Detecting plant diseases at the onset stage

Abstract

The present invention relates to detection of plant diseases at a developmental stage. A computer-implemented method for determining the onset and / or onset time of a disease in a plant (12) in agriculture is provided. The method includes providing (S110) first data including field data (14) associated with the cultivation of the plant and weather data (16) associated with the location where the plant is cultivated to a computer model (20). The method further includes determining (S120) a plant disease presence prediction and a plant disease spread for the plant using the computer model (20), and determining the onset and / or onset time when the plant disease is predicted to develop from the computer model (20) output including the plant disease presence prediction and second data (18) including one or more vegetation indices associated with the plant by using the plant disease presence prediction and the change in the one or more vegetation indices.

Description

[Technical field]

[0001] This application relates to computer-aided agricultural plant treatment. In particular, this application relates to a method and apparatus for determining the onset and / or onset time of plant diseases in agriculture. Furthermore, this application relates to the use of computer models to predict plant diseases. [Background technology]

[0002] In agriculture, cultivated plants, crops, etc. may be affected by diseases that occur between sowing and harvesting, which may reduce yields. Thereby, plant disease occurrence is mainly caused by three factors, namely the host plant, which affects the plant's inherent vulnerability, the pathogen, which represents the agent causing the disease, and the environmental conditions, which may include weather favorable for the disease, etc. These three factors mainly cause the occurrence of the disease, and the progression of the disease is a dynamic process between these factors. This complex correlation of factors alone makes the management of plant diseases a challenge. However, plants can be kept healthy by applying plant protection measures or plant treatments, such as plant protection agents, insecticides, etc. Nevertheless, it is still a challenge, for example, to determine the optimal time for plant protection measures, to identify the appropriate plant protection agent, to determine the optimal amount of plant protection agent, etc.

[0003] To this end, computer models for predicting plant diseases, e.g., the severity of plant diseases, can be used to schedule plant treatments to control the plant disease. When such computer models are used in field or crop management systems used, for example, by farmers, to assist in the care and / or treatment of plants, the system provides rather inaccurate, sometimes too late, times when devices, for example, notify farmers about crop diseases and / or provide work instructions for crop treatment and / or disease control. Summary of the Invention [Means for solving the problem]

[0004] Therefore, when using computer models to predict plant diseases, there may still be a need to improve the prediction accuracy. The object of the present invention is solved by the subject matter of the independent claims, with further embodiments being incorporated in the dependent claims.

[0005] In a first aspect, a computer-implemented method for determining the onset and / or onset timing of a plant disease in agriculture is provided. The method includes providing first data, including field data associated with the cultivation of the plant and weather data associated with the location where the plant is cultivated, to a computer model, and using the computer model to determine a plant disease presence prediction and a plant disease spread for the plant. The method further includes determining, from an output of the computer model including the plant disease presence prediction and second data including one or more vegetation indices associated with the plant, a predicted onset and / or onset timing of the plant disease in the plant by using the plant disease presence prediction and a change in the one or more vegetation indices or by using the plant disease presence prediction together with the change in the one or more vegetation indices, in particular by aligning the disease presence prediction with the change in the one or more vegetation indices.

[0006] In other words, a computer-implemented method for detecting plant disease in agriculture at a stage of plant disease development is provided. The method includes providing first data, including field data associated with the cultivation of the plant and weather data associated with the location where the plant is cultivated, to a computer model, and using the computer model to determine a plant disease presence prediction and a plant disease spread for the plant. The method further includes determining, from an output of the computer model including the plant disease presence prediction and second data, including one or more vegetation indices associated with the plant, a predicted onset and / or onset time of the plant disease on the plant by using the plant disease presence prediction and changes in the one or more vegetation indices or by using the plant disease presence prediction together with changes in the one or more vegetation indices, in particular by aligning the disease presence prediction with changes in the one or more vegetation indices.

[0007] The inventors have found that vegetation indices, which may be any index, e.g. graphical index, that can be used to analyze the vegetation in a field, a section of a field, and / or a plant, can be used as auxiliary features to detect the early onset of a plant disease together with a plant disease presence prediction derived from a computer model based on a first data, or the change in the plant disease presence prediction over time, or the change in the disease severity prediction over time. In other words, the inventors have found a method, a system, etc. herein to estimate the appearance of a plant disease at an early stage and work to control the spread of the disease by analyzing the behavior of the vegetation in the field before the first observation of the disease, taking into account the available vegetation indices. Thus, the method makes it possible to predict early signs of disease progression based on a combination of both the output of a computer model and the changes in the vegetation indices. In other words, the method estimates the onset of a plant disease based on both the trend of disease progression predicted in the model and the vegetation indices. It relies on three types of data: in situ or field (collected) data, meteorological data, and vegetation indices.

