|
(11) | EP 3 643 594 A1 |
| (12) | EUROPEAN PATENT APPLICATION |
|
|
|
|
|||||||||||||||||||||||||||
| (54) | METHOD AND SYSTEM FOR REMOVING FLOATING WASTE FROM WATER |
| (57) Described is a method for removing floating undesired objects from a water mass in
an area. The method comprises of determining an entry of an object into the area.
Properties of the object, which show the undesirability of the object, are then determined
with image recognition. After this, undesired objects are carried in floating state
to a temporary storage, preferably with an autonomously navigating vessel, and then
removed therefrom, preferably once again with an autonomously navigating vessel. A
system with which the method can be performed is also described. The method preferably
proceeds under the control of a computer via a digital network.
|
TECHNICAL FIELD OF THE INVENTION
[0001] The present invention relates to a method for removing floating undesired objects from a water mass in an area. The invention also relates to a system for removing floating undesired objects from a water mass in an area, and to a computer program comprising program instructions for having a computer execute the method. The invention further relates to a computer configured to execute such a computer program.
BACKGROUND OF THE INVENTION
[0002] Water masses such as seas, but also rivers, canals and brooks, are subject to pollution. A known problem is for instance the so-called plastic soup floating around our waterways and seas. Not only is this disastrous for the environment, it can also form an obstacle to shipping traffic. Propellers of ships can for instance be rendered defective by objects of plastic and/or another material becoming caught therein. In rivers and other waterways, wooden objects floating around therein can cause great damage to ships and other waterworks, for instance locks.
[0003] A known method is laborious and ineffective, and comprises of tracking down undesired objects floating around in a waterway with a manned vessel and taking up and removing an object with the vessel, which is for this purpose for instance provided with a grabbing installation.
SUMMARY OF THE INVENTION
[0004] The present invention has for its object to provide a more efficient method and corresponding system for removing floating undesired objects from a water mass in an area.
[0005] These and other objects are achieved by providing a method with the features as stated in claim 1. Particularly provided is a method for removing floating undesired objects from a water mass in an area, comprising the steps of:
- a) determining an entry of an object into the area;
- b) determining properties of the object, which show the undesirability of the object, with image recognition;
- c) carrying an undesired object in floating state to a temporary storage; and
- d) removing undesired objects arranged in the temporary storage.
[0006] The steps of the method are preferably performed under the control of a computer via a digital network.
[0007] Undesired objects are understood to mean objects which do not belong in a water mass and which come from outside the water mass. Examples are (plastic) waste, wood, domestic refuse, containers, and liquids, such as for instance oil, which are foreign to the water mass. The term floating can also comprise a state in which an object hovers in the water mass, possibly below the water surface.
[0008] A useful embodiment of the invention comprises a method wherein the temporary storage floats. Such an embodiment is particularly useful when removing objects from a water mass which is subject to tidal current. A system according to an embodiment in which the temporary storage floats does indeed remain usable at a different water levels and can for instance position itself autonomously relative to these water levels.
[0009] The floating objects lie in the water and, depending on for instance their specific gravity, can float therein but also hover under the water surface. Such objects are also referred to as floating objects.
[0010] Determining of an entry of an object into the area can take place in many ways. Firstly, an area is determined in which the floating undesired objects must be removed from the water mass. This area of course depends on the specific task, but will generally in any case be characterized by a number of boundaries. In some cases, in which the water mass is a water mass subject to tidal current, it is possible that floating objects will enter the area along with the average tide across an entry boundary. Such an entry boundary is then determined by the direction of the (tidal) current, wherein the (tidal) current runs such that the area is situated downstream of the entry boundary.
[0011] The entry of an object into the area can be detected in practical manner by providing a means suitable for this purpose at the position of the entry boundary. Suitable means comprise but are not limited to optical devices such as digital cameras, mechanical feelers, measuring devices which measure distances on the basis of sound, and so on. Providing at least one means for detecting an object floating in the water mass at the position of the entry boundary makes it possible to determine if, when and at which position along the entry boundary an object enters the area. The means configured to detect an object floating in the water mass can preferably also operate at night, or can operate with additional lighting.
