Seeding quality acquisition method and system, electronic device and computer readable storage medium

Abstract

The embodiment of the application provides a sowing quality acquisition method, a system, an electronic device and a computer readable storage medium, and belongs to the technical field of data processing. The method comprises the following steps: inputting a plurality of continuous to-be-detected pictures about sowing conditions of a work field into a preset detection network, detecting a plurality of boundary boxes in each to-be-detected picture and confidence of each boundary box, determining a unique target box of each exposed seed from the boundary boxes according to the confidence, counting the number of the target boxes of each to-be-detected picture, obtaining the number of exposed seeds of each to-be-detected picture, then obtaining a seed exposure rate of each to-be-detected picture according to the number of exposed seeds, and obtaining sowing quality according to the seed exposure rates of the to-be-detected pictures with a preset number of continuous frames, so that the problem that an existing sowing machine is difficult to obtain sowing quality is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of data processing, in particular to a sowing quality acquisition method and system, an electronic device and a computer readable storage medium. BACKGROUND

[0002] With the development of intelligent agricultural technology, the application of seeding machines in agriculture is also becoming mature, and more and more agricultural scenarios use seeding machines for sowing. Seeding machine operation reduces labor costs and improves sowing efficiency.

[0003] However, the seeding machine also has many shortcomings. In the sowing process, the seeding machine is affected by many factors, such as the blocking of the soil covering wheel and the too fast sowing speed, which causes some seeds sown by the seeding machine to be not effectively covered by the soil, resulting in poor sowing. However, the current seeding machine is difficult to obtain the sowing quality when performing sowing operation. SUMMARY

[0004] Therefore, the purpose of the present application is to provide a sowing quality acquisition method and system, an electronic device and a computer readable storage medium, which can improve the problem that the existing seeding machine cannot monitor the sowing quality of seeds and is difficult to obtain the sowing quality when performing sowing operation.

[0005] In order to achieve the above purpose, in a first aspect, the present application provides a sowing quality acquisition method, which adopts the following technical solution.

[0006] A sowing quality acquisition method, the method comprising:

[0007] inputting a plurality of frames of to-be-tested pictures about the sowing situation of a work field into a preset detection network, detecting a plurality of bounding boxes in each frame of the to-be-tested pictures and the confidence of each bounding box, the detection network being trained to output the bounding box in the picture and the confidence of the bounding box when the picture is input, and the bounding box representing the position of the exposed seed;

[0008] determining a unique target box of each exposed seed from the bounding box according to the confidence, counting the number of the target boxes of each frame of the to-be-tested pictures, and obtaining the number of exposed seeds of each frame of the to-be-tested pictures;

[0009] obtaining the seed exposure rate of each frame of the to-be-tested pictures according to the number of exposed seeds, and obtaining the sowing quality according to the seed exposure rate of a plurality of frames of the to-be-tested pictures.

[0010] Further, the step of obtaining the sowing quality according to the seed exposure rate of a plurality of frames of the to-be-tested pictures comprises:

[0011] If the seed exposure rate of the to-be-tested picture of the continuous preset frame number is greater than the preset value, it is determined that the seeding quality is poor, otherwise it is determined that the seeding quality is good.

[0012] Further, after the step of obtaining the seeding quality, the method further comprises:

[0013] If the seeding quality is poor, a warning signal is sent out.

[0014] Further, the step of determining the target frame unique to each exposed seed from the bounding box according to the confidence comprises:

[0015] The bounding box with the maximum confidence among all the bounding boxes of the same exposed seed is determined as the comparison frame of the exposed seed;

[0016] The intersection over union value of the comparison frame of each exposed seed and the remaining bounding boxes of the exposed seed is calculated in turn using the intersection over union loss function;

[0017] For each exposed seed, all the intersection over union values are compared with a threshold value, if none of the intersection over union values is greater than the threshold value, the coordinate information of the comparison frame is taken as the target coordinate information of the exposed seed, and a target frame is determined according to the target coordinate information.

[0018] Otherwise, the comparison frame is excluded, and a new comparison frame is determined among the remaining bounding boxes of the exposed seed until the target coordinate information of the exposed seed is obtained.

[0019] Further, the step of obtaining the seed exposure rate of each frame of the to-be-tested picture according to the number of exposed seeds comprises:

[0020] According to the working field coverage area of the to-be-tested picture, the total amount of seeds involved in each frame of the to-be-tested picture is obtained in combination with the seeding specification;

[0021] According to the number of exposed seeds and the total amount of seeds of each frame of the to-be-tested picture, the seed exposure rate of each frame of the to-be-tested picture is obtained.

[0022] Further, the detection network comprises a convolution layer, a pooling layer, a full connection layer and an output layer, and the output layer comprises a linear activation function.

[0023] In a second aspect, an embodiment of the present application provides a seeding quality acquisition system, which adopts the following technical scheme.

