[0001] The present invention relates to a weeding apparatus arranged to be installed on
a railway vehicle for weeding and a method for weeding railway tracks which avoid
the growth of weed species near the railway tracks.
[0002] In particular, the invention relates to an apparatus for weeding which uses software
implementing artificial intelligence logic to optimize the weeding procedure, so that
it identifies the weeds species and the position thereof.
[0003] Moreover, the invention relates to a method to perform the weeding of said weed species
with an apparatus which can be installed on railway vehicles, which optimizes the
weeding modes, trying to avoid waste and errors.
[0004] In the known art the weeding apparatuses arranged to be installed on railway vehicles
can avail themselves of software based on artificial intelligence logic (AI) which
provide the processing of data arrays, which suitably re-processed provide as output
a probability function of presence of weed species.
[0005] Based on a probability function, a control unit actuates a weeding unit by optimizing
the operation thereof thereby using the weed-killing agent in the required amount
and in the necessary points, defined by the presence of weed species.
[0006] The data arrays processed by the AI software are generally extracted from a digital
input provided by sensors comprised in the apparatus, for example video cameras.
[0007] The types of sensors can vary depending upon the apparatuses, apart from video cameras
which detect the visible spectrum, even infrared video cameras, tachometers, accelerometers,
and so on, may be present.
[0008] Unfortunately, in the known art the arrays are re-processed, preferably through convolution
procedures, without using different multiplicative factors for the different arrays,
based upon considerations regarding the provenance of data and the data quality depending
upon the external conditions which could affect said quality.
[0009] In this way, the data fusion could not be optimum, by leading to imprecise or redundant
interventions with consequent waste of weed-killing agent or to the lack of intervention
on some weed species.
[0010] Therefore, it results clear how important it could be to reprocess the data arrays,
so that the latter are recombined with different multiplicative factors depending
on their quality, which can be determined by different external factors, hereby avoiding
waste and malfunctions.
[0011] The technical problem underlying the present invention is to provide a weeding apparatus
arranged to be installed on a railway vehicle for weeding and a method for weeding
railway tracks allowing to obviate the drawbacks mentioned with reference to the known
art.
[0012] In a first aspect thereof, such problem is solved by a weeding apparatus arranged
to be installed on a railway vehicle for weeding as defined in the enclosed claim
1.
[0013] This weeding apparatus is specifically arranged to be installed on a railway vehicle
for weeding.
[0014] It comprises a first camera which is configured to detect at least the visible spectrum,
and it is apt to capture a first digital image of a railway track portion and said
first image is expressed digitally through a first array of pixel values.
[0015] Moreover, a second camera is provided, which is configured to detect at least the
infrared spectrum, and it is then configured to capture a second digital image of
a railway track portion, and said second image, analogously to the first one, is expressed
digitally through a respective second matrix of pixel values.
[0016] The apparatus then comprises a LIDAR type remote-sensing device, which operates with
a laser emitter and it is configured to capture a digital scan of a railway track
portion, and said digital scan is expressed digitally through a respective third matrix
of pixel values.
[0017] Further to the above-mentioned devices to detect images, the apparatus comprises
a weeding unit, provided for weeding weed species, and a control module implementing
a first software which in turn is capable of receiving a weed species detection signal
and triggering the spraying of chemical agent from said supply module through at least
a nozzle, based upon the weed species detection signal.
[0018] Moreover, a unit for the identification of the weed species is provided which, in
turn, comprises a communication module arranged to receive the first digital image,
the second digital image and the digital scan and arranged to send a weed species
detection signal to said control module of the weeding apparatus. It further comprises
a memory module, configured to store said first and second image and said scan, and
a processing module having a second software arranged to carry out at least a convolution
operation between a pixel array of each digital image and of the digital scan and
a respective predefined kernel stored in the memory module to obtain an array representing
the features of each digital image and of the digital scan.