[0008] In this way, the farmer can be alerted early, even in advance or at least at an early stage of disease development and / or spread, so that there is a better chance of controlling the plant disease. This has a positive impact on field and crop management, e.g. on sustainability or the environment, since equipment for crop treatment has to be operated less frequently due to less spread, less pesticides etc. that have to be applied etc. Once the onset or time of onset is determined, the field and / or plant management system can provide information about the developed plant disease to the farmer and / or determine a treatment schedule or plan for plant treatment and / or treatment against the disease. The field and / or plant management can also include directing or controlling robotic vehicles, unmanned aerial vehicles etc. and / or sending messages to the farmer. Furthermore, better control of plant diseases at an early stage allows for better or maximum yields to be obtained, at least compared to disease-infested plants.

[0009] It is noted that the onset and / or onset time of the plant disease may be determined and / or used in a server or cloud environment that may be or may be part of the field or plant management system described above. The onset and / or onset time may be used to notify, e.g., send as a message, a farmer of the onset of the plant disease, so that the farmer can initiate measures, such as plant treatments, in response to control the plant disease and its spread. Alternatively or additionally, the onset and / or onset time of the plant disease may be determined and / or used in a client device that may be at the farmer's site.

[0010] As used herein, a plant disease can be any undesirable devastating plant disease and / or devastating crop disease. As an example, the plant disease can be caused by a fungus called Puccinia striiformis, but other diseases are possible, at least if they cause visible effects on the plant's physiology, such as reduced biomass, reduced leaf area index (LAI), lesions caused by infection, pigment destruction, wilting, etc. Such plant diseases can affect, for example, the yield of the plant.

[0011] The plants observed and / or treated may be any cultivated plant, sometimes referred to as crops, such as wheat, rice, corn, potato, soybean, etc. Preferably, the plants show visible effects on plant physiology, such as loss of biomass, loss of Leaf Area Index (LAI), lesions caused by infection, pigment destruction, wilting, etc., upon onset and / or spread of a plant disease.

[0012] The computer model may be a regression model, which is preferred since plant disease severity is a continuous variable. The computer model may be trained with training data prior to its use. The computer model may be implemented with a tree-based algorithm, such as XGBoost, Random Forest, CatBoost, etc. For example, the computer model may be used to model disease severity over time as a value between 0 and 1, with 0 indicating no disease and 1 indicating maximum infestation.

[0013] The weather data may include, for example, observed data from one or more weather stations and / or modeled data, which may include one or more parameters, such as temperature (maximum, minimum, and average), temperature at 5 cm height, cloud cover percentage, dew point, precipitation (mm and duration), relative humidity (maximum, minimum, and average), sunshine duration in hours, and wind speed (maximum, minimum, and average), etc. For example, plant diseases may be correlated with the behavior of daily mean and maximum temperature, precipitation, sunshine duration, and other weather parameters that influence the progression of plant diseases.

[0014] The location where the plants are grown can be known from field data, from searches, in situ observations, farmer records provided, etc., and can be given as geographic coordinates, etc.

[0015] According to an embodiment, using the plant disease presence prediction and the change in one or more vegetation indices or using the plant disease presence prediction with the change in one or more vegetation indices, specifically matching, may include applying a prediction rule, which is applied to determine the presence of a plant disease in the plant disease presence prediction and the change in one or more vegetation indices, such as an increase or decrease, as the onset and / or onset time of the plant disease. In other words, if the presence is determined in the plant disease presence prediction of the computer model and a change in the vegetation indices, such as an increase or decrease, is determined, the method signals the onset of the plant disease. Thus, the onset of the plant disease can be determined very accurately and an early warning to the farmer can be generated and provided accordingly.

[0016] In further other words, according to one embodiment, using, specifically matching, the plant disease presence prediction with a change in one or more vegetation indices or using the plant disease presence prediction with a change in one or more vegetation indices may include applying a prediction rule, which defines that if a plant disease is present in the plant disease presence prediction and there is a change in one or more vegetation indices, a decision of the onset of the plant disease is signaled.

