordinate system (160), thus becoming a secondary reference location (140).

6. A method according to claim 5, wherein a current iteration of the determination comprises: • estimating, for each of said at least one new reference location (140), the at least one corresponding line or row in the line-row coordinate system (160), using the transformation determined in the previous iteration of the determination; • solving an optimization problem using a transformation type and a subset of primary and / or secondary reference locations comprising said at least one new reference location (140), thereby providing a new transformation; • evaluating a reprojection error associated with the new transformation.

7. Method according to claim 6, wherein the estimation, for a new reference location (140), of the at least one corresponding line or row comprises: • if the new reference location (140) belongs to both a line and a row of the plot (100), estimating both the corresponding line and row; • if the new reference location (140) belongs to a line and not to a row, respectively to a row and not to a line, of the plot (100), estimating only the corresponding line, respectively only the corresponding row.

8. A method according to claim 6 or 7, wherein the method further comprises: • repeating, for a variety of transformation types, the steps of solving and evaluating; • select an optimal transformation based on the evaluated reprojection errors.

9. Method according to one of claims 6 to 8, in which the subset used further comprises at least one reference location (140) which became "secondary" in the previous iteration of the determination.

10. Method according to one of claims 5 to 9, wherein said at least one new reference location (140) is selected, from among available new reference locations, on the basis of at least one selection criterion.

11. The method of claim 10, wherein said at least one selection criterion comprises a selection criterion based on geographic proximity to reference locations already used in previous iterations of the determination.

12. A method according to claim 10 or 11, wherein said at least one selection criterion comprises a selection criterion based on an amount of affiliation information, where a reference location (140) belonging to a line and a row of the parcel (100) is preferred to a reference location (140) belonging only to a line or only to a row of the parcel (100).

13. Method according to one of claims 10 to 12, wherein said at least one selection criterion comprises a selection criterion based on a relevance of affiliation information, where a reference location (140) belonging to at least one row of the parcel (100) is preferred to a reference location (140) belonging to at least one line of the parcel (100).

14. Method according to one of the preceding claims, wherein the determining step is iterated until a completion condition is met, and wherein the completion condition is met if one or more of the following elementary conditions are met: - the number of iterations reaches an iteration number threshold; - the number of reference locations used belonging to both a line and a row of the plot (100) reaches a first threshold; - the number of reference locations used belonging to a line and not to a row, respectively to a row and not to a line, of the plot (100) reaches a second threshold; - the reprojection error associated with the last determined transformation is less than a reprojection error threshold; - all the reference locations belonging to at least one row of the plot (100) have been used.

15. Method according to one of the preceding claims, in which a reference location (140) of the set represents one of the following elements: - the location of a plant in the plot (100); - the location of an element of the plot (100) within it, associated with one or more plants; - an unoccupied location within the plot (100) associated with one or more plants.

16. Method according to one of the preceding claims, in which the method further comprises at least: • calculating, using the last determined transformation, the spatial positions of one or more plants in the plot (100), in the global and line-row coordinate systems; • calculating, using the last determined transformation, the spatial positions of one or more rows of plants in the plot (100), in the global and line-row coordinate systems.

17. A processing device configured to perform a method according to one of claims 1 to 16.

18. A computer program comprising instructions which, when the program is executed by a computer, cause the computer to implement a method according to one of claims 1 to 16.

19. A computer-readable medium containing a computer program according to claim 18.

Images