[0019] The processing module, through said second software, can calculate at least a first,
a second and a third presence probability and position of a weed species, in a database
of weed species, said first, second and third presence probability and position being
respectively associated to said first digital image, to said second digital image
and to said scan, so that said processing module detects the presence and position
of said weed species.
[0020] At this point, the processing module generates a weed species detection signal based
upon a resulting presence probability and position, calculated as the sum of the first,
second and third presence probability and position, respectively multiplied by a first,
a second and a third multiplicative factor defined based upon the presence of precipitations
and brightness; moreover, the processing module sends said detection signal to the
control module which is arranged to activate said weeding unit, so that it acts only
in areas in which weed species have been detected.
[0021] The main advantage of the weeding apparatus lies in the fact of multiplying each
array representing the features corresponding to a different source of data, in particular
to the first, to the second digital image and to the digital scan, by a multiplicative
factor which is customized based upon the weather conditions and time of the day (data/night).
In this way, more weight is given to more reliable data; for example, during the day
and in absence of rain, the first camera is to be considered very reliable, possibly
more reliable than the second camera and the remote-sensing device; differently, at
night and in presence of atmospheric precipitations or heavy fog, the second camera
is more reliable than the first camera and the remote sensing device; or at night
and under serene conditions, the remote-sensing device is considered the most reliable
one.
[0022] In a second aspect thereof, such problem is solved by a method for weeding railway
tracks as defined in the enclosed claim
9.
[0023] The method for weeding railway tracks provides to use a weeding apparatus arranged
to be installed on a railway vehicle as it is defined in said first aspect.
[0024] Moreover, in the specific case, the weeding method provides the steps of capturing
a first digital image in the visible of a railway track portion, a second digital
image in the infrared of a railway track portion and a laser scan of a railway track
portion, each one of said images and said scan being expressed with a respective pixel
array.
[0025] Then, through said communication module the captured images and the scan are sent
to a processing and storing module and, through said second software, a convolution
is performed between a pixel array of each digital image and of the digital scan and
a respective kernel stored in the memory module to obtain an array representing the
features of each digital image and of the digital scan.
[0026] Subsequently, through said processing module implementing said second software, one
calculates at least a presence probability and position of at least a weed species
in a database of weed species, from an array representing the features of the image,
by detecting the presence and position of said weed species. Through said processing
module, a first, a second and a third multiplicative factor defined based upon the
presence of precipitations and brightness are then calculated.
[0027] The first multiplicative factor is then attributed to a first presence probability
and position of the first digital image, said second multiplicative factor to a second
presence probability and position of the second digital image, and said third multiplicative
factor to a third presence probability and position of the digital scan.
[0028] A resulting presence probability and position is then calculated, through said software,
as the sum of the first, second and third presence probability and position, respectively
multiplied by the first, second and third multiplicative factor, and then, through
said processing module, a weed species detection signal is generated, based upon the
above-mentioned resulting probability, which is sent to the control module of the
weeding apparatus for triggering the weeding procedures through said weeding unit.
[0029] Even the method like the apparatus has the advantage of weighing the received data
with multiplicative factors which can be different therebetween.
[0030] The present invention will be described hereinafter according to a preferred embodiment
example thereof, provided by way of example and not for limitative purposes with reference
to the enclosed drawings wherein:
* figure 1 shows a block diagram of a weeding apparatus according to the invention;
* figure 2 shows the detection of an intervention area;
* figure 3 shows the detection of additional intervention areas apart from the intervention
area of figure 2;
* figure 4a shows a schematic representation of the processing of data input to the
weeding apparatus;
* figure 4b shows a schematic representation of the data convolution process;
* figure 5 shows a front schematic representation of a railway vehicle on which the
weeding apparatus according to figure 1 is installed; and
* figure 6 shows a top schematic representation of a railway vehicle on which the
weeding apparatus according to figure 1 is installed.
[0031] With reference to the figures, a weeding apparatus, arranged to be installed on a
railway vehicle for weeding, is designated as a whole with
1.