[0017] In one embodiment, the estimation rules may be expressed as follows:

number

number

number

[0018] According to an embodiment, using, specifically aligning, the plant disease presence prediction with changes in one or more vegetation indices or with changes in one or more vegetation indices may include convolving a differential operator with one or more vegetation indices in the plant disease presence prediction, which includes a time course of multiple plant disease presence predictions, and determining the point with the maximum slope as the onset and / or onset time of the plant disease. For example, considering a time series of these predictions, it may be observed that the computer model shows a spike in the plant disease presence prediction a few days before the disease is actually observed. Therefore, an efficient method can be applied to detect this spike and then fix a window around this spike that can be used as an observation window to detect the onset of disease severity. To detect these spikes, a differential operator with a length twice the length of the actual time series is convolved with the remotely sensed indices and the point where the maximum spike occurs is the point with the maximum slope. This approach works well to detect spikes in the computer model.

[0019] In one embodiment, the one or more vegetation indices may include one or more of the following: Normalized Difference Vegetation Index, NVDI, Normalized Difference Red Edge Index, NDRE, and Normalized Difference Water Index, NDWI. As mentioned above, these vegetation indices are graphic indices that can be used to analyze remotely sensed data, e.g., measurements to assess whether a field, a field section thereof, and / or observed plants contain living green vegetation, etc., and are good indicators of vegetation status and amount. Of course, any vegetation index that indicates vegetation status and / or amount can be used in this context. For example, healthy plants are usually green because they absorb blue and red light and have a higher reflectance for near infrared (NIR) and green in visible wavelengths. However, when a plant becomes diseased, the spectral reflectance and absorbance are essentially reversed. Thus, one or more vegetation indices can also be used as an adjunct to manual search, acting as a diagnostic tool that can detect and quantify plant diseases in an automated and objective manner compared to manual search, which is less efficient and influenced by human judgment.

[0020] According to an embodiment, the second data and / or one or more vegetation indices may be derived from the remotely sensed data. In other words, the data used herein to determine the onset and / or onset time of the plant disease may specifically include in situ or field data, meteorological data, and remotely sensed data. This allows such useful information of the field, i.e. one or more vegetation indices, to be obtained without any physical contact or significant manual effort. Such non-invasive exploration may be performed, for example, using one or more optical sensors. Thus, by combining the capabilities of optical sensors, for example with Geographic Information Systems (GIS) and Internet of Things (IoT), it is possible to incorporate the obtained exploration results into precision agriculture. The optical sensors used for disease detection may include RGB (red, green, blue) sensors, multi- and hyperspectral reflectance sensors, thermal sensors, and / or fluorescent imaging. More specifically, the RGB sensors may be configured to capture digital images of the plant disease in red, green, and blue channels, and such images may be processed by machine learning and / or computer vision techniques to quantify the disease in the crop. Furthermore, with multi- and hyperspectral reflectance sensors, disease detection is based on the amount of light emitted from the crop canopy at specific wavelengths of the electromagnetic spectrum (EMS). Exemplary types of spectral sensors, namely multispectral and hyperspectral sensors, are based on the number of bands and narrow bands for the EMS. Multispectral sensors provide imaging with 3-10 bands, e.g., red, green, blue, near infrared (NIR), and short-wave infrared. In contrast, hyperspectral sensors have multiple narrower bands (100-1000 bands). Instead of monitoring reflectance from individual bands, they can also be combined in the form of vegetation indices (VIs) as mentioned above. Vegetation indices such as Normalized Difference Vegetation Index (NDVI), Normalized Difference Red Edge Index (NDRE), Normalized Difference Water Index (NDWI), etc. are all good indicators of vegetation condition and amount. Therefore, such spectral sensors are considered valuable for site-specific disease management.Furthermore, thermal sensors are configured to measure radiation emitted from the crop canopy and convert it to temperature, which depends on the plant's evaporative capacity and soil moisture status, and thus shows that the presence of disease can be influenced and detected using thermal imaging. Furthermore, fluorescent imaging is configured to capture the photosynthetic capacity of the plant, allowing disease to be estimated as a function of the difference in photosynthetic capacity between healthy and disease-stressed plants.

[0021] Remote sensing has the ability to provide near real-time information at both field and large scales, thus enabling sustainable plant protection. As used herein, remote sensing may be understood as the process of detecting and monitoring the physical characteristics of an area at a distance, typically from a satellite or an aircraft, unmanned aerial vehicle (UAV), etc., by measuring its reflected and emitted radiation, as long as they provide measurements and / or images suitable for analysis to derive a vegetation index. Sensors and / or cameras on satellites and aircraft, UAVs, etc., can measure, image, etc., large areas on the Earth's surface.