[0024] A seeding quality acquisition system comprises a seeder and a UAV, the UAV is provided with a camera, the seeder is in communication connection with the UAV, and the UAV is provided with a flight control subsystem and an algorithm processing subsystem;

[0025] The seeder is used for seeding a work field according to a set seeding specification, and generates flight following information in real time, and sends the flight following information to the unmanned aerial vehicle;

[0026] The flight control subsystem is used for controlling the unmanned aerial vehicle to follow the seeder to fly based on the flight following information by using a preset dynamic route generation algorithm.

[0027] The camera is used for capturing a to-be-tested picture of the seeded work field in real time, and sending the to-be-tested picture to the algorithm processing subsystem.

[0028] The algorithm processing subsystem is used for acquiring the seeding quality of the seeder according to the to-be-tested picture by using the seeding quality acquisition method in the first aspect.

[0029] Further, the seeding quality acquisition system further includes a cloud platform, and the seeder is installed with an alarm subsystem, and the cloud platform is in communication connection with the seeder.

[0030] The cloud platform is used for setting a work plan of the seeder, and controlling the seeder to perform seeding work according to the work plan.

[0031] The algorithm processing subsystem is further used for sending a warning signal to the alarm subsystem in a case where the seeding quality is suboptimal.

[0032] The alarm subsystem is used for sending an alarm signal to the cloud platform according to the warning signal, so as to prompt the cloud platform to perform alarm.

[0033] In a third aspect, an embodiment of the present application provides a seeding quality acquisition system, which adopts the following technical solution.

[0034] A seeding quality acquisition system includes a seeder, the seeder is installed with a camera and an algorithm processing subsystem, and the camera is in communication connection with the algorithm processing subsystem.

[0035] The seeder is used for seeding a work field.

[0036] The camera is used for capturing a to-be-tested picture of the seeded work field in real time, and sending the to-be-tested picture to the algorithm processing subsystem.

[0037] The algorithm processing subsystem is used for acquiring the seeding quality of the seeder according to the to-be-tested picture by using the seeding quality acquisition method in the first aspect.

[0038] In a fourth aspect, an embodiment of the present application provides a seeding quality system, which adopts the following technical solution.

[0039] The application discloses a sowing quality acquisition system, comprising a sowing machine and a UAV which are communicatively connected, wherein the sowing machine is provided with an algorithm processing subsystem, and the UAV is provided with a camera and a flight control subsystem.

[0040] The sowing machine is used for sowing a work field and generating flight following information in real time, and the flight following information is sent to the UAV.

[0041] The flight control subsystem is used for controlling the UAV to follow the sowing machine to fly based on the flight following information and by using a preset dynamic route generation algorithm.

[0042] The camera is used for shooting a to-be-tested picture of the sowed work field in real time, and the to-be-tested picture is sent to the algorithm processing subsystem.

[0043] The algorithm processing subsystem is used for acquiring the sowing quality of the sowing machine by using the sowing quality acquisition method according to the to-be-tested picture.

[0044] In the fifth aspect, an electronic device is provided, and the technical scheme is as follows.

[0045] The electronic device comprises a processor and a memory, wherein the memory stores machine executable instructions which can be executed by the processor, and the processor can execute the machine executable instructions to implement the sowing quality acquisition method according to the first aspect.

[0046] In the sixth aspect, a computer readable storage medium is provided, and the technical scheme is as follows.

[0047] The computer readable storage medium stores a computer program, and the computer program is executed by a processor to implement the sowing quality acquisition method according to the first aspect.

[0048] The sowing quality acquisition method, system, electronic device and computer readable storage medium provided by the application can input a plurality of continuous work area sowing condition to-be-tested pictures into a detection network, obtain a boundary box representing a bare seed position, determine a unique target box of each bare seed according to the confidence of the boundary box, count the number of bare seeds in each to-be-tested picture according to the target box, and obtain the sowing quality according to the seed bare rate of the to-be-tested pictures of a continuous preset frame number. In the sowing operation, the sowing quality can be monitored in real time, and the sowing quality can be quickly obtained according to the monitored to-be-tested picture, so that the problem that the sowing quality of the seeds cannot be monitored and the sowing quality cannot be obtained when the existing sowing machine performs the sowing operation can be solved. BRIEF DESCRIPTION OF DRAWINGS

[0049] In order to make the technical solutions of the embodiments of the present application clearer, the following will briefly introduce the drawings needed to be used in the embodiments. It should be understood that the following drawings only show some of the embodiments of the present application, and therefore should not be considered as limiting the scope. Other related drawings can also be obtained by those of ordinary skill in the art without creative effort.

[0050] Figure 1 One of the block diagrams of the seeding quality acquisition system provided by the embodiments of the present application is shown.

[0051] Figure 2 One of the flow diagrams of the seeding quality acquisition method provided by the embodiments of the present application is shown.

[0052] Figure 3 Another flow diagram of the seeding quality acquisition method provided by the embodiments of the present application is shown.

[0053] Figure 4 The flow diagram of part of the sub-steps of step S101 in Figure 2 and Figure 3 is shown.

[0054] Figure 5 The flow diagram of part of the sub-steps of step S102 in Figure 2 and Figure 3 is shown.

