Safety protection method of automatic mower

Abstract

The invention relates to a safety protection method of an automatic mower, belongs to the technical field of automatic mowers, and solves the problem of unsafety in mowing of the automatic mower in the prior art. The safety protection method comprises the following steps: S1, collecting a top view to be mowed at a first distance in the advancing direction of the automatic mower and detecting whether an obstacle is included outside a second distance; the first distance is determined according to the advancing speed of the automatic mower and an image acquisition period, and the second distance is greater than the first distance; s2, judging whether a dangerous target is included in the top view to be mowed or not; if the dangerous target is included, the robot continues to advance after the dangerous target is cleared; and S3, if the obstacle is included outside the second distance, re-planning the advancing path, and continuing to advance according to the re-planned advancing path. And the safety of the automatic mower in the mowing process is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of automatic lawn mowers, and particularly to a safety protection method for an automatic lawn mower. Background Art

[0002] In order to ensure the green area of the city, various types of lawns are planted in streets, parks, and schools. Since the grass on the lawn will grow naturally, it is necessary to regularly trim and beautify the lawn. The existing lawn beautification usually uses an intelligent lawn mower to trim the lawn. The intelligent lawn mower can autonomously complete the work of trimming the lawn without direct human control and operation, and has the characteristics of low power, low noise, and delicate and beautiful appearance. The intelligent lawn mower can greatly reduce manual operation.

[0003] When the automatic lawn mower mows grass according to a preset path, various obstacles or dangerous targets in the preset path may cause damage to the automatic lawn mower, resulting in low safety.

[0004] Therefore, a technical solution for safety protection of the automatic lawn mower is needed. Summary of the Invention

[0005] In view of the above analysis, the embodiments of the present invention aim to provide a safety protection method for an automatic lawn mower to solve the problem of unsafe grass mowing of the automatic lawn mower in the prior art.

[0006] The embodiments of the present invention provide a safety protection method for an automatic lawn mower, and the safety protection method includes:

[0007] Step S1: Collect a top view of the grass to be mowed at a first distance in the forward direction of the automatic lawn mower and detect whether there are obstacles outside a second distance; the first distance is determined according to the forward speed of the automatic lawn mower and the image acquisition period, and the second distance is greater than the first distance;

[0008] Step S2: Determine whether there are dangerous targets in the top view of the grass to be mowed; if there are dangerous targets, remove the dangerous targets and then continue to move forward;

[0009] Step S3: If there are obstacles outside the second distance, re-plan the forward path and continue to move forward according to the re-planned forward path.

[0010] Based on a further improvement of the above safety protection method, the collection of the top view of the grass to be mowed at a first distance in the forward direction of the automatic lawn mower includes:

[0011] An image acquisition device is detachably arranged on the top of the automatic lawn mower. When the automatic lawn mower moves forward, the height of the image acquisition device is adjusted so that the image acquisition device can collect the grassland area at the first distance to obtain a top view of the grass to be mowed.

[0012] Based on a further improvement of the above safety protection method, the adjustment of the height of the image acquisition device includes:

[0013] Determine the real-time height of the image acquisition device according to the forward speed of the automatic lawn mower, the image acquisition period, and the downward viewing angle of the image acquisition device;

[0014] By adjusting the height of the support rod between the image acquisition device and the automatic lawn mower, make the image acquisition device reach the real-time height.

[0015] Based on a further improvement of the above safety protection method, the real-time height of the image acquisition device is calculated by the following formula:

[0016] H = (V * T) / tanθ;

[0017] Wherein, V represents the forward speed of the automatic lawn mower, T represents the acquisition period, and θ represents the downward viewing angle of the image acquisition device.

[0018] Based on a further improvement of the above safety protection method, the image acquisition device is any one of the following:

[0019] High-definition camera;

[0020] Micro scanner;

[0021] Video recorder.

[0022] Based on a further improvement of the above safety protection method, the determination of whether the to-be-mowed top view includes dangerous targets includes:

[0023] Input the collected to-be-mowed top view into a pre-trained target discrimination model to obtain the category and position coordinates of the targets included in the to-be-mowed top view; the target discrimination model is a deep neural network based on the attention mechanism;

[0024] Based on the category of the targets included in the to-be-mowed top view, determine whether they are dangerous targets.

[0025] Based on a further improvement of the above safety protection method, if the categories of the targets included in the to-be-mowed top view do not belong to dangerous targets, continue to move forward.