[0032] The apparatus 1 comprises a first camera 3 configured to detect at least the visible
spectrum and being apt to capture a first digital image of a railway track portion
4, said first image comprising a first array of pixel values.
[0033] The apparatus 1 then comprises a second camera 5 configured to detect at least the
infrared spectrum and being apt to capture a second digital image of a railway track
portion 4, said second image comprising a second array of pixel values.
[0034] The apparatus 1 further comprises a remote-sensing device 6, of the LIDAR type, comprising
a laser emitter, the remote-sensing device being apt to capture a digital scan of
a railway track portion 4, said digital scan comprising an array of pixel values.
[0035] The apparatus 1 even comprises a weeding unit arranged for weeding weed species.
[0036] At last, the apparatus 1 comprises a control module 12, implementing a first software
arranged to receive a weed species detection signal and to trigger weeding through
said weeding unit based upon said weed species detection signal and a unit 13 for
the identification of the weed species.
[0037] This identification unit 13, in turn, comprises, a communication module 14 which
is arranged to receive, the first digital image, the second digital image and the
digital scan and to send a weed species detection signal to said control module 12
of the weeding apparatus 1.
[0038] It further comprises a memory module 15 configured to store said first and second
image and said scan, and a processing module 16 comprising a second software which
is arranged to carry out at least a convolution operation between a pixel array of
each digital image and of the digital scan and a respective predefined kernel stored
in the memory module 15, with the purpose of obtaining an array representing the features
of each digital image and of the digital scan.
[0039] The processing module 16, through said second software, calculates at least a first,
a second and a third presence probability and position of a weed species, in a database
of weed species.
[0040] The first, second and third presence probability and position are respectively associated
to said first digital image, to said second digital image and to said scan, so that
said processing module 16 detects the presence and position of said weed species.
[0041] The processing module 16 generates a weed species detection signal based upon a resulting
presence probability and position, calculated as the sum of the first, second and
third presence probability and position, respectively multiplied by a first multiplicative
factor F
1, a second multiplicative factor F
2 and a third multiplicative factor F
3. The multiplicative factors are defined based upon the presence of precipitations
and the brightness conditions.
[0042] The processing module 16 sends the detection signal to the control module 12 which,
in turn, is arranged to activate the weeding unit, so that it acts only in areas in
which weed species have been detected.
[0043] Alternatively, according to another embodiment of the invention which integrates
what described previously, the processing module 16 can comprise a plurality of parallel
processing cores, and each parallel processing core is arranged to carry out, through
said second software, at least a convolution operation between a sub-array constituted
by near pixels of each digital image and of the digital scan and a respective predefined
kernel, which is stored in the memory module 15, to obtain a sub-array representing
the features of each digital image and of the digital scan by assigning a value to
each pixel of the image.
[0044] Preferably, the weeding apparatus comprises an external storage space 100, in case
even in
cloud, on which the data collected when said weeding apparatus 1 is in use are saved and
stored for future uses.
[0045] Moreover, the processing module 16, through the second software, can calculate at
least a presence probability and position of a weed species in a database of weed
species from each array representing the features of the image constituted by the
sub-arrays for representing the features, in turn constituted by the parallel processing
cores, by detecting the presence and position of said weed species, and it can generate
a weed species detection signal based upon a resulting presence probability and position,
calculated as the sum of the first, second and third presence probability and position,
respectively associated to said first digital image, to said second digital image
and to said scan, respectively multiplied by a first, a second and a third multiplicative
factor defined based upon the presence of precipitations and brightness.
[0046] This detection signal is then transmitted to the control module which is arranged
to activate the spraying unit, so that it acts only in areas in which weed species
have been detected.
[0047] According to some embodiments, the weeding apparatus 1 comprises a first device for
detecting brightness arranged to identify a brightness parameter λ, wherein, if λ
> λ
min the presence of light, that is a brightness sufficient to take images, is considered
and otherwise it is dark.