[0012] According to the invention, it is then determined whether the detected object is a desired object or an undesired object. A desired object can remain floating or hovering in the water mass because such an object cannot cause damage or otherwise harm the environment. An undesired object must preferably be removed from the water mass because it could possibly cause damage or otherwise harm the environment further downstream.
[0013] The method provides a step comprising of determining properties of the object, which show the undesirability of the object, with image recognition. The means for detecting an object floating in the water mass is for this purpose connected, optionally via the digital network, to a computer comprising image recognition software. Such software is per se known and commercially available. The image recognition software is configured to determine properties of a passing object, such as the size, the elongation, the shape, and essentially any property which determines the desirability or, preferably exclusively, the undesirability of the object.
[0014] In an embodiment of the invention it is also possible to characterize the method in that a trained artificial neural network is provided, an input layer of which comprises the properties of the object when it enters the area and an output layer of which comprises the desirability or the undesirability of the object, and the desirability or the undesirability of the object is determined from the measured properties.
[0015] In an embodiment of the method the future positions of an entered undesired object are calculated. It is thus possible in an embodiment of the invention for the image recognition software to be configured to determine other properties, such as the speed at which the object moves and the direction in which the object moves at the position of the entry boundary.
[0016] In this way it becomes possible to calculate the future positions of an entered undesired object from the measured positions and velocity vector upon entry. Such a prediction can for instance be carried out by a deterministic or statistical model, wherein current patterns in the water mass in the area are taken into consideration.
[0017] In an embodiment of the invention it is also possible to characterize the method in that a trained artificial neural network is provided, an input layer of which comprises the position and a velocity vector of the object upon entry into the area and an output layer comprises future positions and velocity vectors of the object in the water mass, and that the future positions of an entered undesired object are calculated from the measured position and velocity vector upon entry.
[0018] The area according to the invention can comprise any area of a water mass and in a preferred embodiment comprises a drainage basin of a river system, which term drainage basin will be known to the skilled person. It is however also possible for the area to comprise a part of a river system, more preferably a river, still more preferably a section of a river.
[0019] The invention preferably makes use of an artificial neural network. This network is able to distil patterns from large amounts of data. In the context of the present invention the neural network is in particular able to determine future positions of an undesired object floating in the water mass as a result of the current position and velocity vector of the undesired object when it is situated at the position of the entry boundary of the area.
[0020] The architecture of a neural network comprises in known manner an input layer of input neurons, at least one hidden layer consisting of one or more layers of neurons, and an output layer of output neurons. According to the invention, the input layer of the artificial neural network comprises the value of the input variables and the output layer comprises the predicted future positions of an object which has entered the area. In an embodiment a current position can likewise be an input variable. A neural network can comprise tens of layers which can each comprise a large number of neurons, for instance from several tens to hundreds or even thousands. A complex neural network with more than two intermediate layers is also referred to as a deep neural network.
[0021] The input variables in the form of positions and/or velocity vectors determined at the entry boundary are conveyed to a hidden layer of 'neurons' with a weighting factor, the value of which is obtained by training of the network. The data of the hidden layer of neurons are then conveyed to other hidden layers, always taking into consideration a weighting factor, and finally to the output neurons, once again with a weighting factor. The weighting factors represent the relative importance attributed to a determined item of data (neuron). The neural network has been pre-trained in order to achieve a relatively realistic output. This procedure is known by the name machine learning and comprises among other things of determining the values of the weighting factors per neuron (node) of the neural network.
[0022] In an embodiment of the invention a method is provided which also comprises the step of determining the future positions of an undesired object in the area (in addition to the positions already predicted by the method). Determining these future positions enables the future positions predicted by the artificial neural network to be compared to the measured positions and the parameters of the network to be modified accordingly. Another embodiment of the method according to the invention therefore has the feature that the neural network is self-learning and the value of weighting factors of the neurons is adjusted to measurements of the positions of objects, preferably the measurement of the position of an undesired object when it is intercepted by a vessel, as will be further elucidated hereinbelow. Predictions of future positions are hereby improved. A network according to the present embodiment is also referred to with the term recurrent neural network.
[0023] If desired, the neural network can comprise a convolutional neural network. In such a neural network filters are applied, whereby not all neurons are connected to each other.
[0024] Carrying detected undesired objects in floating state to a temporary storage can take place in different ways. In an embodiment of the method it is thus possible to carry the undesired object in floating state to the temporary storage with a vessel. The vessel here differs from the temporary storage, or from a rig forming the temporary storage.