[0022] In one embodiment, remote sensing data can be obtained from one or more satellites and / or one or more aircraft. In this regard, remote sensing can rely on optical sensors mounted on satellites and / or aircraft, for example, to measure the amount of light reflected and / or emitted from plants at specific wavelengths to estimate chlorophyll content, which is an indication of plant health. As mentioned above, healthy plants are usually green in color because they absorb blue and red light and have higher reflectance for near infrared (NIR) and green in visible wavelengths. However, when a plant becomes diseased, the spectral reflectance and absorption are essentially reversed. Thus, remote sensing can be an adjunct to manual exploration, acting as a diagnostic tool that can detect and quantify plant disease in an automated and objective manner compared to manual exploration, which is less efficient and influenced by human judgment. For example, satellites used as data sources for remote sensing can include Sentinel-2, Airbus OneAtlas, Landsat 7 or 8, etc.

[0023] According to one embodiment, the remote sensing data may include one or more multispectral images. For example, a multispectral sensor may provide imagery having 3-10 bands, e.g., red, green, blue, near infrared (NIR), and short-wave infrared. Instead of monitoring reflectance from individual bands, they may be combined in the form of a vegetation index (VI).

[0024] In one embodiment, the field data may include information regarding one or more of the following: the plant's growth stage, the number of days since the plant was sown, the month the plant was planted, the planting date, the location of the plant, and the previous crop, which data, individually, in groups, or in combination, have been shown to be particularly useful for the early detection of the onset of plant disease.

[0025] According to one embodiment, the weather data may include information regarding one or more of the following: wind speed, average wind speed, relative humidity, average relative humidity, maximum relative humidity, air temperature, maximum air temperature, precipitation, and minimum air temperature at a height of 5 cm, which data, individually, in groups or in combination, have shown to be particularly useful for the early detection of the onset of plant diseases.

[0026] In one embodiment, the computer model may be trained using training data including training data fitted to the first data but not training data fitted to the second data prior to use of the computer model for plant disease presence prediction. In other words, the inventors have found that one or more vegetation indices are used as auxiliary features to the plant disease presence prediction model to determine the early onset of disease without using them as features to train the computational model. As described herein, the computational model may be trained using only temporal, geometric, hyperlocal weather, previous crop and / or growth stage characteristics, and / or other data described herein, in addition to vegetation index related data or remotely sensed data, for example, and the computer model is used to determine when the model begins to predict disease presence, and one or more vegetation indices, such as NDVI and NDRE, are observed to see when they begin to change, for example, increase or decrease. Using the above-mentioned estimation rules based on the model-predicted trends of disease presence and / or progression, and one or more vegetation indices and / or remotely sensed indices, disease onset and / or disease severity may be estimated. Excluding vegetation index and / or remote sensing data when training the computer model surprisingly resulted in particularly accurate disease incidence predictions or estimates. In addition, due to the limited training data, the training effort is lower.

[0027] According to an embodiment, the first data and / or the second data may be provided in a tabular structure. The first data may be selected according to their importance: growth stage of the crop under consideration, latitude of the crop cultivation, country of the crop cultivation, maximum temperature of and / or during the crop cultivation, planting month of the crop under consideration, planting date of the crop under consideration, planting date of the crop under consideration, previous crop of the field of the crop under consideration, planting year of the crop under consideration, days after sowing of the crop under consideration, trial year (i.e. the year the trial was carried out), minimum temperature of and / or during the crop cultivation, minimum temperature at a height of 5 cm above ground of the crop under consideration, duration of precipitation, average percentage of cloud cover during the crop cultivation, maximum dew point during the crop cultivation, maximum wind speed during the crop cultivation, average wind speed during the crop cultivation, minimum relative humidity during the crop cultivation, maximum relative humidity during the crop cultivation, average relative humidity during the crop cultivation, best precipitation, sunshine duration during the crop cultivation. For example, weather can be very decisive for the occurrence and progression of a disease, so for selection its importance is rather high. By selecting a set of features from the above list, the predictions of the computer model can be further improved.

[0028] In one embodiment, the method may further comprise providing the determined onset and / or onset time as output data for use in plant protection or treatment triggers, alarms and / or schedules. From such prediction of onset and / or onset time, for example for disease management, a spraying time or time window of a plant protection agent can be derived. By way of example, the onset prediction can be used to identify the plant protection agent to be used. For this purpose, suitable computer models, or data obtained by databases, etc. may be used. This can further improve the quality of the plant treatment. In one example, the onset prediction can be used as a trigger to inform (e.g. with a message, an alarm, etc.) and / or to instruct a farmer user to perform a specific measure for crop protection, for example at a specific time. Furthermore, by way of example, the onset protection may be used to generate control data adapted to be provided to a robotic device, for example a spraying device, which may be adapted to automatically perform crop treatments, for example at a specific date or time, and / or to apply, for example, a plant protection agent, an insecticide, etc., and may be used to determine a crop protection and / or treatment plan to be provided for crop protection and / or acute treatment. The plan may include a treatment cycle or treatment timing. In other words, by using the onset, it is possible to find a suitable, preferably the most suitable treatment timing, for example, spray timing, at which, for example, a plant protection agent or insecticide can be applied, on the one hand, to at least control, eliminate or prevent the disease, and on the other hand, to minimize the amount of plant protection agent or insecticide required. Thus, the plant protection treatment parameters can be called optimal treatment and / or application timing. This can further improve efficiency in agriculture.