[0055] Figure 6 The flow diagram of part of the sub-steps of step S103 in Figure 2 and Figure 3 is shown.

[0056] Figure 7 The structural diagram of the seeding quality acquisition system provided by the embodiments of the present application is shown.

[0057] Figure 8 The block diagram of the seeding quality acquisition system provided by the embodiments of the present application is shown.

[0058] Figure 9 One of the working flow diagrams of the seeding quality acquisition system provided by the embodiments of the present application is shown.

[0059] Figure 10 Another working flow diagram of the seeding quality acquisition system provided by the embodiments of the present application is shown.

[0060] Figure 11 The third working flow diagram of the seeding quality acquisition system provided by the embodiments of the present application is shown.

[0061] Figure 12A block diagram of a sowing quality acquisition device is shown.

[0062] Figure 13 A block diagram of an electronic device is shown.

[0063] Icon: 100-sowing quality acquisition system; 110-seeder; 120-camera; 130-algorithm processing subsystem; 140-cloud platform; 150-alarm subsystem; 160-unmanned aerial vehicle; 170-tractor; 180-sowing mechanism; 190-positioning subsystem; 200-unmanned driving subsystem; 210-wireless communication subsystem; 220-gimbal; 240-flight control subsystem; 250-wireless communication module; 260-sowing quality acquisition device; 270-detection module; 280-computing module; 290-electronic device. DETAILED DESCRIPTION

[0064] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only some of the embodiments of the present application, but not all the embodiments. The components of the embodiments of the present application described and shown in the accompanying drawings can be arranged and designed in various different configurations.

[0065] Therefore, the following detailed description of the embodiments of the present application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but only represents selected embodiments of the present application. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present application.

[0066] It should be noted that the relational terms such as "first" and "second" and the like are used only to distinguish one entity or operation from another, and do not necessarily require or imply that these entities or operations exist in any such actual relationship or order. Moreover, the terms "comprising", "including" or any other variant thereof are intended to cover non-exclusive inclusion, so that a process, method, article or apparatus including a series of elements includes not only those elements, but also other elements not explicitly listed or inherent to such a process, method, article or apparatus. Without more limitations, the element defined by the statement "comprising a" does not exclude the presence of additional identical elements in the process, method, article or apparatus including the element.

[0067] The seeding quality of the seeding machine 110 is affected by many factors during the seeding process. For example, the blocking of the soil covering wheel and the too fast seeding speed can cause the seeds to be not effectively covered by the soil, and thus the poor seeding can occur. Once the poor seeding occurs, the seeding machine 110 needs to be stopped in time to eliminate the fault and improve the seeding quality.

[0068] However, the current seeding machine 110 cannot know the seeding quality during the seeding operation.

[0069] Based on the above considerations, the present application provides a seeding quality acquisition scheme which can acquire the seeding quality and improve the problem that the current seeding machine 110 cannot know the seeding quality. In the following, the seeding quality acquisition scheme will be introduced in detail from multiple angles.

[0070] The seeding quality acquisition method provided by the embodiment of the present application can be applied to the seeding quality acquisition system 100 as shown in the figure. Figure 1 The detection system includes a seeding machine 110, and the seeding machine 110 is provided with a camera 120 and an algorithm processing subsystem 130, and the camera 120 and the algorithm processing subsystem 130 are in communication connection.

[0071] The seeding machine 110 is used for seeding the work field.

[0072] The seeding machine 110 can be unmanned seeding, that is, the work field is seeded according to the preset seeding specification, and the seeding path is planned in advance. Alternatively, the seeding machine 110 can be manually driven seeding, and in this case, the seeding path is determined by the driver.

[0073] The camera 120 is used for real-time shooting of the to-be-tested picture of the seeded work field, and the to-be-tested picture is sent to the algorithm processing subsystem 130.

[0074] The camera 120 can be arranged at a position close to the seeding position of the seeding machine 110, so as to be able to seed the work field.

[0075] Since the to-be-tested picture is shot in real time, the to-be-tested picture is continuous.

[0076] The algorithm processing subsystem 1301 acquires the seeding quality of the seeding machine 110 according to the to-be-tested picture by using the seeding quality acquisition method provided by the embodiment of the present application.

[0077] The above seeding quality acquisition system 100 shoots the seeded work field in time, and acquires the seeding quality of the seeding machine 110 by using the seeding quality acquisition method according to the continuous to-be-tested picture, so as to monitor the seeding condition during the seeding process, control the seeding quality, and improve the seeding efficiency.

[0078] For timely processing of abnormal seeding, please continue to refer to Figure 1 The seeding quality acquisition system 100 can also include a cloud platform 140, and the seeding machine 110 is installed with an alarm subsystem 150, and the cloud platform 140 and the seeding machine 110 can be connected through network communication.

[0079] The algorithm processing subsystem 130 is also used to send a pre-warning signal to the alarm subsystem 150 in the case of poor seeding quality.

[0080] A preset value can be set, and the seeding quality is poor when the seeding quality is less than the preset value. The judgment method of the seeding quality can also be adaptively adjusted according to the actual application. In the embodiment, no specific limitation is made.