[0048] Moreover, a second device for detecting precipitations in form of rain and/or hail
and/or snow is provided, which is arranged to identify an atmospheric parameter µ,
wherein if µ > µ
min one establishes that there are precipitations so as to influence the weeding process,
or otherwise it is clear.
[0049] If the first and the second detection device are present, the communication module
can be arranged to receive the brightness parameter λ and the atmospheric parameter
µ and wherein said processing module 16, through said software, establishes said first,
second and third multiplicative factors based upon the brightness parameter λ and
the atmospheric parameter µ,
wherein
- if λ > λmin and µ < µmin then F1 > F2 ≥ F3, or
- if λ > λmin and µ > µmin then F2 > F1 > F3, or
- if λ < λmin and µ < µmin then F3 ≥ F2 > F1, or
- if λ < λmin and µ > µmin then F2 > F3 > F1.
[0050] In this way, based upon the atmospheric conditions and brightness, it is possible
to define three multiplicative factors F
1, F
2, F
3, which will weigh the respective arrays depending on the greater reliability of each
one thereof.
[0051] In fact, in presence of rain, heavy fog, snow or hail the scan obtained with the
remote-sensing device 6 is to be considered less reliable than the first and second
digital image which, in turn, could be considered with different multiplicative factors
depending on whether they are detected during the day or at night.
[0052] According to the embodiments shown in figure, the weeding unit comprises a spraying
unit 7 having a supply module 8 with a tank 9 of chemical agent and a dispensing valve
10, and at least a nozzle 11, to spray a chemical agent from said supply module 8.
[0053] The control module 10 is arranged to receive a weed species detection signal and
to command the spraying of chemical agent from the supply module 8 through the nozzle
11 based on the weed species detection signal.
[0054] Moreover, the processing module 16 sends the detection signal to the control module
12 which is arranged to activate the spraying unit 4, so that it acts only in areas
in which weed species have been detected.
[0055] According to other preferred embodiments of the invention, the processing module
16, through said second software, calculates at least a resulting probability which,
apart from the presence and position, considers the type of at least one weed species,
by detecting, indeed, the presence, position and type of the weed species.
[0056] The processing module 16 then generates a weed species detection signal based upon
the resulting presence probability, position and type, and sends said detection signal
to said control module 12 arranged to activate, through the second software, the weeding
unit, so that it acts only in areas in which weed species have been detected and preferably
with a dosage of chemical agent suitable to the type of weed species.
[0057] Apart from a customized dosage, for weeding specific types of weed species, the supply
module 8 can comprise at least two tanks, each one with a chemical agent different
from the other ones, and the processing module, through said second software, can
generate a signal for detecting the different weed species based upon said resulting
presence probability, position and type and sending said detection signal to said
control module 12 arranged to activate the spraying unit 7, thereby different chemical
agents are sprayed based upon the weed species type.
[0058] According to additional preferred embodiments of the invention, the weeding unit
comprises a weeding device arranged to emit an electric discharge and/or to emit a
laser beam.
[0059] When the weeding device is present, the control module 10 is arranged to receive
a weed species detection signal and to command the activation of said weeding device,
based upon said weed species detection signal and wherein said processing module 16
sends said detection signal to said control module 12 which is arranged to activate
said weeding device, so that it acts only in areas in which weed species have been
detected.
[0060] When both the spraying unit 7 and the weeding device are present, the processing
module 16, through said second software, calculates at least a resulting presence
probability, position, type of at least one weed species, by detecting the presence,
position and type of said at least one weed species and generates a signal for detecting
the different weed species based upon said resulting presence probability, position
and type and sends said detection signal to said control module 12 arranged to activate
selectively, through said software, the spraying unit 7 and/or the weeding device
based upon which one is considered to be most suitable for weeding a weed species
type, in a determined position, and so that they act only in areas wherein in which
the weed species have been detected.