[0029] A second aspect of the present application relates to the use of disease onset and / or onset time determined by the method according to the first aspect for scheduling or planning plant treatments in agriculture. From such prediction of disease onset and / or onset time, spraying time points or time windows of plant protection agents can be derived, for example for disease management. By way of example, the onset prediction can be used to identify the plant protection agent to be used. For this purpose, suitable computer models, or data obtained by databases, etc. can be used. This can further improve the quality of plant treatments. In an embodiment, the onset prediction may be used as a trigger, for example to inform a farmer via a message, an alarm, etc., and / or to instruct a farmer user to carry out a specific operation for crop protection, for example at a specific time. Furthermore, by way of example, disease protection may be used to generate control data, which may be adapted to be provided to a robotic device, for example a spraying device, adapted to automatically perform crop treatments, for example at a specific date or time, and / or to apply, for example, plant protection agents, pesticides, etc. This can further improve efficiency in agriculture.

[0030] In a third aspect, there is provided an apparatus for determining the onset and / or timing of onset of a plant disease in agriculture, the apparatus may be configured to carry out a method according to the first aspect.

[0031] In other words, an apparatus is provided for detecting plant diseases in agriculture at the developmental stage of the plant disease.

[0032] The apparatus includes a data processor and a data interface connected to the data processor. The data interface is configured to receive first data including field data associated with the cultivation of the plant and weather data associated with a location where the plant is cultivated, and second data including one or more vegetation indices associated with the plant, and the data processor is configured to execute a computer model for determining a plant disease presence prediction and a spread of the plant disease for the plant based on the received first data. The data processor is configured to determine an expected onset and / or onset time of the plant disease in the plant based on an output of the computer model including the disease presence prediction and the second data by using the plant disease presence prediction and changes in the one or more vegetation indices or using the plant disease presence prediction together with changes in the one or more vegetation indices, in particular by aligning the plant disease presence prediction with changes in the one or more vegetation indices.

[0033] A fourth aspect provides a computer program element for determining the onset and / or timing of onset of a plant disease in agriculture, the computer program being adapted, when executed by a data processor and / or a computing device, to perform a method according to the first aspect and / or to control an apparatus according to the third aspect.

[0034] According to a fifth aspect, there is provided a computer readable storage or transmission medium having stored thereon or comprising a computer program element according to the fourth aspect.

[0035] A sixth aspect provides a server or cloud environment configured for determining onset and / or onset timing of a plant disease in agriculture according to any one of the above aspects. The server or cloud environment may include at least one data processor and at least one data and / or communication interface preferably configured for providing and / or receiving one or more of the first and / or second data, i.e. field data, weather data, and one or more vegetation indices.

[0036] A seventh aspect provides a client device configured for determining the onset and / or onset time of a plant disease in agriculture according to any one of aspects 1 to 4. The client device may include at least one data processor and at least one data and / or communication interface preferably configured for providing and / or receiving one or more of the first and / or second data, i.e. field data, weather data, and one or more vegetation indices.

[0037] An eighth aspect provides a system comprising a server or cloud environment according to the fifth aspect and a client device according to the seventh aspect. The system may be a distributed computer system and the computational steps may be divided between the distributed components. Also, the above apparatus of the third aspect may be part of the system or may be distributed between different components of the system.

[0038] It should be noted that several embodiments of the present invention are described with reference to different subject matters. In particular, some embodiments are described with reference to method type claims, while other embodiments are described with reference to apparatus or device type or system type claims. However, a person skilled in the art will infer from the above and following description that, unless otherwise noted, any combination of features belonging to one type of subject matter, as well as any combination between features relating to different subject matters, is also disclosed in the present application. Moreover, the combination of all features can achieve more synergistic effects than the mere sum of the features.

[0039] These and other aspects of the invention will be apparent from and elucidated with reference to the embodiments described hereinafter.