[0026] Based on a further improvement of the above safety protection method, if it includes dangerous targets, after clearing the dangerous targets, continue to move forward, including:

[0027] Determine the position coordinates of the dangerous targets, and control the broom of the automatic lawn mower to clear the dangerous targets to prevent the cutting blade of the automatic lawn mower from contacting the dangerous targets.

[0028] Based on a further improvement of the above safety protection method, if there are obstacles within the second distance, re-planning the forward path includes:

[0029] Determine the third distance between the obstacle and the automatic lawn mower, as well as the first deflection angle and the second deflection angle between both sides of the obstacle and the automatic lawn mower;

[0030] Determine the rotation angle of the automatic lawn mower according to the third distance, the first deflection angle and the second deflection angle, so that the automatic lawn mower rotates to reach the rotation angle;

[0031] The automatic lawn mower continues to move forward according to the rotated angle. When the automatic lawn mower crosses the obstacle, it returns to the initial path and continues to move forward.

[0032] Based on a further improvement of the above safety protection method, the target discrimination model is any one of the following networks:

[0033] Multi-view convolutional neural network MVCNN;

[0034] Convolutional neural network for grouped views GVCNN;

[0035] Multi-view long short-term memory network MV-LSTM;

[0036] Multi-view deformable neural network MVTransformer.

[0037] Compared with the prior art, the present invention can at least achieve one of the following beneficial effects:

[0038] 1. During the process of the automatic lawn mower mowing the lawn, collect the mowing images in the forward direction in real time, judge whether there are dangerous targets in the mowing images, reduce the damage to the automatic lawn mower, and improve the safety of the automatic lawn mower during the mowing process;

[0039] 2. Use a deep neural network to identify the mowing images and quickly obtain the recognition results, so that the automatic lawn mower improves the mowing safety without affecting the mowing efficiency;

[0040] 3. At the same time, judge whether there are obstacles in the forward path of the automatic lawn mower. If there are obstacles, the automatic lawn mower bypasses the obstacles to improve the safety of the automatic lawn mower during the mowing process.

[0041] In the present invention, the above technical solutions can also be combined with each other to achieve more preferred combination schemes. Other features and advantages of the present invention will be described in the subsequent specification, and some advantages can be made obvious from the specification, or understood by implementing the present invention. The objectives and other advantages of the present invention can be realized and obtained from the content specifically pointed out in the specification and the drawings. Description of the Drawings

[0042] The drawings are only for the purpose of showing specific embodiments and are not considered as a limitation to the present invention. Throughout the drawings, the same reference signs denote the same components.

[0043] Figure 1 It is a schematic flow chart of the safety protection method for the automatic lawn mower provided by the embodiment of the present invention;

[0044] Figure 2 It is a schematic structural diagram of the deflection angle of the automatic lawn mower provided by the embodiment of the present invention. Detailed Embodiments

[0045] The following will specifically describe the preferred embodiments of the present invention with reference to the drawings. The drawings form a part of this application and are used together with the embodiments of the present invention to explain the principle of the present invention, rather than to limit the scope of the present invention.

[0046] A specific embodiment of the present invention discloses a safety protection method for an automatic lawn mower, as Figure 1 shown, the safety protection method includes:

[0047] Step S1: Collect the top view of the grass to be mowed at a first distance in the forward direction of the automatic lawn mower and detect whether there are obstacles beyond a second distance; the first distance is determined according to the forward speed of the automatic lawn mower and the image acquisition period, and the second distance is greater than the first distance;

[0048] Step S2: Determine whether there are dangerous targets in the top view of the grass to be mowed; if there are dangerous targets, remove the dangerous targets and then continue to move forward;

[0049] Step S3: If there are obstacles beyond the second distance, re-plan the forward path and continue to move forward according to the re-planned forward path.

[0050] Specifically, in step S1, the top view of the grass to be mowed at a first distance in the forward direction of the automatic lawn mower is collected at regular intervals of a period of time T. The top view of the grass to be mowed is the grassland area to be mowed on the forward path of the automatic lawn mower. It can be understood that if there are dangerous targets on the top view of the grass to be mowed, such as steel pipes, stone blocks, etc., when the cutting blade of the automatic lawn mower touches the dangerous target, it will accelerate the damage of the cutting blade, making the mowing process of the automatic lawn mower unsafe and reducing the service life of the automatic lawn mower.

[0051] Specifically, the first distance changes with the forward speed of the automatic lawn mower and the image acquisition period. The faster the forward speed, the greater the first distance; the larger the image acquisition period, the greater the first distance.