[0061] In fact, apart from the weed species type, in this case even the position is relevant,
since, preferably, the weeding device is positioned below the railway vehicle
2. This preferential positioning could make the weeding device to be arranged to act
below the railway vehicle 2, then on the tracks 4 and in proximity thereof, without
reaching areas very far from the railway vehicle 2. In particular, this limitation
could involve mainly the possible implementation of the weeding device which acts
through electric discharge.
[0062] According to more advanced embodiments, the weeding apparatus comprises an anemometer
arranged to detect a wind intensity parameter W, and/or a tachymeter arranged to detect
a speed parameter V of the vehicle on which the apparatus travels.
[0063] In the presence of anemometer and/or tachymeter, the communication module 14 can
receive said wind intensity parameter W and/or said speed parameter V and said memory
module can store them and the processing module 16, through said software, verifies
that the conditions W
min ≤ W ≤ W
max and V ≥ V
min are fulfilled, W
min and W
max being threshold values of the wind intensity parameter stored in said memory module
15, and V
min being a threshold value of the speed parameter stored in said memory module 15.
[0064] If the conditions are fulfilled, the control module receives said detection signal
and triggers the weeding of the weed species through said weeding unit, based upon
said weed species detection signal.
[0065] An example of convolution procedures is schematically shown in figure 4, wherein
one obtains: a parameter y ∈ (0,1) designating the activation or not of the weeding
apparatus 1, a parameter P1 and a parameter P2 designating the activation of the spraying
unit 7 and of the weeding device, respectively, and a parameter
lj designating the intensity to be applied to several nozzles.
[0066] To perform weeding of the railway tracks 4 the first and the second software can
avail of artificial intelligence logic.
[0067] A preferred method to perform weeding of the railway tracks 4 provides to activate
the weeding apparatus, capturing a first digital image in the visible of a railway
track portion 4, a second digital image in the infrared of a railway track portion
4 and a laser scan of a railway track portion 4, provides that each one of said images
and said scan is expressed as a respective array of pixel.
[0068] To be able to process the pixel arrays, the processing module 16 receives through
said communication module 14 the captured images and the scan in turn stored in the
memory module 14.
[0069] Preferably, the processing module 16 performs a convolution, through the second software,
between a pixel array of each digital image and of the digital scan and a respective
kernel stored in the memory module 15, to obtain an array representing the features
of each digital image and of the digital scan.
[0070] The method then provides to calculate, through the processing module 16 implementing
the second software, at least a presence probability and position of at least one
weed species in a database of weed species, from an array representing the features
of the image, by detecting the presence and position of said weed species.
[0071] Moreover, the processing module 16 can calculate a first, a second and a third multiplicative
factor defined based upon the presence of precipitations and brightness and can multiplicate
them, respectively, with a first presence probability and position of the first digital
image, with a second presence probability and position of the second digital image
and with a third presence probability and position of the digital scan in order to
calculate a resulting presence probability and position.
[0072] At last, the processing module 16 generates a weed species detection signal based
upon said resulting presence probability and position of a weed species and sends,
through said communication module 14, said weed species detection signal to the control
module 12 which can trigger the weeding of the weed species, through said weeding
unit, based upon the weed species detection signal.
[0073] The weeding unit can comprise the spraying unit 7 and/or the weeding device already
described previously.
[0074] The detection signal can comprise even information about the amount of chemical agent
to be dispensed based upon the weed species type.
[0075] If the spraying unit 7 comprises two or more tanks, each one containing a chemical
agent different from the other ones, the detection signal is based upon a presence
probability position and type and it can comprise information about different week-killers
to be used based upon the weed species type and it triggers said spraying unit 7 thereby
different chemical agents are sprayed based upon the weed species type.
[0076] If the weeding apparatus 1 comprises both the spraying unit 7 and the weeding device,
the processing module can make the selection between which is better to use or whether
to use both of them and sends the detection signal, comprising the selection, to said
control module which can trigger selectively, through said first software, said spraying
unit 7 and/or said weeding device.
