Image monitoring for control of invasive grasses
The invention relates to weed mapping, monitoring and control, in particular to the mapping, monitoring and control of invasive annual grasses. A computer system comprises a receiving unit providing a processing unit with images of a geographical area, the images displaying the geographical area at points in time during a phenological cycle of a weed. At least two images depicting a weed at two different phenological stages may be used because two sequentially collected images are required in order to identify a characteristic temporal change. The processing unit analyzes the images to identify image areas showing a spectral signature characteristic of the weed. The processing unit identifies geographical sub-areas corresponding to the identified image areas and to create a weed distribution map with the identified geographical sub-areas marked as areas affected by the weed. The output unit is configured to output the map.
1 . A computer system, the computer system comprising:
a receiving unit;
a processing unit; and
an output unit;
wherein the receiving unit is configured to provide the processing unit with at least two images of a geographical area, wherein the images display the geographical area at two or more points in time during a phenological cycle of a weed, including a first point in time and a second point in time, the first point in time being a time point at which the weed is at a first phenological stage and the second point in time being a time point at which the weed is at a second phenological stage;
wherein the processing unit is configured to analyze the images to identify one or more image area(s) in the images the image area(s) showing a spectral signature, wherein the spectral signature is characteristic of the weed at the two or more points in time during the phenological cycle of the weed, and wherein the spectral signature is caused by a transition of the weed from the first phenological stage into the second phenological stage;
wherein the processing unit is configured to identify one or more geographical subarea(s) corresponding to the identified image area(s);
wherein the processing unit is configured to create a weed distribution map of the geographical area on which the identified geographical subarea(s) is/are marked as (an) area(s) affected by the weed; and
wherein the output unit is configured to output the map.
2 . The computer system according to claim 1 , wherein the weed is a grass weed.
3 . The computer system according to claim 1 , wherein the weed is downy brome being in the phenological stage in which the color of the weed is purple, or the weed is ventenata being in the phenological stage in which the color of the weed is white with a silvery sheen, or the weed is medusahead being in the phenological stage in which the weed is strawcolored.
4 . The computer system according to claim 1 , wherein the weed is downy brome changing its color from green to purple or from purple to straw-color, or the weed is ventenata changing its color from green to white with a silvery sheen, or the weed is medusahead changing its color from green to straw-color.
5 . The computer system according to claim 1 , wherein the processing unit is configured to feed the at least two images into a classifier, and to receive, as an output from the classifier, information about one or more subareas showing the one or more graphical subareas infested by the weed.
6 . The computer system according to claim 1 , wherein the processing unit is configured to feed the at least two images into an-artificial neural network, and to receive, as an output from the artificial neural network, information about one or more subareas showing the one or more graphical subareas infested by the weed.
7 . The computer system according to claim 6 , wherein the artificial neural network comprises convolutional layers and feedback cycles.
8 . The computer system according to claim 1 , wherein the processing unit is configured to determine a degree of infestation of the identified geographical subarea.
9 . The computer system according to claim 1 wherein the processing unit is configured to determine a criticality score for the identified geographical subarea, the criticality score being representative of a risk of a fire outbreak caused by the identified geographical subarea.
10 . The computer system according to claim 1 , wherein the processing unit is configured to determine a criticality score for the identified geographical subarea, the criticality score being representative of a potential amount of damages caused by the identified geographical subarea.
11 . The computer system according to claim 1 , wherein the processing unit is configured to calculate a cost for treating the identified geographical subarea with a herbicide.
12 . The computer system according to claim 1 , wherein the processing unit is configured to generate a prescription map on the basis of the weed distribution map, wherein the prescription map displays areas to be treated with a herbicide.
13 . The computer system according to claim 12 , wherein the areas to be treated are selected on the basis of a criticality score and/or a cost for treating.
14 . A system for controlling weed, the system comprising;
an image acquisition unit, wherein the image acquisition unit is configured to collect at least two while the image acquisition unit is moved relative to a geographical area, wherein the images display the geographical area at two or more points in time during a phenological cycle of a weed, including a first point in time and a second point in time, the first point in time being a time point at which the weed is at a first phenological stage and the second point in time being a time point at which the weed is at a second phenological stage;
a spraying unit, comprising a receptable containing a herbicide for controlling a weed;
locomotion means for moving the image acquisition unit and the spraying unit relative to the geographical area; and
a processing unit configured to:
receive the at least two images from the image acquisition unit;
analyze the received and identify one or more areas in the images based on a spectral signature, the area(s) in the images showing geographical subareas, the geographical subareas being infested by the weed, wherein the spectral signature is characteristic of the weed at the two or more points in time during the phenological cycle of the weed, and wherein the spectral signature is caused by a transition of the weed from the first phenological stage into the second phenological stage; and
cause the spraying unit to apply the herbicide to the geographical subareas.
15 . A method for mapping a weed, the method comprising:
receiving at least two images of a geographical area, wherein the images display the geographical area at two or more points in time during a phenological cycle of a weed, including a first point in time and a second point in time, the first point in time being a time point at which the weed is at a first phenological stage and the second point in time being a time point at which the weed is at a second phenological stage;
analyzing the at least two and identifying one or more image area(s) in the at least two images, the image area(s) showing a spectral signature, wherein the spectral signature is characteristic of the weed at the two or more points in time during the phenological cycle of the weed, and wherein the spectral signature is caused by a transition of the weed from the first phenological stage into the second phenological stage;
identifying one or more geographical subareas, the geographical subarea(s) corresponding to the identified image area(s);
creating a weed distribution map of the geographical area on which the identified geographical subarea(s) is/are marked as (an) area(s) affected by the weed;
displaying the weed distribution map on a monitor and/or storing the map in a data storage and/or transmitting the map to a separate computer system.