Weeding method and device of laser weeding robot used in crop seedling drilling period
Through laser weeding robots to identify and accurately weed the problem of low efficiency and environmental pollution in existing agricultural weeding methods, achieving efficient and environmentally friendly weeding effects, and are suitable for rural areas where labor shortages are needed.
Patent Information
- Application Number
- CN202411070859.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-06
- Publication Date
- 2024-11-19
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
The existing agricultural weeding methods are inefficient and have environmental pollution and health risks, especially in rural areas where labor shortages are particularly prominent.
Laser weeding robots are used to obtain crop images in real time, and to identify crops and weeds using position characteristics, color characteristics and texture characteristics to perform precise weeding to avoid pollution and health risks caused by chemical weeding.
It improves the weeding efficiency, reduces human resource demand, avoids the negative impact of chemical weeding on the environment and health, and achieves efficient and environmentally friendly weeding effects.
Smart Images

Figure CN118975545A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of new agricultural technology, and in particular to a weeding method and a weeding device of a laser weeding robot used during row sowing of crop seedlings. Background Art
[0002] Weeds have many impacts on crop growth, such as resource competition, parasitic spread, quality pollution and management barriers, which seriously affect crop production. Therefore, weeds need to be effectively controlled to ensure high yield and quality of crops.
[0003] At present, the mainstream weed control methods in agricultural production are manual weed control and chemical weed control, but manual weed control is inefficient and requires a lot of human resources, especially in rural areas facing a shortage of labor. Although chemical weed control is more efficient, it will bring many invisible problems, such as environmental pollution, possible impact on human health, increased weed resistance, deterioration of soil texture and affecting agricultural sustainability. Summary of the invention
[0004] In order to solve the technical problems existing in the background technology, the present invention provides a weeding method and device of a laser weeding robot used during row sowing of crop seedlings.
[0005] The present invention proposes a weeding method for a laser weeding robot used during row sowing of crop seedlings, comprising:
[0006] Real-time acquisition of ground crop images during the seedling stage of row-sown crops;
[0007] Segment the ground crop image to obtain the green plant image and soil background image;
[0008] Using position features, color features and texture features to identify green plant images to obtain crops;
[0009] Subtract the green plant map from the crop map to obtain the weed map;
[0010] Weed control using lasers based on weed maps.
[0011] Preferably, the crops are identified from the green plant image using position features, color features and texture features, specifically including:
[0012] According to the pixel histogram of the green plant image, the center line and width of the crop row are extracted to identify the crop row;
[0013] Identify crops within a crop row based on color and texture features.
[0014] Preferably, before acquiring the ground crop image information in the seedling stage of row-sown crops, the method further includes:
[0015] Obtain surface characteristics and environmental information of ground crops during the seedling stage of row-sown crops;
[0016] Lighting is performed on ground crops based on surface characteristics information and environmental information of ground crops.
[0017] Preferably, the crops on the ground are illuminated according to the acquired surface property information and environmental information, specifically including:
[0018] Determine intensity adjustment parameters and angle adjustment parameters of the light source according to the acquired surface property information and environmental information;
[0019] According to the intensity adjustment parameters and angle adjustment parameters of the light source, the light source is controlled to illuminate the target product to be measured.
[0020] Preferably, weeding is performed using laser according to a weed map, specifically comprising:
[0021] Determine the amount and location of weeds based on the weed map;
[0022] The laser weeding robot is used to carbonize the weeds and their rhizomes according to the number and location of the weeds.
[0023] Preferably, after determining the quantity and location of weeds according to the weed map, the method further comprises:
[0024] Determine whether the leaves of weeds next to crops overlap with the crops;
[0025] If so, the air flow is used to separate the weed leaves from the crop.
