Real-time identification system for weeds in crop planting area

Abstract

The invention belongs to the technical field of crop planting area weed real-time identification systems, and particularly relates to a crop planting area weed real-time identification system which comprises an image collector and a monitor. The image collector is used for collecting pictures of a crop planting area in real time; the monitor is provided with a weed graph collection and storage module, a weed graph data enhancement processing module, a weed graph recognition model construction module and a crop planting area picture recognition module; the weed graph collection and storage module is used for collecting and storing weed pictures; the weed graphic data enhancement processing module is used for carrying out data enhancement processing on the weed pictures; the weed graph recognition model construction module is used for training a YOLO v5 algorithm by using the weed pictures to obtain a weed recognition model; and the crop planting area picture recognition module is used for performing weed recognition on the crop planting area picture by using a weed recognition model.

Description

Technical Field

[0001] The present application belongs to the technical field of weed identification in crop planting areas, and specifically relates to a real-time weed identification system in crop planting areas. Background Art

[0002] Weeds often grow in crop planting areas, such as pigweed, sowwort, wild oats, veronica and prickly lettuce, which compete with crops for fertilizer, water and sunlight, seriously affecting the growth of crops. Therefore, it is necessary to remove the weeds growing in the crop planting areas in a timely manner.

[0003] To remove weeds growing in crop planting areas, you first need to identify the weeds growing in the crop planting areas, and then you can remove them in a targeted manner.

[0004] At present, weeds growing in crop planting areas are mostly identified by manual inspection. This method consumes a lot of manpower and is prone to omissions. It is difficult to ensure that weeds growing in crop planting areas can be discovered and identified in a timely manner over a large area, and the opportunity for removal is missed, which can easily seriously affect crop growth.

[0005] This application is proposed in view of the above-mentioned technical defects.

[0006] It should be noted that the disclosure of the above background technology content is only used to assist in understanding the inventive concept and technical solution of the present invention, and it does not necessarily belong to the prior art of this patent application. In the absence of clear evidence that the above content has been disclosed on the filing date of this application, the above background technology should not be used to evaluate the novelty and creativity of the present application. Summary of the invention

[0007] The purpose of the present application is to provide a real-time weed identification system for crop planting areas to overcome or alleviate at least one of the known technical deficiencies.

[0008] The technical solution of this application is:

[0009] A real-time weed recognition system for crop planting areas, comprising an image collector and a monitor;

[0010] The image collector is used to collect pictures of crop planting areas in real time;

[0011] The monitor has a weed pattern collection and storage module, a weed pattern data enhancement processing module, a weed pattern recognition model building module, and a crop planting area picture recognition module;

[0012] The weed image collection and storage module is used to collect and store weed images;

[0013] The weed image data enhancement processing module is used to perform data enhancement processing on the weed image;

[0014] The weed pattern recognition model building module is used to train the YOLO v5 algorithm using weed pictures to obtain a weed recognition model;

[0015] The crop planting area image recognition module is used to use the weed recognition model to perform weed recognition on the crop planting area image.

[0016] Optionally, in the above-mentioned real-time weed identification system for crop planting areas, the image collector is specifically a camera that can rotate freely, or a drone with a camera;

[0017] A monitor is specifically a high-performance computer with a display.

[0018] Optionally, in the above-mentioned real-time weed identification system for crop planting areas, the weed image collection and storage module collects and stores weed images, including images of Clerodendrum sphaerocephalum, Artemisia selengensis, Oats, Veronica officinalis, and Eupatorium truncatum.

[0019] Optionally, in the above-mentioned real-time weed recognition system in the crop planting area, the weed image data enhancement processing module performs data enhancement processing on the weed image, including:

[0020] Rotate the weed image, specifically 90°, 180°, and 270° counterclockwise;

[0021] Flip the weed image horizontally;

[0022] Add noise to the weed image, including Gaussian noise, salt and pepper noise, and scattered noise.

