BR0101172B1 - Automatic spray control system through scanned image analysis. - Google Patents

Description

"AUTOMATIC SPRAY CONTROL SYSTEM THROUGH SCANNED IMAGE ANALYSIS".

This report refers to a patent for automatic spray control system through scanned image analysis and more specifically to an automatic spray control system to be adapted to agricultural implements so that through image analysis can be automatically controlled for the steps that make up a spray.

As is well known to those skilled in the art, agriculture in general has undergone major technological transformations in the application of chemicals in the field in order to reduce application costs and make the most of the efficiency of the chemicals used. as well as increasing productivity and increasingly extracting top quality products from the ground.

Conventionally, crop application systems do not operate without the help or interference of the machine operator, who is responsible for opening and closing the valves, pumps and / or injectors that make up the system. Thus, all operation of chemical application on crops is carried out by the agricultural machine operator, who visually controls and manipulates the controls of the equipment, all of this operation being performed at a very low speed.

There are currently two ways of applying chemicals to the crop, which depends on the type of product to be used for spraying. One of these types of application throughout a crop is the application without any form of interruption. In this case the application of the chemical is made over the whole cultivation area without distinguishing where is the planted crop line, where is the pest in the soil and even where there is soil without any leaf cover.

This form of chemical application in the plantation area, although fast, presents as inconvenient the fact that it wastes a lot of chemical because the chemical, for example a pesticide, is applied where it is not necessary; In addition to causing damage to the crop, too, an insufficient amount of pesticide can be applied inappropriately to a weed area, which is harmless to the weed and consequently harmful to the crop. It is noteworthy that the damage and disastrous effects are more pronounced when considering the selective type pesticides, which are much more expensive than the traditional ones, adding to this waste an increase in application costs.

There is another conventional way of applying pesticide to the crop, such as phytosanitary application, where plants are applied pesticides only where necessary. Thus for the region of plants a fungicide or foliar fertilizer is applied thereon; However, for a region with weeds herbicides are applied, usually in the region between plants. With this other application, and with row-grown plants, the space between adjacent rows should be kept free of weeds in order to optimize crop yield and minimize problems with harvesting.

In fact, traditionally, this space between rows has been kept clean by applying herbicide for weed eradication on an uninterrupted basis, regardless of whether or not weed can contaminate the environment with excess produce and raise production costs.

The major drawback of this type of pesticide application is basically the difficulty of accurately identifying different types of plants and pests or even knowing or being aware of the differences between soil and plant, thus falling into the same problems of inadequate spraying. above.

It is an object of the present invention to provide an automatic spray control system through scanned image analysis that identifies and classifies the types of plants and soil whose images were captured and automatically directs to specific points and identifies the type of fertilizer and / or herbicide and / or crop protection product to be released by the spray nozzles of the equipment. In the following the present invention will be described with reference to

the accompanying drawings in which:

Figure 1 is a schematic plan view of an agricultural implement acting on any crop; and

Figure 2 is a schematic block diagram of the automatic spray control system in question.

According to these illustrations, the automatic spray control system through scanned image analysis object of the present invention basically consists of an S sensor responsible for capturing the digital images of the soil and / or the plant images in the latter; This sensor S is also responsible for sending these captured images to a processing center 3, where software that uses neural learning networks or equivalent systems that, through repetitive training with a data set, learns to recognize the image classes (plants ) directly by form and / or physical properties. With this automatic control system it is possible to recognize one particular plant from another, thus allowing to eliminate or give a different treatment or even to recognize the existence of plants or not of the soil, where the spray implement is in operation. These digital images can be captured by normal or infrared lenses.

It should be borne in mind that the figures presented are simplified schematics, where details have been omitted without prejudice to the understanding of the scope of the present invention.

As can be seen from the figures, the automatic spray control system through scanned image analysis is comprised of the installation of a support bar 1 arranged at a fixed height in relation to the ground or target plant and in a position in front of the spray bar. spraying 2 of implement 10, said support bar 1 serving as a support for the attachment of sensors S1, S2, S3, ...., Sn along the path of implement 10.

Preferably, sensors S1, S2, S3, ......., Sn are mounted on the support bar 1 interchangeably when the spray agricultural implement 10 acts on row-planted crops C, so that a certain sensor S1 acts on crop line C, while sensor S2 subsequently or adjacent mounted on support bar 1 acts in the gap between rows F of plants and so on.