[0025] It is advantageous here for the vessel to navigate autonomously or to be remotely navigated, preferably using virtual reality techniques, which are per se known.
[0026] In a further improved embodiment a method is provided wherein the vessel navigates or is navigated such that the vessel reaches the future position of an entered undesired object. The future position of the entered object can for instance be known from the prediction of the above described trained neural network, or from a prediction of the future position obtained in other manner. It can be advantageous here to characterize the method according to the invention in that the predicted future position of an undesired object is visualized digitally for a navigator of the vessel. If desired, other quantities, such as the entry position of an object or the current in the water mass, can also be digitally visualized.
[0027] In order to further support the favourable environmental effects of the method the vessel is preferably electrically driven. In order to increase the searchability of an undesired object the vessel is in an embodiment provided with a camera which is preferably incorporated in the digital network.
[0028] The vessel is configured to intercept an undesired object which has entered the area. According to an embodiment of the invention, the method is for this purpose characterized in that the vessel comprises means for pushing along a floating object, for instance a pusher blade, more preferably a pusher blade in V-shape or U-shape.
[0029] In another embodiment of the method the carrying of the undesired object in floating state to the temporary storage is carried out by having a current in the water mass entrain the undesired object into a trap which debouches in the temporary storage.
[0030] In an embodiment in which the water mass has a width, for instance the width of a river, the trap opening preferably covers only a part of the width of the water mass. Shipping traffic can hereby continue unhindered in the width part of the water mass not covered by the trap opening. The above described vessel which is configured to carry a floating undesired object to the temporary storage can in this embodiment not only be operational only in the width part of the water mass not covered by the trap opening, but furthermore need only carry an object which has entered the area along this width part to a position in which the object is entrained with the current of the water mass into the trap opening which debouches in the temporary storage.
[0031] The trap according to an embodiment arranged in the water mass can take any form. A particularly suitable trap is a floating barrier comprising for instance two screens, together forming a trap, of a material which retains undesired objects, for instance of layered, fibre-reinforced rubber which is used for conveyor belts. In an embodiment the two screens arranged as a trap taper toward the temporary storage. In order to orient the screens substantially vertically (upright) they can be provided with buoyant elements at the top and weights at the bottom. The height of the screens (the depth over which they extend in the water mass) can vary depending on the objects to be caught. A preferred height lies between 10 and 150 cm, more preferably between 20 and 100 cm, still more preferably between 30 and 80 cm, depending among other things on the depth of the water mass.
[0032] In an embodiment the temporary storage to which the trap is connected comprises a floating collecting container with water-permeable side walls. The side walls can for instance be embodied in gauze-like material.
[0033] In another embodiment of the method the temporary storage comprises a floating collecting container with a front wall which can be opened and/or closed by a current in the water mass. Such a front wall can for this purpose for instance be provided on sides with oar-like elements which hang in the water and which can transmit a torque to the front wall under the influence of the current.
[0034] Another embodiment according to the invention provides a method wherein the temporary storage comprises a grabbing crane with which undesired objects arranged in the temporary storage can be taken up and removed. An alternative embodiment provides a method wherein the temporary storage comprises a screen belt whereby undesired objects arranged in the temporary storage can be removed.
[0035] The removal of the undesired objects arranged in the storage can take place in any manner. In an embodiment of the method it is thus possible to remove the undesired objects arranged in the temporary storage by removing them to a removal vessel provided in the vicinity of the temporary storage.
[0036] An efficient embodiment is formed by a method wherein the removal vessel comprises the (same) vessel with which an undesired object is carried to the temporary storage.
[0037] The invention likewise relates to a system for removing floating undesired objects from a water mass in an area with a device suitable for this purpose. The system preferably also comprises a central computer which is connected via a digital network to the device. According to the invention, the device comprises:
- a) means for determining an entry of an object into the area;
- b) image recognition means for determining properties of the object which show the undesirability of the object;
- c) means for carrying an undesired object in floating state to a temporary storage; and
- d) means for removing undesired objects arranged in the temporary storage.