[0081] The alarm subsystem 150 is used to send an alarm signal to the cloud platform 140 according to the pre-warning signal, so as to prompt the cloud platform 140 to alarm.

[0082] The cloud platform 140 alarms to timely inform the operation and maintenance personnel to process the seeding machine 110.

[0083] In order to timely remind the driver or staff on the seeding machine 110, the seeding machine 110 can also be installed with a sound generator and / or a signal lamp in communication connection with the alarm subsystem 150.

[0084] The alarm subsystem 150 sends an alarm signal to the sound generator and the signal lamp according to the pre-warning signal, the sound generator responds to the alarm signal to perform a beeping alarm, and the signal lamp responds to the alarm signal to perform a warning light alarm. In order to remind the driver or staff to stop seeding and repair the seeding machine 110.

[0085] In order to more specifically introduce the seeding quality acquisition method, in an embodiment, referring to Figure 2 , a flowchart of a seeding quality acquisition method provided by the present application. The present embodiment mainly takes the method applied to the algorithm processing subsystem 130 in the seeding quality acquisition system 100 as an example. In the present embodiment, the method can include the following steps. Figure 1

[0086] S101, input a plurality of frames of to-be-tested pictures about the seeding situation of the working field into a preset detection network, and detect a plurality of bounding boxes in each frame of to-be-tested picture and the confidence of each bounding box.

[0087] The detection network is trained to output the bounding box in the picture and the confidence of the bounding box when the picture is input. The bounding box represents the position of the exposed seed, and the confidence represents the probability that the bounding box is the real bounding box of the exposed seed.

[0088] ​Specifically, the algorithm processing subsystem 130 inputs the continuous multiple frames of the to-be-tested pictures about the sowing situation of the work field into a preset detection network, detects multiple bounding boxes in each frame of the to-be-tested pictures and the confidence of each bounding box, and each frame of the to-be-tested pictures may have multiple bounding boxes.

[0089] S102, according to the confidence, the target box unique to each exposed seed is determined from the bounding boxes, the number of target boxes of each frame of the to-be-tested pictures is counted, and the number of exposed seeds of each frame of the to-be-tested pictures is obtained.

[0090] Specifically, for each frame of the to-be-tested pictures, the algorithm processing subsystem 130 determines the target box unique to each exposed seed from the multiple bounding boxes of the frame of the to-be-tested pictures, counts the number of target boxes of the frame of the to-be-tested pictures, and obtains the number of exposed seeds of the frame of the picture. The above steps are repeated for each frame of the to-be-tested pictures to obtain the number of exposed seeds of each frame of the to-be-tested pictures.

[0091] S103, according to the number of exposed seeds, the seed exposure rate of each frame of the to-be-tested pictures is obtained, and according to the seed exposure rates of the continuous preset number of frames of the to-be-tested pictures, the sowing quality is obtained.

[0092] The preset number of frames can be three frames, and since the three frames of the to-be-tested pictures are continuously taken, the shooting objects of the three frames of the to-be-tested pictures are basically the same. In addition, the to-be-tested ranges of the multiple frames of the to-be-tested pictures taken by the camera 120 are the same or different.

[0093] Specifically, the algorithm processing subsystem 130 obtains the seed exposure rate of each frame of the to-be-tested pictures according to the number of seeds exposed in each frame of the to-be-tested pictures. And according to the seed exposure rates of the continuous three frames of the to-be-tested pictures, the sowing quality of the sowing machine 110 is obtained.

[0094] In the above sowing quality obtaining method, by inputting the multiple frames of the to-be-tested pictures about the sowing situation of the work field into the detection network, the bounding boxes representing the positions of the exposed seeds are obtained, and according to the confidence of the bounding boxes, the target box unique to each exposed seed is determined, the number of exposed seeds in each frame of the to-be-tested pictures is counted according to the target box, and the sowing quality is obtained according to the seed exposure rates of the continuous preset number of frames of the to-be-tested pictures. In unmanned operation, the sowing quality can be monitored in real time, and the sowing quality can be quickly obtained according to the monitored to-be-tested pictures, so as to solve the problem that the existing unmanned sowing machine 110 cannot monitor the sowing quality of the seeds in the unmanned operation and is difficult to obtain the sowing quality.

[0095] On the basis of the above, Figure 3 The sowing quality obtaining method provided by the present application further comprises a step S104.

[0096] S104, if the sowing quality is poor, a warning signal is sent out.

[0097] Specifically, if the sowing quality obtained by the algorithm processing subsystem 130 is "poor", a signal "1" is sent to the alarm subsystem 150.

[0098] In the case of analyzing that the sowing quality is poor, the algorithm processing subsystem 130 sends a warning signal to the alarm subsystem 150, and the alarm subsystem 150 sends an alarm signal to the cloud platform 140, the signal lamp and the sound generator, etc. The cloud platform 140, the signal lamp and the sound generator immediately alarm to remind the control personnel to stop the work of the sowing machine 110 and remind the maintenance personnel to timely repair, so as to timely eliminate the fault and improve the sowing quality.