[0040] Exemplary embodiments of the invention are described below with reference to the following drawings: [Brief description of the drawings]

[0041] [Figure 1] 1 shows a schematic block diagram of the principle of determining the onset and / or onset time of a plant disease in agriculture according to the present disclosure. [Diagram 2] 1 shows in a schematic block diagram the principle of determining the onset and / or onset time of plant diseases in agriculture using remote sensing. [Diagram 3] FIG. 3 shows a schematic block diagram of a use case for applying the principle of determining the onset and / or onset time of a plant disease in agriculture according to FIG. 1 or FIG. 2. [Figure 4] FIG. 3 shows a schematic block diagram of a use case for applying the principle of determining the onset and / or onset time of a plant disease in agriculture according to FIG. 1 or FIG. 2. [Diagram 5] FIG. 1 shows a flow diagram of a computer-implemented method for determining the onset and / or onset time of a plant disease in agriculture in accordance with the present disclosure. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0042] The drawings are only schematic representations and serve only to illustrate the invention. Identical or equivalent elements have been given the same reference numbers throughout.

[0043] Figure 1 shows in a schematic block diagram the principle of determining the onset and / or onset timing of a plant disease in agriculture according to the present disclosure. The principle shown in Figure 1 can be implemented on a computer and applied to the methods, apparatus, devices and systems described herein.

[0044] The starting point of the principle is the field 10, or one or more field sections 10a thereof, or one or more plants 12 grown thereon, which can be localized, for example, using geographic coordinates, farmer data, etc. The field 10 or the plants 12 grown thereon may become infected with a plant disease between its sowing and its harvesting, which may reduce or impair the yield. Determining the onset and / or onset time of a plant disease as described herein helps, for example, to detect and / or control the disease at an early stage, when it can be more easily controlled and the yield can be maintained or increased.

[0045] As shown in Figure 1, the principle of determining the onset and / or onset time of a plant disease relies on field data 14 associated with the cultivation of the plant, meteorological data 16 associated with the location of the plant 12, i.e. the location of the field 10 or the respective field section 10a etc, and vegetation index data 18 indicative of the vegetation condition and / or amount. The field data 14 and meteorological data 16 may be referred to as first data and the vegetation index data 18 may be referred to as second data.

[0046] The field data 14, weather data 16, and vegetation index data 18 are provided to a computer model 20 that is trained to determine and output predictions of plant disease presence and spread for plants.

[0047] The output of the computer model 20, including the above-mentioned plant disease presence predictions, which may include disease severity predictions over time, and the vegetation index data 18, are processed, for example, by a data processor 30. The computer model 20 may be run by the data processor 30 or by any other suitable computing means. The data processor 30, and further the computer model 20, may be configured to receive or provide data such as the results of the calculations or the above-mentioned data, for example via a data interface, indicated by an arrow in Figure 1.

[0048] In particular, the output of computer model 20, which includes the plant disease presence prediction described above, is matched with or derived from changes in one or more vegetation indices of vegetation index data 18 to determine the expected onset and / or time of onset of the plant disease in the plant.

[0049] For example, to use, and in particular to coordinate, a plant disease presence prediction with a change in one or more vegetation indices or a plant disease presence prediction with a change in one or more vegetation indices, the data processor 30 may apply a prognostic rule that is adapted to determine, when the plant disease presence prediction indicates that a plant disease is present and when there is a change, e.g., an increase or decrease, in one or more vegetation indices, as the onset and / or time of onset of the plant disease. The rule may be implemented in computer instructions.

[0050] In other words, using, specifically matching, a plant disease presence prediction with a change in one or more vegetation indices or using a plant disease presence prediction with a change in one or more vegetation indices may include applying a prediction rule that defines that if a plant disease is present in the plant disease presence prediction and there is a change in one or more vegetation indices, a decision of the onset of the plant disease is signaled.

[0051] Therefore, the inference rule can be expressed as follows:

number

number

number

[0052] The field data 14 may include information regarding one or more of the following: the plant's growth stage, the number of days since the plant was sown, the month the plant was planted, the planting date, the location where the plant was grown, and the previous crop.

[0053] Weather data 16 may include, for example, observed data from one or more weather stations, modeled data, etc., which may be obtained from weather stations, a weather station network, a database, or may be provided or obtained via an application programming interface (API).

[0054] Vegetation index data 18 may include one or more vegetation indices such as the Normalized Difference Vegetation Index, NVDI, the Normalized Difference Red Edge Index, NDRE, and the Normalized Difference Water Index, NDWI, etc., and may be provided or obtained via an application programming interface (API).