[0038] According to the invention, the computer is configured to perform the method and is for this purpose loaded with a computer program comprising program instructions for having the method be performed. The advantages of such a system have already been elucidated on the basis of the above discussed method, and will not be repeated here. The system according to the invention can collect the necessary input variables in the form of positions and velocity factors of objects entering the area and calculate the future positions and/or velocity vectors of these objects as a result of the change of input variables. In an embodiment of the device, the device comprises the properties of an entered object to properties which are typical for an undesired object. If (a majority of) these properties match, the object is catalogued as undesired. If (a majority of) these properties do not match, the object is catalogued as desired, and allowed to pass. In another embodiment the model output can be made visible on a digital screen, on the basis of which a possible operator of the vessel will optionally adjust the route of the vessel remotely.
[0039] Finally, it is stated that the embodiments of the invention described in this patent application can be combined in any possible combination of these embodiments.
BRIEF DESCRIPTION OF THE FIGURES
[0040] The invention will now be further elucidated on the basis of the exemplary embodiments shown in the following figures, without otherwise being limited thereto. Herein:
Figure 1 is a schematic top view of a section of a river and a system for removing floating undesired objects from the river section according to an embodiment of the invention;
Figure 2A is a schematic top view of a detail of the river section shown in figure 1 and a step in the method according to an embodiment of the invention;
Figure 2B is a schematic top view of a detail of the river section shown in figure 1 and another step in the method according to an embodiment of the invention;
Figure 2C is a schematic top view of a detail of the river section shown in figure 1 and yet another step in the method according to an embodiment of the invention;
Figure 3A is a schematic top view of a detail of the river section shown in figure 1 and a further step in the method according to an embodiment of the invention;
Figure 3B is a schematic top view of a detail of the river section shown in figure 1 and a further step in the method according to an embodiment of the invention;
Figure 4A is a schematic top view of a temporary storage according to an embodiment of the invention;
Figure 4B is a schematic perspective top view from a rear side of a temporary storage according to an embodiment of the invention;
Figure 4C is a schematic perspective top view from a front side of a temporary storage according to an embodiment of the invention;
Figure 4D is a schematic perspective bottom view from a front side of a temporary storage according to an embodiment of the invention;
Figure 4E is a schematic side view from a temporary storage according to an embodiment of the invention;
Figure 5 is a schematic top view of an undesired object as seen through a camera with image recognition according to an embodiment of the invention, arranged at the position of an entry boundary of the area; and, finally
Figure 6 is a schematic perspective rear view of an autonomously navigated vessel for carrying undesired objects in floating state to the temporary storage according to an embodiment of the invention.
DESCRIPTION OF EXEMPLARY EMBODIMENTS
[0041] Referring to figure 1, an area 7 which comprises a river section 8 bounded by two riversides (8a, 8b) is shown schematically. The river is subject to tidal current, but the average flow direction 9 is from the left to the right. Situated on an upstream side of the area 7 is a bridge 70 over the river 8. Bridge 70 defines and lies above an entry boundary 71 for area 7. Entry boundary 71 is in this case determined by the direction 9 of the tidal current, wherein tidal current 9 runs such that area 7 is located downstream of entry boundary 71.
[0042] According to the invention, a temporary storage 2 is firstly provided in river 8. A possible embodiment of this storage 2 will be further elucidated below and is configured to collect undesired objects 80 floating in river 8. Temporary storage 2 is provided for this purpose with a floating trap 1 which extends to a position under bridge 70, or more specifically to entry boundary 71, while fanning out from temporary storage 2. In order not to impede the shipping traffic through river 8 too much, an entry portion 10 of trap 1 extends over a part 81 of the width 82 of river 8. Every object which enters area 7 by crossing entry boundary 71 in the portion 81 covered by trap 1 is collected by trap 1 and carried to temporary storage 2. Trap 1 can be configured such that only undesired objects 80 are caught therein and that desired objects are allowed to pass through, as will be further described below.
[0043] Every object 80 which enters area 7 by crossing entry boundary 71 into the portion 83 left clear by trap 1 cannot be automatically collected by trap 1. According to the invented method, in this portion it will firstly be determined if and when objects 80 enter area 7, as well as the positions where and the speed at which this happens. The speed and direction of an object 80 can for instance be represented by means of a velocity vector 801 (see figure 5). An object 80 entering area 7 can be detected in practical manner by providing a means suitable for this purpose at the position of entry boundary 71, for instance in the form of one or more digital cameras 5. Camera 5 makes it possible to determine if, when and at which position along entry boundary 71 an object 80 enters area 7. A digital camera 5 can preferably also operate at night, or can operate with additional lighting (not shown).