Based on the physical mounting of the sensors S1, S2, S3, ..., Sn next to the agricultural implement 10, and as shown schematically in the diagram of figure 2, each of the sensors S1, S2, S3, ..., Sn sends simultaneously and in real time the images captured in the area of culture C and digitized to an electronic processor 3, which is provided with software based on neural learning networks or equivalent systems, which makes comparisons with pre-established plant and soil patterns. , being able to accurately recognize along the scanning path of the sensor S1, S2, S3, .... Sn whether there are plants or not and what are their different types.

The electronic processor 3 analyzes the images and sends command signals to each set of electronic valves, pumps and / or injectors V1, V2, V3, ...., Vn installed along the spray bar 2 and covering the range. action of each sensor S1, S2, S3, ...., Sn.

Since there is a distance between the sensor clamping bar 1 and the spray bar 2, the system automatically compensates for this distance so that spraying is started or terminated at the right time, with the signal sending time to start spraying. a function of the speed and / or distance traveled by the implement. Thus, processor 3 receives signal from velocity Sv or displacement sensors Sd, preferably installed in implement 10, which are processed and used for processor 3 to trigger the control signal at the exact moment.

The system also foresees the installation of Si inclination sensors, preferably installed at the ends of the busbar 1, which constantly send signals to the processor 3 allowing it to know, at any moment, the inclination of the busbar 1. This type of information is important. so that the processor 3 can correct the images according to possible distortions and / or increase / decrease of the image, caused by the variation of the distance between the sensors and the plant or the soil, in the point where the image was generated, being such variation of distance measured at the time when bar 1 was inclined. In this way, processor 3 can compare the images generated across all sensors with the selected pattern in

equal conditions.

The system is provided with a digital display 4, where the implement operator 10 can select plant images and also the type of treatment that will be performed on each one of them.

Thus, once the images have been processed and compared with the images selected by the operator, the electronic processor 3 sends a signal at a given time, which is previously calculated as a function of signals received from the Sv speed sensors and / or Sd1 distance commander. the opening and closing of electrically controlled valves, pumps or injectors V1, V2, V3, ...., Vn1 along the spray bar 2, and may also act on valves, pumps or injectors of different defensive lines, for the treatment of different plants.

With this constructive and functional design it is possible to selectively dispose of or treat individual plants, considerably reducing the consumption of sprayed product by reaching the target plant precisely and at the proper dosage.

Still with this constructive and functional concept it is possible to install at least one sensor generating digital images in a range corresponding to and parallel to the length of the spray bar 2, so that the processor 3 decides which and where a particular product will be applied at any given moment. . This is achieved by simultaneously switching valves, pumps and / or injectors present throughout the spray bar 2.

Although a construction and a preferred operational and functional concept have been described and illustrated, it should be noted that other forms of mounting can be obtained, for example, by arranging sensors in other ways than along a horizontal bar, may be used on non-bar sprayers without departing from the scope of the present invention.

Claims (2)

1. "AUTOMATIC SPRAYING CONTROL SYSTEM THROUGH SCANNED IMAGE ANALYSIS" containing sensors (S) for capturing digital soil images (F) and / or plant images (C) therein, and sending these images to a processor (3), where software that recognizes and classifies plants by shape and / or physical properties, and directs the spray assembly for a differentiated treatment of plants or soil, where the spray implement is operating, these being sensors (S) installed on a support bar arranged at a fixed height in relation to the ground or target plant and in a position in front of the implement spray bar (10) and characterized by having inclination sensors (Si), preferably , installed at the ends of the bar (1), which constantly send signals of the inclination of the bar (1) to the processor (3), which in turn corrects the images for possible distortion. and / or increase / decrease of the image, caused by the variation of the distance between the sensors (S1, S2, ... Sn) and the plant or the soil, at the point where the image was generated. moment when the bar (1) was inclined; the distance between the fixing bar (1) of the sensors (S) and the spray bar (2) being automatically compensated by the processor (3) which starts or ends the spraying at the right time; The signal sending time to start spraying is a function of the speed and / or distance traveled by the implement; It is also provided with a control interface (4), where the operator selects plant images and also the type of treatment that will be performed on each of them.