[0055] FIG. 2 shows in a schematic block diagram the principle of determining the onset and / or onset timing of plant diseases in agriculture, involving the use of remote sensing to obtain vegetation index data 18 from remote sensing.

[0056] In this exemplary embodiment, remote sensing may utilize a satellite or satellite data source 100, such as Sentinel-2, Airbus OneAtlas, Landsat, 7, Landsat, 8, or an aircraft, to obtain sensing data from which one or more vegetation indices can be derived. For example, remote sensing may be performed by utilizing an optical sensor on the satellite or aircraft to obtain one or more images of the field 10, field section 10a, and plants 14. From these images, one or more of the vegetation indices described above may be determined, e.g., calculated, and provided to the computer model 20. For example, the image obtained by remote sensing may be a multispectral image, having multiple bands, such as red, green, blue, near infrared (NIR, or short wave infrared). Combining the individual bands, a vegetation index may be determined.

[0057] 3 illustrates a use case in which the determined onset and / or onset time of a plant disease is provided as information for, for example, a farmer. The farmer may utilize a client device 40, such as any type of user equipment, such as a computer, smartphone, etc., configured to receive a message including the onset and / or onset time of the plant disease and determine the onset and / or onset time of the plant disease according to the principles described herein.

[0058] For example, the computer model 20 and data processor 30 described above may be part of a client device 40. In this case, the client device 40 may comprise a data and / or communication interface configured to obtain, e.g. receive, at least the vegetation index data 18 and the weather data 16, e.g. via a communication network such as the Internet. It is noted that the vegetation index data 18 may also be determined by the client device 18 based on survey data, which may be derived remotely.

[0059] The client device 40, e.g., data processor 30, may be configured to execute a suitable application having a graphical user interface and may be configured to facilitate deriving the field data 16, weather data 16, and vegetation index data 18, which may be entered manually or via a search application or may be received at least in part remotely. The client device 40 may be configured to control the computer model 20 and data processor 30 to determine the onset and / or timing of onset of a plant disease as described above.

[0060] The determined onset and / or onset time may then be signaled to a farmer, i.e., a user of the client device 40, for example via a display, i.e., GUI, loudspeaker, etc. In some embodiments, the determined onset and / or onset time may be used to determine a treatment plan for the plant 12, and / or the field 10 or field section 10a thereof.

[0061] Figure 4 shows in a schematic block diagram another use case, where the determined onset and / or onset time of a plant disease is provided as information for, for example, a farmer. Different from figure 3, in this exemplary embodiment the computer model 20 and the data processor 30 are located in a server or cloud environment 50, which is thus configured to determine the onset and / or onset time of a plant disease as described above. Furthermore, the server or cloud environment 50 may use the determined onset and / or onset time to determine a treatment plan for the plant 12, and / or the field 10 or a field section 10a thereof. The latter may be provided to a client device 40.

[0062] For example, the field data 14 may be obtained, e.g., received, from a client device 40 via a communication interface of a server or cloud environment 50 .

[0063] It should be noted that some tasks for determining the onset and / or timing of plant disease and further processing, e.g. plant treatment planning, signaling to farmers, etc., may be distributed between the client device 40 and the server or cloud environment 50.

[0064] Figure 5 shows in a flow diagram a method for determining the onset and / or timing of onset of a plant disease in agriculture. As before, the method follows the principles described above and may be applied to any of the above-mentioned systems shown in Figures 1 to 4 or combinations thereof.

[0065] In step S110, first data including field data 14 associated with the cultivation of the plant and weather data 16 associated with the location where the plant is cultivated is provided to the computer model 20. As described above, the field data 14 may be obtained from the client device 40, or another field in-situ device, or observation device or station, and may be recorded in a database or the like. Additionally, the weather data 16 may be obtained from a weather station, a weather data provider, or the like.

[0066] In step S120, a plant disease presence prediction and plant disease spread for the plant 12 is determined using the computer model 20. For example, the data processor 30 may run the computer model 20 or any other suitable computing device.

[0067] In step S130, from the output of the computer model 20 including the plant disease presence prediction and second data including one or more vegetation indices associated with the plant, i.e., vegetation index data 18, a predicted onset and / or time of onset of the plant disease in the plant is determined by using the plant disease presence prediction and changes in the one or more vegetation indices or using the plant disease presence prediction together with changes in the one or more vegetation indices, in particular by aligning the plant disease presence prediction with changes in the one or more vegetation indices.