[0052] By collecting the top view of the grass to be mowed and identifying the targets included in the top view of the grass to be mowed, damage to the automatic lawn mower is reduced and the use safety is improved.

[0053] Preferably, the collection of the top view of the grass to be mowed at the first distance in the advancing direction of the automatic lawn mower includes:

[0054] An image acquisition device is detachably arranged on the top of the automatic lawn mower. When the automatic lawn mower advances, the height of the image acquisition device is adjusted so that the image acquisition device can collect the grassland area at the first distance to obtain the top view of the grass to be mowed.

[0055] Specifically, by arranging a detachable image acquisition device on the top of the automatic lawn mower, different image acquisition devices can be reasonably set according to the situation of the grassland. If the environment of the grassland is good, a higher mowing speed can be set to mow the grassland; on the contrary, if the environment of the grassland is poor, a lower mowing speed needs to be set to mow the grassland to prevent the automatic lawn mower from jolting, etc., and reduce the mowing effect of the automatic lawn mower.

[0056] When mowing the grassland at different speeds, the height of the image acquisition device can be adjusted so that the image acquisition device can collect grasslands of different areas.

[0057] Preferably, the adjustment of the height of the image acquisition device includes:

[0058] Determine the real-time height of the image acquisition device according to the advancing speed of the automatic lawn mower, the image acquisition period and the downward viewing angle of the image acquisition device;

[0059] By adjusting the height of the support rod of the image acquisition device and the automatic lawn mower, the image acquisition device reaches the real-time height.

[0060] Specifically, the real-time height of the image acquisition device is calculated by the following formula:

[0061] H = (V * T) / tanθ;

[0062] Wherein, V represents the advancing speed of the automatic lawn mower, T represents the acquisition period, and θ represents the downward viewing angle of the image acquisition device.

[0063] Specifically, the speed of the automatic lawn mower is different during the mowing process, resulting in different first distances V * T. In order to collect all the regional images involved in the mowing area within this period, the height H of the image acquisition device needs to be adjusted to obtain the image of the grass to be mowed, that is, the top view of the grass to be mowed. During the mowing process, the downward viewing angle θ of the image acquisition device remains unchanged.

[0064] Specifically, the image acquisition device is any one of the following:

[0065] High-definition camera;

[0066] Microscopic scanner;

[0067] Video recorder.

[0068] Preferably, determining whether the top view of the grass to be mowed includes a dangerous target includes:

[0069] Inputting the collected top view of the grass to be mowed into a pre-trained target discrimination model to obtain the category and position coordinates of the targets included in the top view of the grass to be mowed; the target discrimination model is a deep neural network based on the attention mechanism;

[0070] Based on the category of the targets included in the top view of the grass to be mowed, determine whether it is a dangerous target.

[0071] Specifically, in step S2, input the top view of the grass to be mowed obtained by the image acquisition device into a pre-trained target discrimination model. The target discrimination model is a deep neural network based on the attention mechanism, and the target discrimination model is any one of the following networks:

[0072] Multi-view Convolutional Neural Network MVCNN;

[0073] Group-view Convolutional Neural Network GVCNN;

[0074] Multi-view Long Short-Term Memory Network MV-LSTM;

[0075] Multi-view Transformer Network MVTransformer.

[0076] Specifically, the target recognition model provided by the embodiments of the present invention may be MVCNN (Multi-view Convolutional Neural Networks), GVCNN (Group-view Convolutional Neural Networks), MV-LSTM (Long Short-Term Memory), or MVTransformer (Multi-view Transformer Networks).

[0077] A neural network is a computational model composed of a set of interconnected simple units (neurons), which can perform various tasks, such as classification, regression, clustering, etc., by learning the relationship between inputs and outputs.

[0078] A neural network consists of multiple layers of neurons, typically including an input layer, hidden layers, and an output layer. Each neuron receives inputs from the previous layer, applies a weighted sum to these inputs using an activation function, and outputs a non-linear activation value. The connection weights between neurons can be trained using the backpropagation algorithm to minimize the error between the model's predictions and the true outputs.

[0079] The attention mechanism in deep learning can mimic the human visual and cognitive systems, allowing the neural network to focus on relevant parts when processing input data. By introducing the attention mechanism, the image object detection model can automatically learn and selectively focus on important information in the input, improving the performance and generalization ability of the image object detection model and enhancing the recognition of two-dimensional projection images.