[0044] According to the invention, it is then determined whether an object 80 detected by a camera 5 is a desired object or an undesired object. A desired object can remain floating or hovering in river 8 because such an object cannot cause damage or otherwise harm the environment. According to the invention, an undesired object 80 is removed from river section 8 because it could possibly cause damage or otherwise harm the environment further downstream.
[0045] In order to determine whether a detected object 80 is desired or undesired, the method provides a step comprising of determining properties of the object 80 with image recognition, which properties show the undesirability of object 80. Camera 5 is (or a plurality of cameras 5 are) for this purpose connected, optionally via the digital network, to a computer (not shown) comprising image recognition software. With the image recognition software it is possible to determine properties of a passing object 80, particularly the size and/or the shape of object 80. Comparing the measured properties to limit values for these properties predetermined for an undesired object makes it possible to determine whether an object 80 is desired or undesired. The comparison can for instance take place by means of a prediction of the trained neural network, or in other manner. An input layer of such a trained artificial neural network comprises the properties of an object 80, determined by a camera 5, when an object 80 enters area 7, and an output layer determines whether object 80 is desired or undesired, wherein the desirability or the undesirability of the object is determined from the measured properties.
[0046] Every object 80 labelled undesired is in principle carried in floating state to temporary storage 2. This takes place in an embodiment by means of a vessel 6 which preferably navigates autonomously or is remotely navigated, preferably using virtual-reality techniques. Vessel 6 can here navigate or be navigated such that vessel 6 reaches the future positions 802 of an entered undesired object 80. The future positions 802 of the entered object 80 can for instance be determined from a prediction of a trained neural network, or in other manner. An input layer of such a trained neural network comprises the positions and velocity vectors 81 of an object 80 when it enters area 7, while an output layer comprises future (downstream) positions 802 and velocity vectors of object 80 in river 8. The future positions 802 of an entered undesired object 80 are then calculated (predicted) from the measured positions and velocity vectors 801 upon entry into area 7. The predicted future positions 802 of an undesired object 80 can if desired be digitally visualized for an operations manager, or for a virtual navigator of vessel 7.
[0047] As can be seen in figure 6, a suitable vessel 6 is configured to intercept an undesired object 80 which has entered area 7. Vessel 6 can for this purpose comprise on a front side pushing means for a floating object 80, for instance in the form of a U-shaped pusher blade 60 which extends partially under water so that it is also able to move along hovering objects 80. Vessel 6 is preferably electrically driven and can be provided with a camera (not shown) which is preferably incorporated in the digital network. This increases the searchability of an undesired object 80.
[0048] Vessel 6 is preferably deployed as shown in figures 2A and 2B. Once an undesired object 80 has been detected by camera 5, this object 80 is intercepted by vessel 6 and carried with pusher blade 60 toward the entry portion 10 of trap 1 as according to the arrow 61, where it is entrained into trap 1 by current 9 in the direction of temporary storage 2.
[0049] In a final step of the method, the undesired objects 80 arranged in temporary storage 2 are removed from the temporary storage. Referring to figure 3C, this can for instance be done by means of a removal vessel 6a which is positioned adjacently of temporary storage 2, is filled from this position with the undesired objects 80, and then navigates along route 62 to a disposal storage onshore. The undesired objects 80 are dropped off there for further transport and the removal vessel then returns to its position in the vicinity of temporary storage 2 along route 63. Vessel 6 is preferably used as removal vessel 6a. Removal of objects 80 arranged in storage 2 can for instance take place at low tide or, in general, during periods when not many or no objects 80 are expected in the area. As shown in figures 3A and 3B, the transfer can for instance take place by placing a container 64, placed on vessel 6a and filled with objects 80, onshore, and taking up an empty container 64a, already provided onshore, as according to arrow 65 and taking it to storage 2 along route 63.