[0068] As mentioned above, using, and in particular aligning, a plant disease presence prediction with a change in one or more vegetation indices or a plant disease presence prediction with a change in one or more vegetation indices may involve applying the above estimation rules to determine where a plant disease is present in the plant disease presence prediction and determining, if there is a change in one or more vegetation indices, e.g. an increase or decrease, as the onset and / or time of onset of the plant disease.

[0069] Furthermore, as described above, the onset and / or onset time determined as output data may be used for plant protection or treatment triggers, alarms and / or schedules.

[0070] While the invention has been illustrated and described in detail in the drawings and the foregoing description, such illustration and description are to be considered as illustrative or exemplary and not restrictive. The invention is not limited to the disclosed embodiments. Other variations to the disclosed embodiments can be understood and effected by those skilled in the art in practicing the claimed invention, from a study of the drawings, the disclosure, and the dependent claims.

Claims

1. A computer-implemented method for determining the onset and / or onset time of a plant (12) disease in agriculture, said method comprising: providing (S110) first data to a computer model (20) including field data (14) associated with the cultivation of the plant and meteorological data (16) associated with the location where the plant is cultivated; and using the computer model (20) to determine (S120) a plant disease presence prediction for the plant and a spread of the plant disease; from the plant disease presence prediction output by the computer model (20) and from second data (18) including one or more vegetation indices associated with the plant, determining the onset and / or onset time when the plant disease is predicted to occur in the plant by using the plant disease presence prediction and changes in the one or more vegetation indices.

2. Using the plant disease presence prediction and the change in one or more vegetation indices, 2. The method of claim 1, further comprising applying a prediction rule, the prediction rule being applied to determine as the onset and / or onset time of the plant disease when the plant disease is present in the plant disease presence prediction and when there is a change in the one or more vegetation indices.

3. The estimation rule is: [0010] It is expressed by During the ceremony, [0025] is the gradient of the prediction of the computer model, [0030] 3. The method of claim 2, wherein x is the gradient of the one or more vegetation indices and * is a logical AND operator.

4. Using the plant disease presence prediction and the change in one or more vegetation indices, 4. The method according to claim 1, comprising convolving a differential operator with the one or more vegetation indices in the plant disease presence prediction comprising changes over time of a plurality of plant disease presence predictions, and determining a point having a maximum gradient as the onset and / or onset time of the plant disease.

5. 5. The method of claim 1, wherein the one or more vegetation indices include one or more of a Normalized Difference Vegetation Index, NVDI, a Normalized Difference Red Edge Index, NDRE, and a Normalized Difference Water Index, NDWI.

6. The method of any one of claims 1 to 5, wherein the second data and / or the one or more vegetation indices are derived from remotely sensed data.

7. The method of claim 6 , wherein the remotely sensed data is obtained from one or more satellites and / or one or more aircraft.

8. The method of claim 6 or 7, wherein the remotely sensed data comprises one or more multispectral images.

9. 9. The method according to any one of claims 1 to 8, wherein the field data comprises information on one or more of the following: the growth stage of the plant, the number of days since sowing of the plant, the month of planting of the plant, the planting date, the growing location of the plant, and previous crops.

10. 10. The method according to any one of claims 1 to 9, wherein the weather data comprises information on one or more of the following: wind speed, average wind speed, relative humidity, average relative humidity, maximum relative humidity, air temperature, maximum air temperature, precipitation, and minimum air temperature at a height of 5 cm.

11. 11. The method of any one of claims 1 to 10, wherein the computer model is trained using training data including training data fitted to the first data but not training data fitted to the second data prior to use of the computer model for predicting plant disease presence.

12. The method according to any one of the preceding claims, wherein the first data and / or the second data are provided in a tabular structure.

13. 13. The method according to any one of claims 1 to 12, further comprising providing the determined onset and / or onset time as output data for use for plant protection or treatment triggers, alarms and / or schedules.

14. Use of the onset and / or onset time determined by the method according to any one of claims 1 to 13 for scheduling the treatment of plants in agriculture.

15. 1. An apparatus for determining the onset and / or onset time of a plant disease in agriculture, said apparatus comprising: A data processor (30); a data interface connected to the data processor; the data interface is configured to receive first data including field data associated with growing the plant and weather data associated with a location where the plant is grown, and second data including one or more vegetation indices associated with the plant; the data processor (30) is configured to execute a computer model (20) for determining a plant disease presence prediction and a spread of the plant disease for the plant based on the received first data; The data processor (30) is configured to determine, based on the output of the computer model including the disease presence prediction and the second data, the onset and / or time of onset at which the plant disease is expected to develop in the plant by using the disease presence prediction and changes in the one or more vegetation indices.

Images