[0099] Further, the detection network includes a convolution layer, a pooling layer, a full connection layer, a Softmax layer and an output layer, and the output layer includes a linear activation function. By adding a linear activation function to the output layer, the nonlinearity of the detection network is increased, and the prediction accuracy is improved.

[0100] In an embodiment, the local features of the to-be-detected picture are extracted by the convolution layer, the global features of the to-be-detected picture are extracted by the pooling layer, the local features and the global features are integrated by the full connection layer to obtain a plurality of bounding boxes, the confidence of each bounding box is obtained by the Softmax layer, and the bounding boxes and the confidence of each bounding box are output by the output layer.

[0101] In order to better understand the processing process of step S101, refer to Figure 4 The above step S101 can include the following steps.

[0102] S101-1, the to-be-detected picture is divided into a grid including S*S cells.

[0103] S101-2, based on each cell, a plurality of bounding boxes of the exposed seeds falling in the cell are predicted, and the coordinate information and the confidence of each bounding box are obtained.

[0104] Each bounding box can be represented by four values, which can be (x, y, w, h), wherein (x, y) is the center coordinate of the bounding box, w is the width of the bounding box, and h is the height of the bounding box. (x, y, w, h) is the coordinate information. The confidence is represented by c. In fact, the predicted value of each bounding box includes five elements (x, y, w, h, c).

[0105] On this basis, if the to-be-detected picture is divided into S*S grids, (B*5+C) values are predicted for each cell, and the final predicted value is a tensor of S*S*(B*5+C), that is, an array.

[0106] The calculation formula of the confidence is:

[0107] wherein, represents the confidence of the jth bounding box of the ith cell. represents the intersection over union value of the predicted current bounding box and the real bounding box of the object when the current bounding box has the object. Pr(Object) represents the probability of whether the current bounding box has the object. The acquisition methods of Pr(Object) and Pr(Non-object) are common algorithms in target detection, which will not be described one by one here.

[0108] Please refer to Figure 5 The flowchart of part of the sub-steps of step S102 is shown in the following sub-steps, which can determine the target box unique to each exposed seed from the bounding boxes according to the confidence.

[0109] S102-1, the bounding box with the highest confidence among all the bounding boxes of the same exposed seed in the to-be-detected picture is determined as the comparison box of the exposed seed.

[0110] Specifically, the algorithm processing subsystem 130 determines the bounding box with the highest confidence among all the bounding boxes of the same exposed seed in the to-be-detected picture as the comparison box of the exposed seed for each frame of the to-be-detected picture.

[0111] For example, if there are 8 exposed seeds in a frame of to-be-detected picture, there are 8 comparison boxes.

[0112] S102-2, the intersection over union value of each exposed seed comparison box and the remaining bounding boxes of the exposed seed is calculated in turn by using the intersection over union loss function.

[0113] Specifically, the algorithm processing subsystem 130 calculates the intersection over union between the comparison box of each exposed seed and each of the remaining bounding boxes of the exposed seed, and obtains a plurality of intersection over union values.

[0114] For example, there are 8 exposed seeds on a frame of to-be-detected picture, each exposed seed has 8 bounding boxes, and there are 7 remaining bounding boxes after removing the comparison box. Therefore, each exposed seed will calculate 7 intersection over union values.

[0115] S102-3, for each exposed seed, compare all the intersection over union values with the threshold value.

[0116] If none of the intersection over union values is greater than the threshold value, step S102-4 is performed, otherwise step S102-5 is performed, and steps S102-1 to S102-3 are repeated after S102-5 until the target coordinate information of the exposed seed is obtained.

[0117] Wherein, the intersection over union value greater than the threshold value means that the coincidence degree of the bounding box and the comparison box is too high.

[0118] S102-4, taking the coordinate information of the comparison frame as target coordinate information of the exposed seed, and determining a target frame according to the target coordinate information.

[0119] The target coordinate information is the center coordinate and the length and width of the target frame, and the four corner coordinates of the target frame can be quickly obtained according to the center coordinate and the length and width, so that the position and range of the target frame can be determined.

[0120] S102-5, eliminating the comparison frame.

[0121] After eliminating the comparison frame, the comparison frame is retained, and the coincidence degree is reduced.

[0122] In the above S102-1 to S102-5, one of the boundary frames with high coincidence degree is eliminated, and finally a unique target frame is obtained, which can improve the accuracy of the target frame.

[0123] It should be understood that since the target coordinate information is the coordinate information of the comparison frame, the target coordinate information is also (x, y, w, h).

[0124] In order to introduce the seed exposure rate in more detail, in an embodiment, please refer to Figure 6 The seed exposure rate of each frame of the to-be-tested picture can be obtained according to the number of exposed seeds in S103 by the following steps.

[0125] S103-1, according to the working field coverage area of the to-be-tested picture, and combining the sowing specification, the total amount of seeds involved in each frame of the to-be-tested picture is obtained.

[0126] The working field coverage area of the to-be-tested picture can be obtained through picture processing, or can be the preset shooting area of the camera 120. The sowing specification is preset. The working field coverage area and the sowing specification are combined to obtain the total amount of seeds involved in the to-be-tested picture.