[0023] Optionally, in the above-mentioned real-time weed recognition system for crop planting areas, in the weed pattern recognition model construction module, 80% of the weed images are randomly used for training, 10% of the weed images are used for testing, and 10% of the weed images are used for verification.

[0024] Optionally, in the above-mentioned real-time weed identification system for crop planting areas, in the weed pattern recognition model construction module, the network structure of the YOLO v5 algorithm model includes an input layer, a backbone layer, a network layer, and a detection layer.

[0025] Optionally, in the above-mentioned real-time weed identification system in the crop planting area, in the weed pattern recognition model construction module, in the network structure of the YOLO v5 algorithm model, the backbone layer and the network layer extract and fuse the features of the weed image, and pass the feature image of the weed image to the detection layer for weed identification.

[0026] Optionally, in the above-mentioned real-time weed recognition system for crop planting areas, in the weed pattern recognition model construction module, in the network structure of the YOLO v5 algorithm model, an anchor box is used in the feature picture of the weed picture output by the network layer, and GIOU is used as the loss function for bounding box training, specifically:

[0027]

[0028] in,

[0029] I O U is the intersection area of ​​the predicted box and the real box;

[0030] A c is the minimum enclosed area of ​​the predicted box and the true box;

[0031] U is the area of ​​the region in the closure region that does not belong to the predicted box or the true box.

[0032] This application has at least the following beneficial technical effects:

[0033] A real-time weed recognition system for crop planting areas is provided. By collecting weed pictures, performing data enhancement processing on the weed pictures, and training the YOLO v5 algorithm, a weed recognition model is obtained. Then, the weed recognition model can be used to identify weeds in crop planting area pictures to determine whether weeds grow in the crop planting area. The weeds grown in the crop planting area can be discovered and identified in a timely manner over a large area, and the weeds grown in the crop planting area can be removed to avoid serious impact on crop growth. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Figure 1 is a schematic diagram of a real-time weed identification system for crop planting areas provided in an embodiment of the present application;

[0035] Figure 2 It is a schematic diagram of the network structure of the YOLO v5 algorithm model provided in an embodiment of the present application;

[0036] Figure 3 is a schematic diagram of a weed picture provided in an embodiment of the present application;

[0037] Figure 4 Schematic diagram of a weed picture with feature extraction provided in an embodiment of the present application.

[0038] In order to better illustrate this embodiment, some parts of the drawings may be omitted, enlarged or reduced, and do not represent the size of the actual product. In addition, the drawings are only used for illustrative purposes and cannot be understood as limitations on this patent. DETAILED DESCRIPTION

[0039] In order to make the technical solution and advantages of the present application clearer, the technical solution of the present application will be further described in detail in detail and in detail in conjunction with the accompanying drawings. It can be understood that the specific embodiments described here are only partial embodiments of the present application, which are only used to explain the present application, not to limit the present application. It should be noted that, for the convenience of description, only the parts related to the present application are shown in the accompanying drawings, and other related parts can refer to the general design. In the absence of conflict, the embodiments in the present application and the technical features in the embodiments can be combined with each other to obtain new embodiments.

[0040] In addition, unless otherwise defined, the technical terms or scientific terms used in the description of this application shall be the common meanings understood by the general technicians in the field to which this application belongs. The words "upper", "lower", "left", "right", "center", "vertical", "horizontal", "inside", "outside" and other words indicating orientation used in the description of this application are only used to indicate the relative direction or positional relationship, and do not imply that the device or element must have a specific orientation, be constructed and operated in a specific orientation. When the absolute position of the described object changes, its relative positional relationship may also change accordingly, so it cannot be understood as a limitation on this application. The "first", "second", "third" and similar terms used in the description of this application are only used for descriptive purposes to distinguish different components, and cannot be understood as indicating or implying relative importance. The similar words "one", "one" or "the" used in the description of this application should not be understood as an absolute limitation on quantity, but should be understood as the existence of at least one. The similar words "including" or "comprising" used in the description of this application mean that the elements or objects appearing in front of the word include the elements or objects listed after the word and their equivalents, without excluding other elements or objects.