[0026] The present invention also proposes a weeding device of a laser weeding robot used for row sowing of crop seedlings, comprising:
[0027] An acquisition module is used to acquire ground crop images in the seedling stage of row-sown crops in real time;
[0028] The segmentation module is used to segment the ground crop image to obtain the green plant image and the soil background image;
[0029] A recognition module is used to identify the green plant image to obtain crops using position features, color features and texture features;
[0030] A weed determination module is used to perform a subtraction operation on the green plant map and the crop map to obtain a weed map;
[0031] A weeding module is used to remove weeds using lasers according to a weed map.
[0032] Preferably, the crops are identified from the green plant image using position features, color features and texture features, specifically including:
[0033] According to the pixel histogram of the green plant image, the center line and width of the crop row are extracted to identify the crop row; the crops in the crop row are identified according to the color features and texture features.
[0034] Preferably, it also includes:
[0035] The lighting module is used to obtain surface property information and environmental information of ground crops in the seedling stage of row-sown crops, and to light the ground crops according to the surface property information and environmental information of the ground crops.
[0036] Preferably, weeding is performed using laser according to a weed map, specifically comprising:
[0037] Determine the amount and location of weeds based on the weed map;
[0038] The laser weeding robot is used to carbonize the weeds and their rhizomes according to the number and location of the weeds.
[0039] In the present invention, the proposed weeding method and device for a laser weeding robot during row sowing of crop seedlings first utilizes position features, color features, and texture features to identify a green plant image to obtain crops; then a subtraction operation is performed between the green plant image and the crops to obtain a weed image, which greatly reduces the difficulty of identification and the required computing power, improves the identification speed, and can use the laser weeding robot to quickly weed according to the weed image, avoiding the problems caused by chemical weeding. BRIEF DESCRIPTION OF THE DRAWINGS
[0040] Figure 1 The present invention is a flow chart of a weeding method for a laser weeding robot used during row sowing of crop seedlings. DETAILED DESCRIPTION
[0041] It should be noted that, in the absence of conflict, the embodiments of the present invention and the features in the embodiments can be combined with each other. The present invention will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.
[0042] Reference Figure 1 The present invention proposes a weeding method for a laser weeding robot used during row sowing of crop seedlings, comprising:
[0043] Real-time acquisition of ground crop images during the seedling stage of row-sown crops;
[0044] Segment the ground crop image to obtain the green plant image and soil background image;
[0045] Using position features, color features and texture features to identify green plant images to obtain crops;
[0046] Subtract the green plant map from the crop map to obtain the weed map;
[0047] Weed control using lasers based on weed maps.
[0048] In specific implementation, the present invention acquires ground crop images of row-sown crops in the seedling stage in real time, and segments the ground crop images to obtain green plant images and soil background images; then, the green plant images are identified using position features, color features, and texture features to obtain crops; then, the green plant images and crops are subtracted to obtain a weed map; and finally, laser is used to remove weeds based on the weed map.
[0049] The present invention first uses position features, color features and texture features to identify green plant images to obtain crops; then subtracts the green plant images from the crops to obtain a weed map, which greatly reduces the difficulty of identification and the required computing power, improves the identification speed, and can use a laser weeding robot to quickly weed according to the weed map, avoiding the problems caused by chemical weeding.
[0050] In this embodiment, the position feature, color feature and texture feature are used to identify the green plant image to obtain crops, specifically including:
[0051] According to the pixel histogram of the green plant image, the center line and width of the crop row are extracted to identify the crop row, so as to preliminarily determine the position of the crop to be identified;
[0052] Identify crops within crop rows based on color and texture features to quickly identify crops.
[0053] Before identification, the color features and texture features of a large number of crops of known crop species are used to train the identification model to obtain a trained recognition model, and then the trained recognition model is used to identify the crops in the crop row according to the color features and texture features to quickly identify the crops.