[0077] According to some preferred embodiments, wherein the weeding apparatus comprises
an anemometer and a tachymeter, the triggering of the weeding apparatus 1 provides
to measure the wind intensity at a certain time instant t with the anemometer and
storing a wind intensity parameter W(t), measuring the speed of the railway vehicle
at a certain time instant t and storing a speed parameter V(t) in the memory module
15 and communicating them, through said communication module 14, to said processing
module 16.
[0078] At last, the triggering of the weeding apparatus provides to verify, through said
second software implemented on said processing module 16 that the conditions W
min ≤ W(t) ≤ W
max and V(t) ≥ V
min are fulfilled, if these conditions are fulfilled, the weeding apparatus 1 proceeds
with subsequent actions, otherwise it does not proceed.
[0079] When the weeding apparatus 1 comprises a tachymeter, the weeding method can provide
to measure the speed of the railway vehicle 2 and to store a speed parameter V in
the memory module 15 and to communicate it through said communication module 14 to
said processing module 16.
[0080] The memory module 15 can even communicate a time t
e for processing the digital images and the scan and a time t
a for activating the weeding unit, both being saved in the memory module, to said processing
module 16.
[0081] The processing module 16 can calculate, through said second software, a minimum distance
of a detection area 200 X
Dmin = V(t
e+t
a) to be considered when processing the weed species detection signal.
[0082] The fact of identifying the detection area 200 allows to process images and scans
of the surrounding reality present at a distance from the installation point of the
weeding unit on the railway vehicle 2, and this allows the weeding apparatus to act
effectively for weeding the weed species.
[0083] Preferably, the weeding apparatus 1 allows to identify a detection area 200 and neighbouring
areas 201, which correspond to different distances from the installation point of
the weeding unit on the railway vehicle 2.
[0084] To the above-described weeding apparatus and related method a person skilled in the
art, with the purpose of satisfying additional and contingent needs, could introduce
several additional modifications and variants, all however comprised within the protective
scope of the present invention, as defined by the enclosed claims.
1. A weeding apparatus (1) arranged to be installed on a railway vehicle (2) for weeding,
comprising:
• a first camera (3) configured to detect at least the visible spectrum and being
apt to capture a first digital image of a railway track portion (4), said first image
comprising a first array of pixel values;
• a second camera (5) configured to detect at least the infrared spectrum and being
apt to capture a second digital image of a railway track portion (4), said second
image comprising a second array of pixel values;
• a LIDAR type remote-sensing device (6), comprising a laser emitter, the remote-sensing
device being apt to capture a digital scan of a railway track portion (4), said digital
scan comprising a third array of pixel values;
• a weeding unit (77) arranged for weeding weed species;
• a control module (12) implementing a first software arranged to receive a weed species
detection signal and to trigger weeding through said weeding unit based upon said
weed species detection signal;
• a unit (13) for the identification of the weed species comprising:
* a communication module (14) arranged to receive, the first digital image, the second
digital image and the digital scan and to send a weed species detection signal to
said control module (12) of the weeding apparatus (11),
* a memory module (15) configured to store said first and second image and said scan,
and
* a processing module (16) comprising a second software arranged to carry out at least
a convolution operation among the pixel arrays of each digital image and of the digital
scan and a respective predefined kernel stored in the memory module (15) to obtain
a further array representing the features of each digital image and of the digital
scan,
wherein the processing module (16), through said second software, calculates at least
a first, a second and a third presence probability and position of a weed species,
in a database of weed species, said first, second and third presence probability and
position being respectively associated to said first digital image, to said second
digital image and to said scan, so that said processing module (16) detects the presence
and position of said weed species,
and wherein the processing module generates a weed species detection signal based
upon a resulting presence probability and position, calculated as the sum of the first,
second and third presence probability and position, respectively multiplied by a first
(F1), a second (F2) and a third multiplicative factor (F3) defined based upon the presence of precipitations and brightness, and wherein said
processing module sends said detection signal to said control module arranged to activate
said weeding unit so that it acts only in areas in which weed species have been detected.