[0080] Specifically, the trained target discrimination model can quickly and accurately identify multiple targets included in the top view of the lawn to be mowed, as well as the categories of the multiple targets, and determine whether there are dangerous targets among the multiple targets. If there are dangerous targets, the dangerous targets need to be cleared. When training the target discrimination model, it includes: ① constructing a training set; ② selecting a neural network type; ③ training the neural network using the training set until a preset end condition is met.

[0081] When constructing the training set in ①, determine several categories of objects, throw the objects into the grassland areas in different environments in sequence, take pictures through an image acquisition device, obtain a large number of training images, set the category labels and position coordinates of each target for each training image, and use each training image, as well as the categories and position coordinates of the targets included in the training image, as a training sample, and determine multiple training samples as the training set.

[0082] When selecting a neural network type in ②, any one of the following networks can be selected:

[0083] Multi-View Convolutional Neural Network (MVCNN);

[0084] Group View Convolutional Neural Network (GVCNN);

[0085] Multi-View Long Short-Term Memory Network (MV-LSTM);

[0086] Multi-View Deformable Neural Network (MVTransformer).

[0087] It should be noted that the target discrimination model also needs to output the position coordinates of multiple targets in the top view of the lawn to be mowed for subsequent possible dangerous target clearance tasks.

[0088] Specifically, if the categories of the targets included in the top view of the lawn to be mowed do not belong to dangerous targets, continue to move forward.

[0089] Preferably, if a dangerous target is included, after clearing the dangerous target, continue to move forward, including:

[0090] Determine the position coordinates of the dangerous target, and control the broom of the automatic lawn mower to clear the dangerous target to prevent the cutting blade of the automatic lawn mower from contacting the dangerous target.

[0091] Specifically, brooms are arranged on both sides of the automatic lawn mower. When it is detected that the top view of the area to be mowed includes a dangerous target, that is, the area to be mowed includes a dangerous target, the dangerous target needs to be cleared first, that is, the dangerous target is cleared by the broom to prevent the dangerous target from damaging the automatic lawn mower and reducing the service life of the automatic lawn mower.

[0092] Specifically, in step S3, if there is an obstacle outside the second distance, re-plan the forward path and continue to move forward according to the re-planned forward path.

[0093] The outside of the second distance means a distance farther than the first distance on the forward path of the automatic lawn mower. At this time, detecting whether there is an obstacle outside the second distance is to prevent the automatic lawn mower from hitting the obstacle, causing irreparable damage to the automatic lawn mower, and at the same time avoiding damage to the obstacle. The obstacle may be an agricultural tool or a sign, etc.

[0094] Preferably, if there is an obstacle outside the second distance, re-plan the forward path, including:

[0095] Determine the third distance between the obstacle and the automatic lawn mower, and the first deflection angle and the second deflection angle between both sides of the obstacle and the automatic lawn mower;

[0096] Determine the rotation angle of the automatic lawn mower according to the third distance, the first deflection angle and the second deflection angle, so that the automatic lawn mower rotates to reach the rotation angle;

[0097] The automatic lawn mower continues to move forward according to the rotated angle. When the automatic lawn mower passes over the obstacle, it returns to the initial path and continues to move forward.

[0098] Specifically, when there is an obstacle outside the second distance, if the automatic lawn mower does not make an evasive maneuver, it will cause serious consequences. At this time, it is necessary to re-design the forward path of the automatic lawn mower to avoid the obstacle.

[0099] Specifically, such as Figure 2As shown in the figure, measure the distance L1 between the left side of the automatic lawn mower and the left side of the obstacle and the distance R2 from the right side of the obstacle, and measure the distance L2 between the right side of the automatic lawn mower and the left side of the obstacle and the distance R1 from the right side of the obstacle. Calculate the first deflection angle and the second deflection angle through trigonometric functions, compare the magnitudes of the first deflection angle and the second deflection angle, determine the direction in which the automatic lawn mower rotates, and determine the re-planned path.

[0100] Specifically, select the larger one of the first deflection angle and the second deflection angle as the rotation angle of the automatic lawn mower, so that the automatic lawn mower rotates. When the rotation angle is reached, the automatic lawn mower continues to move forward according to the rotated angle. When the automatic lawn mower crosses the obstacle, that is, when the driving distance is greater than the third distance, at this time, the automatic lawn mower continues to drive along the initial path and will not collide with the obstacle.