[0041] In addition, it should be noted that, unless otherwise clearly specified and limited, the words "installed", "connected", "connected" and similar terms used in the description of this application should be understood in a broad sense. For example, the connection can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, an indirect connection through an intermediate medium, or the internal connection of two components. Technical personnel in the field can understand their specific meanings in this application according to the specific circumstances.

[0042] A real-time weed identification system for crop planting areas, such as Figure 1 As shown, it includes an image collector and a monitor.

[0043] The image collector and the monitor are wirelessly connected, and specifically can be connected via a signal antenna to enable signal transmission.

[0044] Image collectors and monitors can be powered by solar panels.

[0045] The image collector can specifically be a freely rotating camera, or a drone with a camera, for collecting pictures of the crop planting area in real time.

[0046] The monitor can be a high-performance computer with a display, which has a control module, a weed pattern collection and storage module, a weed pattern data enhancement processing module, a weed pattern recognition model building module, and a crop planting area image recognition module.

[0047] The control module can control the image collector. For example, when the image collector is a freely rotating camera, the control module can control the time, frequency, and angle of the camera's rotation. When the image collector is a drone with a camera, the control module can control the time, frequency, and flight trajectory of the drone.

[0048] The weed image collection and storage module is used to collect and store weed images, including images of Glechoma longituba, Artemisia selengensis, Oats ovata, Veronica officinalis, and Eupatorium truncatum, which can be collected from the Internet or by photographing.

[0049] The weed image data enhancement processing module is used to perform data enhancement processing on the weed image.

[0050] The weed image can be rotated, for example, 90°, 180°, or 270° counterclockwise.

[0051] Flip the weed image horizontally.

[0052] Add noise to the weed image, for example, you can add an appropriate amount of Gaussian noise, salt and pepper noise, or scattered noise.

[0053] The image annotation tool Labelmg can be used to manually annotate the path, width and height, number of channels, and location information of crop and weed annotation boxes of weed images and store them in PASCAL VOC format.

[0054] The weed pattern recognition model building module is used to train the YOLO v5 algorithm using weed pictures to obtain a weed recognition model.

[0055] Data augmentation processing of weed images can increase the data for YOLO v5 algorithm training and promote the efficient and accurate weed recognition model.

[0056] Randomly use 80% of the weed pictures for training, 10% of the weed pictures for testing, and 10% of the weed pictures for validation.

[0057] The core idea of ​​the YOLO v5 algorithm is to divide the image into multiple grids and use a separate neural network on each grid to predict the bounding box and category of the target object through reverse propagation. The network structure of the YOLO v5 algorithm model includes the input layer (Input), the backbone layer (Backbone), the network layer (Neck), and the detection layer (Detect). Figure 2 shown.

[0058] The backbone layer and network layer extract and fuse the features of the weed images, and pass the feature images of the weed images to the detection layer for weed recognition.

[0059] The anchor box is used in the feature picture of the weed picture output by the network layer, and GIOU is used as the loss function for bounding box training:

[0060]

[0061] in,

[0062] I O U is the intersection area of ​​the predicted box and the real box;

[0063] A c is the minimum enclosed area of ​​the predicted box and the true box;

[0064] U is the area of ​​the region in the closure region that does not belong to the predicted box or the true box.

[0065] In a specific example, a weed image such as Figure 3 As shown, the feature-extracted weed pictures are as follows Figure 4 shown.

[0066] After the weed recognition model is constructed, the weed pattern collection and storage module, the weed pattern data enhancement processing module, and the weed pattern recognition model construction module can be stopped.

[0067] The crop planting area image recognition module is used to use the weed recognition model to perform weed recognition on the crop planting area image.