[0054] Since the surface trait information of crops and the information of the environment in which they are located are different, there will be different requirements for information such as light sources and lighting angles. For example, surface trait information includes surface roughness, surface color and surface reflectivity. The surface color includes a variety of colors, such as green or yellow. The surface reflectivity is divided into 1-5 levels, from mirror reflection to basically no reflection. Therefore, in order to obtain the color characteristics and texture characteristics of crops with distinct information characteristics, crops with different surface trait information must have different lighting requirements. In this embodiment, before obtaining the ground crop image information during the seedling stage of row-sown crops, it also includes:
[0055] Obtain surface characteristics and environmental information of ground crops during the seedling stage of row-sown crops;
[0056] Lighting is performed on ground crops based on surface characteristics information and environmental information of ground crops.
[0057] Among them, the crops on the ground are illuminated according to the acquired surface property information and environmental information, including:
[0058] Determine intensity adjustment parameters and angle adjustment parameters of the light source according to the acquired surface property information and environmental information;
[0059] The light source is controlled to illuminate the crops on the ground by adjusting the parameters according to the intensity and angle of the light source.
[0060] Specifically, the surface property information of the ground crops is manually input into the laser weeding robot in advance, so as to quickly determine the crops for this weeding task, thereby facilitating the rapid identification of weeds.
[0061] In this embodiment, weeding is performed using laser according to a weed map, specifically including:
[0062] Determine the amount and location of weeds based on the weed map;
[0063] The laser weeding robot is used to carbonize the weeds and their rhizomes according to the number and location of the weeds.
[0064] In a further embodiment, after determining the quantity and location of weeds according to the weed map, the method further includes:
[0065] Determine whether the leaves of weeds next to crops overlap with the crops;
[0066] If so, the airflow is used to separate the weed leaves from the crop;
[0067] The laser weeding robot is then used to carbonize the rhizomes of the weeds so that the airflow can completely separate the weeds from the crops after the roots are broken, thus preventing the secondary growth of the weed roots.
[0068] The present invention provides a weeding device of a laser weeding robot used for row sowing of crop seedlings, comprising:
[0069] An acquisition module is used to acquire ground crop images in the seedling stage of row-sown crops in real time;
[0070] The segmentation module is used to segment the ground crop image to obtain the green plant image and the soil background image;
[0071] A recognition module is used to identify the green plant image to obtain crops using position features, color features and texture features;
[0072] A weed determination module is used to perform a subtraction operation on the green plant map and the crop map to obtain a weed map;
[0073] A weeding module is used to remove weeds using lasers according to a weed map.
[0074] In this embodiment, the position feature, color feature and texture feature are used to identify the green plant image to obtain crops, specifically including:
[0075] According to the pixel histogram of the green plant image, the center line and width of the crop row are extracted to identify the crop row; the crops in the crop row are identified according to the color features and texture features.
[0076] In this embodiment, it also includes:
[0077] The lighting module is used to obtain surface property information and environmental information of ground crops in the seedling stage of row-sown crops, and to light the ground crops according to the surface property information and environmental information of the ground crops.
[0078] In this embodiment, weeding is performed using laser according to a weed map, specifically including:
[0079] Determine the amount and location of weeds based on the weed map;
[0080] The laser weeding robot is used to carbonize the weeds and their rhizomes according to the number and location of the weeds.
[0081] Among them, laser weeding is carried out according to the weed map, including:
[0082] Determine the amount and location of weeds based on the weed map;
[0083] The laser weeding robot is used to carbonize the weeds and their rhizomes according to the number and location of the weeds.
[0084] In a further embodiment, an airflow blowing module is also included;
[0085] After determining the number and location of weeds based on the weed map, the weed control module is also used to determine whether the leaves of weeds located next to the crop overlap with the crop;
[0086] If so, the airflow blowing module is used to separate the weed leaves from the crops;
[0087] Lasers are then used to carbonize the rhizomes of the weeds so that the airflow can completely separate the weeds from the crops after the roots are broken, thus preventing the secondary growth of the weed roots.
[0088] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical scheme and inventive concept of the present invention within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.