[0127] S103-2, according to the number of exposed seeds of each frame of the to-be-tested picture and the total amount of seeds, the seed exposure rate of each frame of the to-be-tested picture is obtained.

[0128] The seed exposure rate of each frame of the to-be-tested picture can be obtained by dividing the number of exposed seeds by the total amount of seeds.

[0129] The seed exposure rate of each frame of the to-be-tested picture can be quickly obtained through the above S103-1 to S103-2.

[0130] It should be understood that the method of the above S103-1 to S103-2 is only an example of one way, and is not the only limitation. In actual application, adaptive adjustment can be made according to actual situation.

[0131] To facilitate a more clear understanding of how to obtain the seeding quality, in an embodiment, the seeding quality can be obtained by the following method: if the seed exposure rates of the preset number of continuous images are all greater than the preset value, it is determined that the seeding quality is poor, otherwise, it is determined that the seeding quality is good.

[0132] The above method for determining the seeding quality is only an example, not the only one. In actual application, adaptive adjustment can be made according to actual needs, for example, when the seeding area needs to be considered, the seeding quality can be determined according to the ratio of the seed exposure rate to the seeding area.

[0133] It should be understood that, although Figures 2-6 The steps in the flowchart of the method for determining the seeding quality are displayed in sequence according to the arrows, but these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified herein, the execution of these steps is not strictly limited in sequence, and these steps can be executed in other orders. Moreover, Figures 2-6 At least part of the steps in the method for determining the seeding quality can include multiple sub-steps or multiple stages, which are not necessarily executed at the same time, but can be executed at different times, and the execution order of these sub-steps or stages is not necessarily sequential, but can be alternated or alternated with at least part of other steps or sub-steps or stages of other steps.

[0134] Based on the above concept of the method for obtaining the seeding quality, in an embodiment, referring to Figure 7 and Figure 8 The present application also provides a seeding quality obtaining system 100. The seeding quality obtaining system 100 comprises a seeder 110 and a UAV 160, the UAV 160 is provided with a camera 120, the seeder 110 and the UAV 160 are in communication connection, and the UAV 160 is provided with a flight control subsystem 240 and an algorithm processing subsystem 130.

[0135] The seeder 110 is used for seeding the working field and generating flight following information in real time, and the flight following information is sent to the UAV 160.

[0136] The seeder 110 provided in the above Figure 1 The seeder 110 can be unmanned or manually driven. When unmanned, the seeder 110 seeds the working field according to the preset seeding specifications, and the seeding path can be obtained by a seeding navigation algorithm.

[0137] The flight control subsystem 240 is used for controlling the UAV 160 to follow the seeder 110 to fly based on the flight following information and by using a preset dynamic route generation algorithm.

[0138] The flight following information includes position information, attitude information, flight speed information, flight height information, and following distance.

[0139] The camera 120 is configured to capture a to-be-tested image of the sowed work field in real time and send the to-be-tested image to the algorithm processing subsystem 130.

[0140] The algorithm processing subsystem 130 is configured to acquire the seeding quality of the seeding machine 110 by using a seeding quality acquisition method according to the to-be-tested image.

[0141] In more detail, the seeding machine 110 can include a tractor 170, a seeding mechanism 180, a positioning subsystem 190, an unmanned subsystem 200, and a wireless communication subsystem 210. The tractor 170 is configured to provide power, the seeding machine 110 is configured to automatically sow seeds, the positioning subsystem 190 is configured to provide accurate position coordinates and attitude angles for the seeding machine 110, the unmanned subsystem 200 is configured to control automatic driving control of the seeding machine 110, and the wireless communication subsystem 210 is configured to communicate with the unmanned aerial vehicle 160 to transmit information.

[0142] The positioning subsystem 190 can include a GPS-RTK satellite positioning receiver and an inertial measurement unit.

[0143] The unmanned aerial vehicle 160 includes a wireless communication module 250 configured to communicate data with the seeding machine 110.

[0144] Please continue to refer to Figure 1 and Figure 2 The unmanned aerial vehicle 160 further includes a gimbal 220 configured to mount and stabilize the camera 120.

[0145] Similar to the seeding quality acquisition system 100 described above Figure 1 Please continue to refer to Figure 7 and Figure 8 The seeding quality acquisition system 100 further includes a cloud platform 140, and the seeding machine 110 is further provided with an alarm subsystem 150, a signal lamp, and a sound generator. The cloud platform 140 is in communication connection with the seeding machine 110 through the wireless communication subsystem 210 to transmit information.

[0146] The cloud platform 140 is configured to set a work plan for the seeding machine 110 and control the seeding machine 110 to perform a seeding work according to the work plan.

[0147] In more detail, the unmanned subsystem 200 controls automatic driving control of the seeding machine 110 according to various work instructions issued by the cloud platform 140 according to the work plan, and the cloud platform 140 controls seeding work of the seeding mechanism 180 according to the work plan.