[0050] Finally, referring to figures 4A-4E, an embodiment is shown of a temporary storage 2. This comprises two floating docks 20, provided if desired with a fence 21. The floating docks 20 have a modular construction of blocks and are mutually connected with walkways 22 running transversely of docks 20. The whole is connected to the riverbed by means of one or more steel anchor piles 23. An electrically driven and autonomously navigable vessel 6 (6a) is located on a side of storage 2. This vessel 6 (6a) is provided with a container 64 in which objects 80 arranged in storage 2 can be placed. This can be done with a remotely controllable grabbing crane 25 provided on the pontoon of storage 2. Arranged between the two docks 20 is a collecting container 27 which extends partially in the water. Collecting container 27 comprises water-permeable side walls, for instance embodied in gauze-like material. Collecting container 27 also comprises a front wall 28 which can be opened and/or closed by current 9 in river 8. A rear wall can also be embodied in this way. Front wall 28 can for this purpose be provided on sides with oar-like elements which hang in the water and which can transmit a torque to front wall 28 under the influence of current 9.
[0051] A trap 1 is further connected to front wall 28. Trap 1 comprises two screens (11, 12) floating in the water and for instance constructed from a number of floating bodies 13 from which rubber net is suspended. Floating bodies 13 ensure that the screens (11, 12) remain oriented substantially vertically (upright) in use, and do not sink. The two screens (11, 12) positioned as a trap taper from a relatively wide entry side 10 to temporary storage 2. The height of the screens (the depth over which they extend in the water mass) can vary depending on the objects to be caught, and in the shown exemplary embodiment amounts to about 60 cm. The screens (11, 12) are per se known, for instance under the name floating litter boom.
1. Method for removing floating undesired objects from a water mass in an area, comprising
the following steps of:
a) determining an entry of an object into the area;
b) determining properties of the object, which show the undesirability of the object, with image recognition;
c) carrying an undesired object in floating state to a temporary storage; and
removing undesired objects arranged in the temporary storage, wherein carrying the undesired object in floating state to the temporary storage is done with a vessel which navigates autonomously or is remotely navigated, this such that the vessel reaches the future position of an entered undesired object.2. Method according to claim 1, wherein the temporary storage floats.
3. Method according to any one of the foregoing claims, wherein a trained artificial
neural network is provided, an input layer of which comprises the properties of the
object when it enters the area and an output layer of which comprises the desirability
or the undesirability of the object, and the desirability or undesirability of the
object is determined from the measured properties.
4. Method according to any one of the foregoing claims, wherein the future positions
of an entered undesired object are calculated.
5. Method according to claim 4, wherein a trained artificial neural network is provided,
an input layer of which comprises the position and a velocity vector of the object
upon entry into the area and an output layer comprises future positions and velocity
vectors of the object in the water mass, and wherein the future positions of an entered
undesired object are calculated from the measured position and velocity vector upon
entry.
6. Method according to any one of the foregoing claims, wherein the vessel comprises
means for pushing along a floating object.
7. Method according to any one of the foregoing claims, wherein carrying the undesired
object in floating state to the temporary storage is carried out by having a current
in the water mass entrain the undesired object into a trap which debouches in the
temporary storage.
8. Method according to any one of the foregoing claims, wherein the removal of undesired
objects arranged in the temporary storage comprises of removal to a removal vessel
provided in the vicinity of the temporary storage.
9. Computer program comprising program instructions for having a computer execute a method
according to any one of the foregoing claims.
10. Computer program according to claim 9, characterized in that the computer program is arranged on a physical carrier.
11. Computer program according to claim 10, characterized in that the computer program is stored at least partially in a computer memory.
12. Computer configured to execute a computer program according to any one of the claims
9-11.
13. System for removing floating undesired objects from a water mass in an area with a
device suitable for this purpose, wherein the device comprises:
a) means for determining an entry of an object into the area;
b) image recognition means for determining properties of the object which show the undesirability of the object;
c) means for carrying an undesired object in floating state to a temporary storage; and
d) means for removing undesired objects arranged in the temporary storage,
wherein the means for carrying the undesired object in floating state to the temporary storage comprise a vessel which navigates autonomously or is remotely navigated, this such that the vessel reaches the future position of an entered undesired object.14. System according to claim 13, wherein the temporary storage floats.
15. System according to any one of the claims 13-14, further comprising means configured
to calculate the future positions of an entered undesired object.
16. System according to any one of the claims 13-14, further comprising a central computer
which is connected via a digital network to the device.