2. The weeding apparatus (1) according to claim 1, comprising a first device for detecting
brightness and/or a second device for detecting precipitations in form of rain and/or
hail and/or snow, wherein said first detection device is arranged to identify a brightness
parameter λ, wherein if λ > λmin there is light, vice versa it is dark and wherein said second detection device is
arranged to identify an atmospheric parameter µ, wherein if µ > µmin there are precipitations, otherwise it is clear.
3. The weeding apparatus according to claim 2, wherein said communication module (14)
is arranged to receive the brightness parameter λ and the atmospheric parameter µ
and wherein said processing module (16), through said software, establishes said first,
second and third multiplicative factors based upon the brightness parameter λ and
the atmospheric parameter µ,
wherein
- if λ > λmin and µ < µmin then F1 > F2 ≥ F3, or
- if λ > λmin and µ > µmin then F2 > F1 > F3, or
- if λ < λmin and µ < µmin then F3 ≥ F2 > F1, or
- if λ < λmin and µ > µmin then F2 > F3 > F1.
4. The weeding apparatus according to any one of the preceding claims, wherein said weeding
unit has a spraying unit (4) comprising:
* a supply module (8), including a tank of chemical agent (11) and a dispensing valve
(12); and
* at least a nozzle (9) to spray a chemical agent from said supply module (8);
wherein said control module (10) is arranged to receive a weed species detection signal
to command the spraying of chemical agent from said supply module (8) through said
at least a nozzle (11) based upon said weed species detection signal and wherein said
processing module (16) sends said detection signal to said control module (12) which
is arranged to activate said spraying unit, so that it acts only in areas in which
weed species have been detected.
5. The weeding apparatus according to claim 4, wherein said supply module (8) comprises at least two tanks each one with a chemical
agent different from the other ones and wherein said processing module (16), through
said second software, generates a signal for detecting the different weed species
based upon a resulting probability of presence, position and type and sends said detection
signal to said control module (12) arranged to activate said spraying unit (7), thereby
different chemical agents are sprayed based upon the weed species type.
6. The weeding apparatus (1) according to any one of claims 1 to 3, wherein said weeding
unit comprises a weeding device arranged to emit an electric discharge and/or to emit
a laser beam, and wherein said control module (10) is arranged to receive a weed species
detection signal and to command the activation of said weeding device, based upon
said weed species detection signal and wherein said processing module (16) sends said
detection signal to said control module (12) which is arranged to activate said weeding
device, thereby it acts only in areas in which weed species have been detected.
7. The weeding apparatus (1) according to claims 4 and 6, wherein the processing module (16), through said second software, calculates at least
a resulting presence probability, position, type of at least a weed species, by detecting
the presence, position and type of said at least a weed species and, wherein the processing
module (16) generates a signal for detecting the different weed species based upon
said resulting presence probability, position and type and sends said detection signal
to said control module (12) arranged to activate selectively, through said software,
said spraying unit (7) and/or said weeding device based upon which one is considered
to be most suitable for weeding a weed species type, in a determined position, and,
hereby they act only in areas in which the weed species have been detected.
8. The weeding apparatus according to any one of the preceding claims, comprising an
anemometer arranged to detect a wind intensity parameter W, and/or comprising a tachymeter
arranged to detect a speed parameter V.
9. The weeding apparatus according to claim 8, wherein said communication module (14)
is arranged to receive said wind intensity parameter W and said speed parameter V
and, wherein said memory module (15) is configured to store said wind intensity parameter
and said speed parameter and wherein the processing module (16), through said software,
verifies that the conditions Wmin ≤ W ≤ Wmax and V ≥ Vmin are fulfilled, Wmin and Wmax being threshold values of the wind intensity parameter stored in said memory module
(15) and Vmin being a threshold value of the speed parameter stored in said memory module; if the
conditions are fulfilled, said control module (12) receives said detection signal
and triggers the weeding of the weed species through said weeding unit, based upon
said weed species detection signal.