[0148] The algorithm processing subsystem 130 is also configured to send a warning signal to the alarm subsystem 150 when the seeding quality is poor.

[0149] For example, the algorithm processing subsystem 130 outputs a signal “1” to the alarm subsystem 150 when the seeding quality is poor, and outputs no signal or a signal “0” otherwise.

[0150] The alarm subsystem 150 is configured to send an alarm signal to the cloud platform 140, the sound generator, and the signal lamp according to the warning signal, so as to prompt the cloud platform 140, the sound generator, and the signal lamp to alarm.

[0151] More specifically, when the algorithm processing subsystem 130 outputs a signal “1”, the alarm subsystem 150 sends an alarm command to the cloud platform 140, the sound generator, and the signal lamp, and the cloud platform 140, the sound generator, and the signal lamp start to alarm.

[0152] In the seeding quality acquisition system 100, when the seeding machine 110 is performing the seeding operation, the wireless communication subsystem 210 continuously sends the flight following information to the unmanned aerial vehicle 160. After the flight control subsystem 240 on the unmanned aerial vehicle 160 receives the flight following information, the flight control subsystem 240 controls the unmanned aerial vehicle 160 to follow the seeding machine 110 at a preset distance according to the dynamic route generation algorithm. The camera 120 on the unmanned aerial vehicle 160 captures the to-be-tested pictures of the farmland that has been seeded by the seeding machine 110 in real time. The algorithm processing subsystem 130 processes the to-be-tested pictures to obtain the seeding quality of the seeding machine 110, and sends a warning according to the seeding quality, so as to prompt the cloud platform 140, the sound generator, and the signal lamp to alarm, so as to remind the seeding machine 110 to pause the seeding operation. This can greatly improve the problem that the existing unmanned seeding machine 110 cannot monitor the seeding quality of the seeds during the unmanned operation, and thus cannot know the seeding quality.

[0153] Reference Figure 9 FIG. 4 is a flowchart of the working process of the seeding quality acquisition system 100. When the seeding starts, the unmanned aerial vehicle 160 follows the seeding machine 110 according to the flight following information sent by the seeding machine 110, and obtains the seed exposure rate according to the to-be-tested pictures captured by the camera 120. When the seed exposure rate of a continuous preset number of frames is higher than a threshold value, a warning signal is sent. After receiving the alarm signal, the maintenance personnel starts to troubleshoot and handle the fault. After the fault is resolved, the seeding continues until the seeding is completed.

[0154] In other embodiments, the algorithm processing subsystem 130 can also be arranged on the seeding machine 110.

[0155] And, in other embodiments, after the reservation algorithm processing subsystem 130 and the camera 120, the unmanned aerial vehicle 160 is omitted, and the camera 120 is installed on the seeder 110 away from the vehicle head, and the camera 120 is directed to the working field, and the camera 120 communicates with the algorithm processing subsystem 130. At the same time, the algorithm processing subsystem 130 is arranged on the seeder 110, and the functions realized by the above-mentioned seeding quality acquisition system 100 can be realized.

[0156] At this time, refer to Figure 10 , the working flow chart after the unmanned aerial vehicle 160 in the seeding quality acquisition system 100. When the seeding starts, the camera 120 transmits the pictures in real time to the algorithm processing subsystem 130 on the seeder 110, and the algorithm processing subsystem 130 obtains the seed exposure rate according to the pictures taken by the camera 120. When it is judged that the seed exposure rate of a continuous preset number of frames is higher than the threshold value, a warning signal is sent out. After receiving the alarm signal, the maintenance personnel starts to troubleshoot and handle the fault, and continues to seed after the fault is removed, until the seeding is completed.

[0157] In addition, an image transmission module can also be added to the unmanned aerial vehicle 160, an image receiving module is added to the seeder 110, and the algorithm processing subsystem 130 is installed on the seeder 110. The seeder 110 can bear a larger and heavier algorithm processing subsystem 130 installation, and increase the operation speed. The pictures taken by the camera 120 are sent in real time to the image receiving module of the seeder 110 through the image transmission module. The functions realized by the above-mentioned seeding quality acquisition system 100 can be realized.

[0158] Refer to Figure 11 , the working flow chart when the image transmission module is added to the unmanned aerial vehicle 160, and the algorithm processing subsystem 130 and the image receiving module are located on the seeder 110. When the seeding starts, the unmanned aerial vehicle 160 follows the seeder 110 according to the flight following information sent by the seeder 110, and transmits the pictures taken by the camera 120 to the image receiving module of the seeder 110 through the image transmission module. The algorithm processing module obtains the seed exposure rate according to the pictures received by the image receiving module. When it is judged that the seed exposure rate of a continuous preset number of frames is higher than the threshold value, a warning signal is sent out. After receiving the alarm signal, the maintenance personnel starts to troubleshoot and handle the fault, and continues to seed after the fault is removed, until the seeding is completed.

[0159] Based on the above-mentioned seeding quality acquisition method, in one embodiment, a seeding quality acquisition system 100 is also provided, comprising a seeder 110, the seeder 110 is provided with a camera 120 and an algorithm processing subsystem 130, the camera 120 is in communication connection with the algorithm processing subsystem 130;

[0160] The seeder 110 is used for seeding the working field.