[0101] Compared with the prior art, the safety protection method of the automatic lawn mower provided in this embodiment collects the mowing image in the forward direction in real time during the mowing process of the automatic lawn mower, judges whether the mowing image includes dangerous targets, reduces the damage of the automatic lawn mower, and improves the safety of the automatic lawn mower during the mowing process; uses a deep neural network to identify the mowing image and quickly obtains the recognition result, so that the automatic lawn mower improves the mowing safety without affecting the mowing efficiency; at the same time, it judges whether there are obstacles on the forward path of the automatic lawn mower. If there are obstacles, the automatic lawn mower bypasses the obstacles, improving the safety of the automatic lawn mower during the mowing process.

[0102] Those skilled in the art can understand that all or part of the processes for implementing the methods of the above embodiments can be completed by instructing relevant hardware through a computer program, and the program can be stored in a computer-readable storage medium. Among them, the computer-readable storage medium is a magnetic disk, an optical disk, a read-only memory or a random access memory, etc.

[0103] The above is only a specific and preferred embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily thought of by those skilled in the art within the technical scope disclosed by the present invention should be covered within the protection scope of the present invention.

Claims

1. A safety protection method for an automatic lawn mower, characterized in that, The described safety protection method includes: Step S1: Collect a top view of the grass to be cut at a first distance in the forward direction of the automatic lawn mower and detect whether there are obstacles within a second distance; the first distance is determined according to the forward speed of the automatic lawn mower and the image acquisition period, and the second distance is greater than the first distance; Step S2: Determine whether there are dangerous targets in the top view of the grass to be cut; if there are dangerous targets, remove the dangerous targets and then continue to move forward; Step S3: If there are obstacles outside the second distance, re-plan the forward path and continue to move forward according to the re-planned forward path.

2. The safety protection method according to claim 1, wherein The collection of the top view of the grass to be cut at a first distance in the forward direction of the automatic lawn mower includes: An image acquisition device is detachably arranged on the top of the automatic lawn mower. When the automatic lawn mower moves forward, the height of the image acquisition device is adjusted so that the image acquisition device can collect the grassland area at the first distance to obtain a top view of the grass to be cut.

3. The safety protection method according to claim 2, wherein The adjustment of the height of the image acquisition device includes: Determine the real-time height of the image acquisition device according to the forward speed of the automatic lawn mower, the image acquisition period, and the downward viewing angle of the image acquisition device; By adjusting the height of the support rod between the image acquisition device and the automatic lawn mower, the image acquisition device reaches the real-time height.

4. The safety protection method according to claim 3, characterized in that, The real-time height of the image acquisition device is calculated by the following formula: H = (V * T) / tanθ; Wherein, V represents the forward speed of the automatic lawn mower, T represents the acquisition period, and θ represents the downward viewing angle of the image acquisition device.

5. The safety protection method according to any one of claims 2-4, characterized in that, The image acquisition device is any one of the following: High-definition camera; Micro scanner; Video recorder.

6. The safety protection method according to claim 1, wherein, The determination of whether there are dangerous targets in the top view of the grass to be cut includes: Input the collected top view of the grass to be cut into a pre-trained target discrimination model to obtain the category and position coordinates of the targets included in the top view of the grass to be cut; the target discrimination model is a deep neural network based on the attention mechanism; Based on the category of the targets included in the top view of the grass to be cut, determine whether they are dangerous targets.

7. The safety protection method according to claim 6, wherein If the categories of the targets included in the top view of the grass to be cut do not belong to dangerous targets, continue to move forward.

8. The safety protection method according to claim 6, characterized in that, The "if there are dangerous targets, remove the dangerous targets and then continue to move forward" includes: Determine the position coordinates of the dangerous targets, and control the broom of the automatic lawn mower to remove the dangerous targets to prevent the cutting blade of the automatic lawn mower from contacting the dangerous targets.

9. The safety protection method according to claim 1, wherein The "if there are obstacles outside the second distance, re-plan the forward path" includes: Determine the third distance between the obstacle and the automatic lawn mower, and the first deflection angle and the second deflection angle on both sides of the obstacle with respect to the automatic lawn mower; Determine the rotation angle of the automatic lawn mower according to the third distance, the first deflection angle, and the second deflection angle, so that the automatic lawn mower rotates to reach the rotation angle; The automatic lawn mower continues to move forward according to the rotated angle. When the automatic lawn mower crosses the obstacle, it returns to the initial path and continues to move forward.

10. The safety protection method according to claim 6, wherein The target discrimination model is any one of the following networks: Multi-view convolutional neural network MVCNN; Group-view convolutional neural network GVCNN; Multi-view long short-term memory network MV-LSTM; Multi-View Deformable Neural Network MVTransformer.

Images