Claims
1. A weeding method using a laser weeding robot during row sowing of crop seedlings, characterized in that: include: Real-time acquisition of ground crop images during the seedling stage of row-sown crops; Segment the ground crop image to obtain the green plant image and soil background image; Using position features, color features and texture features to identify green plant images to obtain crops; Subtract the green plant map from the crop map to obtain the weed map; Weed control using lasers based on weed maps.
2. The weeding method of the laser weeding robot used during row sowing of crop seedlings according to claim 1, characterized in that: The position features, color features and texture features are used to identify the crops in the green plant image, including: According to the pixel histogram of the green plant image, the center line and width of the crop row are extracted to identify the crop row; Identify crops within a crop row based on color and texture features.
3. The weeding method of the laser weeding robot used during row sowing of crop seedlings according to claim 1, characterized in that: Before obtaining the ground crop image information in the seedling stage of row-sown crops, the following steps are also included: Obtain surface characteristics and environmental information of ground crops during the seedling stage of row-sown crops; Lighting is performed on ground crops based on surface characteristics information and environmental information of ground crops.
4. The weeding method of the laser weeding robot used during row sowing of crop seedlings according to claim 3, characterized in that: Based on the acquired surface characteristics and environmental information, the crops on the ground are illuminated, including: Determine intensity adjustment parameters and angle adjustment parameters of the light source according to the acquired surface property information and environmental information; According to the intensity adjustment parameters and angle adjustment parameters of the light source, the light source is controlled to illuminate the target product to be measured.
5. The weeding method of the laser weeding robot used during row sowing of crop seedlings according to claim 1, characterized in that: Weed control using lasers based on weed maps includes: Determine the amount and location of weeds based on the weed map; The laser weeding robot is used to carbonize the weeds and their rhizomes according to the number and location of the weeds.
6. The weeding method of the laser weeding robot used during row sowing of crop seedlings according to claim 5, characterized in that: After determining the number and location of weeds based on the weed map, also include: Determine whether the leaves of weeds next to crops overlap with the crops; If so, the air flow is used to separate the weed leaves from the crop.
7. A weeding device of a laser weeding robot used during row sowing of crop seedlings, characterized in that: include: An acquisition module is used to acquire ground crop images in the seedling stage of row-sown crops in real time; The segmentation module is used to segment the ground crop image to obtain the green plant image and the soil background image; A recognition module is used to identify the green plant image to obtain crops using position features, color features and texture features; A weed determination module is used to perform a subtraction operation on the green plant map and the crop map to obtain a weed map; A weeding module is used to remove weeds using lasers according to a weed map.
8. The weeding device of the laser weeding robot used during row sowing of crop seedlings according to claim 7, characterized in that: The position features, color features and texture features are used to identify the crops in the green plant image, including: According to the pixel histogram of the green plant image, the center line and width of the crop row are extracted to identify the crop row; the crops in the crop row are identified according to the color features and texture features.
9. The weeding device of the laser weeding robot used during row sowing of crop seedlings according to claim 1, characterized in that: Also includes: The lighting module is used to obtain surface property information and environmental information of ground crops in the seedling stage of row-sown crops, and to light the ground crops according to the surface property information and environmental information of the ground crops.
10. The weeding device of the laser weeding robot used during row sowing of crop seedlings according to claim 7, characterized in that: Weed control using lasers based on weed maps includes: Determine the amount and location of weeds based on the weed map; The laser weeding robot is used to carbonize the weeds and their rhizomes according to the number and location of the weeds.
Citation Information
Patent Citations
Method for recognizing inline crops and weeds of seedling stage in farmland
CN101859375A
Method and device for automatically extracting cotton seeds
CN102339378A
Intelligent weeding machine and weed removing method thereof
CN105230224A
Method for recognizing inline crops and weeds in seedling stage of farmland
CN106683069A
Cultivation method for interplanting forage mulberries and forage grass
CN113039986A