10. A method for weeding railway tracks (4) performed by using a weeding apparatus (1)
arranged to be installed on a railway vehicle (2), the apparatus (1) comprising a
first camera (3) configured to detect at least the visible spectrum, a second camera
(5) configured to detect at least the infrared spectrum, a LIDAR type remote-sensing
device (6), comprising a laser emitter, a weeding unit, a control module (12) implementing
a first software, a unit (13) for the identification of species of weeds including
a communication module (14), a memory unit (15) and a processing module (16) implementing
a second software based upon artificial intelligence logic, the method comprising
the steps of:
* activating said weeding apparatus (1);
* capturing a first digital image in the visible of a railway track portion (4), a
second digital image in the infrared of a railway track portion (4) and a laser scan
of a railway track portion (4), each one of said images and said scan comprising a
pixel array,
* receiving through said communication module (14) the captured images and the scan
in the processing module (16) and storing said images and said scan in the memory
module (15);
* performing, through said second software, at least a convolution among a pixel array
of each digital image and of the digital scan and a respective kernel stored in the
memory module (15) to obtain an array representing the features of each digital image
and of the digital scan;
* calculating, through said processing module (16) implementing said second software,
at least a presence probability and position of at least a weed species in a database
of weed species, from an array representing the features of the image, by detecting
the presence and position of said weed species;
* calculating, through said processing module (16), a first (F1), a second (F2) and a third (F3) multiplicative factor defined based upon the presence of precipitations and brightness;
* attributing, through said processing module, said first multiplicative factor (F1) to a first presence probability and position of the first digital image, said second
multiplicative factor (F2) to a second presence probability and position of the second digital image, said
third multiplicative factor (F3) to a third presence probability and position of the digital scan;
* calculating a resulting presence probability and position, through said software,
as the sum of the first, second and third presence probability and position, respectively
multiplied by the first (F1), second (F2) and third (F3) multiplicative factor;
* generating, through said processing module (16), a weed species detection signal
based upon said resulting presence probability and position of a weed species and
sending, through said communication module (14), said detection signal of weed species
to the control module (12); and
* triggering, through said control module (12) the weeding of the weed species through
said weeding unit based upon the weed species detection signal.
11. The method according to claim 10, wherein said weeding unit comprises:
* a spraying unit (7) including:
* a supply module (8), including a tank of chemical agent (11) and a dispensing valve
(12);
* at least a nozzle (9) to spray a chemical agent from said supply module (8); and
* a weeding device arranged to emit an electric discharge and/or to emit a laser beam,
wherein the method comprises the steps of:
* calculating, through said processing module, a resulting presence probability, position
and type of at least a weed species, by detecting the presence, the position of said
at least a weed species;
* generating, through said processing module, a weed species detection signal based
upon said resulting presence probability, position and type;
* selecting, through said second software, which one between said spraying unit and/or
said weeding device is most suitable for weeding a weed species type and including
the information of the selection in said detection signal; and
* sending said detection signal, through said communication module to said control
module arranged to activate selectively, through said software, said spraying unit
and/or said weeding device based upon the type and the position of the weed species,
so that they act only in areas in which the weed species have been detected.
12. The method according to claim 10 or 11, wherein the weeding apparatus comprises an
anemometer and a tachymeter, and wherein the activation of said weeding apparatus
comprises the steps of:
* measuring the wind intensity at a certain time instant t with said anemometer and
storing a wind intensity parameter W(t);
* measuring the speed of the railway vehicle at a certain time instant t and storing
a speed parameter V(t) in the memory module and communicating them through said communication
module to said processing module;
* verifying, through said second software implemented on said processing module that
the conditions Wmin ≤ W(t) ≤ Wmax and V(t) ≥ Vmin are fulfilled, if these conditions are fulfilled, the weeding apparatus proceeds
with subsequent actions, otherwise it does not.