[0068] The weed recognition model can be used to identify the presence of weeds in images of crop planting areas, and then determine the type and quantity of weeds.

[0069] The real-time weed recognition system for crop planting areas disclosed in the above embodiments collects weed images, performs data enhancement processing on the weed images, and then trains the YOLO v5 algorithm to obtain a weed recognition model. The weed recognition model can then be used to identify weeds in crop planting area images collected in real time to determine whether weeds have grown in the crop planting area. The system can timely detect and identify weeds grown in crop planting areas over a large area, remove the weeds grown in the crop planting areas, and avoid serious impact on crop growth.

[0070] In addition, those skilled in the art should also be able to realize that the various modules of the real-time weed identification system for crop planting areas disclosed in the embodiments of the present application can be implemented by electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the present application generally describes them according to their functions. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can choose to use different methods to implement the described functions for each specific application and its actual constraints, but such implementation should not be considered to be beyond the scope of the present application.

[0071] So far, the technical solution of the present application has been described in conjunction with the preferred embodiments shown in the accompanying drawings. Those skilled in the art should understand that the scope of protection of the present application is obviously not limited to these specific embodiments. Without departing from the principles of the present application, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will fall within the scope of protection of the present application.

Claims

1. A real-time weed identification system for crop planting areas, characterized in that: Including image collector and monitor; The image collector is used to collect pictures of crop planting areas in real time; The monitor has a weed pattern collection and storage module, a weed pattern data enhancement processing module, a weed pattern recognition model building module, and a crop planting area picture recognition module; The weed image collection and storage module is used to collect and store weed images; The weed image data enhancement processing module is used to perform data enhancement processing on the weed image; The weed pattern recognition model building module is used to train the YOLO v5 algorithm using weed pictures to obtain a weed recognition model; The crop planting area image recognition module is used to use the weed recognition model to perform weed recognition on the crop planting area image.

2. The real-time weed identification system for crop planting areas according to claim 1, characterized in that: The image collector is specifically a camera that can rotate freely, or a drone with a camera; A monitor is specifically a high-performance computer with a display.

3. The real-time weed identification system for crop planting areas according to claim 2, characterized in that: In the weed image collection and storage module, weed images are collected and stored, including images of pigweed, sophora japonica, wild oats, veronica and prickly lettuce.

4. The real-time weed identification system for crop planting areas according to claim 3, characterized in that: In the weed image data enhancement processing module, data enhancement processing is performed on the weed image, including: Rotate the weed image, specifically 90°, 180°, and 270° counterclockwise; Flip the weed image horizontally; Add noise to the weed image, including Gaussian noise, salt and pepper noise, and scattered noise.

5. The real-time weed identification system for crop planting areas according to claim 4, characterized in that: In the weed pattern recognition model construction module, 80% of the weed images are randomly used for training, 10% of the weed images are used for testing, and 10% of the weed images are used for verification.

6. The real-time weed identification system for crop planting areas according to claim 5, characterized in that: In the weed pattern recognition model building module, the network structure of the YOLO v5 algorithm model includes input layer, backbone layer, network layer, and detection layer.

7. The real-time weed identification system for crop planting areas according to claim 6, characterized in that: In the weed pattern recognition model building module, in the network structure of the YOLO v5 algorithm model, the backbone layer and the network layer extract and fuse the features of the weed image, and pass the feature images of the weed image to the detection layer for weed recognition.

8. The real-time weed identification system for crop planting areas according to claim 7, characterized in that: In the weed pattern recognition model building module, in the network structure of the YOLO v5 algorithm model, the anchor box is used in the feature picture of the weed picture output by the network layer, and GIOU is used as the loss function for bounding box training, specifically: in, I O U is the intersection area of ​​the predicted box and the real box; A c is the minimum enclosed area of ​​the predicted box and the true box; U is the area of ​​the region in the closure region that does not belong to the predicted box or the true box.

Images