[0161] Camera 120 is used to capture images of the sown field in real time and send the images to the algorithm processing subsystem 130.

[0162] The algorithm processing subsystem 130 is used to obtain the sowing quality of the seeder 110 based on the image to be tested and using the sowing quality acquisition method provided in the above embodiment.

[0163] Compared to the seeding quality acquisition system 100 in the previous embodiment, this embodiment mounts the camera 120 on the seeder 110. Other limitations regarding the seeding quality acquisition system 100 can be found in the limitations described in the previous embodiment, and will not be repeated here.

[0164] Based on the above concept of the method for obtaining sowing quality, in one embodiment, referring to Figure 12 A seeding quality acquisition device 260 is provided, including a detection module 270 and a calculation module 280.

[0165] The detection module 270 is used to input multiple frames of test images related to the sowing situation in the work field into a preset detection network, and detect multiple bounding boxes and the confidence level of each bounding box in each frame of the test image.

[0166] The detection network is trained to output bounding boxes and their confidence scores in an image when the image is taken as input. The bounding boxes represent the locations of exposed seeds.

[0167] The calculation module 280 is used to determine the unique target box of each exposed seed from the bounding box based on the confidence level, count the number of target boxes in each frame of the test image, obtain the number of exposed seeds in each frame of the test image, obtain the seed exposure rate of each frame of the test image based on the number of exposed seeds, and obtain the sowing quality based on the seed exposure rate of the test images for a consecutive preset number of frames. The aforementioned sowing quality acquisition device 260, through the test images of the sown field, uses a detection network to detect the number of exposed seeds in each frame of the test image, and obtains the sowing quality based on the seed exposure rate of each frame of the test image, which can improve the problem of not being able to obtain the sowing quality of the seeder 110.

[0168] Furthermore, the seeding quality acquisition device 260 may also include an early warning module.

[0169] The early warning module is used to issue an early warning signal if the sowing quality is poor.

[0170] The specific definition of the sowing quality acquisition device 260 can be referred to the definition of the sowing quality acquisition method above, which will not be repeated here. Each module in the sowing quality acquisition device 260 described above can be realized by software, hardware and combination thereof in whole or in part. The modules described above can be embedded in or independent of the processor in the electronic device 290 in hardware form, or stored in the memory in the electronic device 290 in software form, so that the processor calls and executes the operations corresponding to each module.

[0171] In one embodiment, a computer device is provided, and an internal structure diagram of the computer device can be as shown in Figure 13 The electronic device 290 includes a processor, a memory and a network interface connected by a system bus. The processor of the electronic device 290 is used to provide computing and control capabilities. The memory of the electronic device 290 includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operating system and the computer program in the non-volatile storage medium to run. The database of the electronic device 290 is used to store the number of exposed seeds, the exposure rate of seeds and the sowing quality. The network interface of the electronic device 290 is used to communicate with the external terminal through the network connection. The computer program is executed by the processor to implement a sowing quality acquisition method.

[0172] Those skilled in the art can understand that Figure 13 The structure shown in the figure is only a block diagram of part of the structure related to the scheme of the present application, and does not constitute a limitation on the computer device to which the scheme of the present application is applied. The specific computer device can include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0173] In one embodiment, the sowing quality acquisition device 260 provided by the present application can be realized in the form of a computer program, which can run on the electronic device 290 as shown in Figure 13 The memory of the electronic device 290 can store various program modules constituting the sowing quality acquisition device 260, such as the detection module 270 and the calculation module 280 as shown in Figure 12 The computer program constituted by each program module enables the processor to execute the steps in the sowing quality acquisition method of each embodiment of the present application described in the specification.

[0174] For example, Figure 13 The electronic device 290 can execute step S101 through the detection module 270 in the sowing quality acquisition device 260 as shown in Figure 12 The electronic device 290 can execute steps S102-S103 through the calculation module 280.

[0175] In one embodiment, an electronic device 290 is provided, comprising a memory and a processor, the memory storing a computer program, the processor implementing the following steps when executing the computer program: inputting a plurality of frames of to-be-detected pictures about seeding conditions of a work field in succession into a preset detection network, detecting a plurality of bounding boxes in each frame of to-be-detected pictures and confidence of each bounding box, the detection network being trained to output, in a case of taking a picture as input, a bounding box in the picture and confidence of the bounding box, the bounding box representing a position of a bare seed; determining a unique target box of each bare seed from the bounding boxes according to the confidence, counting a number of target boxes of each frame of to-be-detected pictures, and obtaining a number of bare seeds of each frame of to-be-detected pictures; obtaining a seed bare rate of each frame of to-be-detected pictures according to the number of bare seeds, and obtaining a seeding quality according to the seed bare rates of a plurality of frames of to-be-detected pictures in succession.

[0176] In one embodiment, a computer readable storage medium is provided, storing a computer program, the computer program being executed by a processor to implement the following steps: inputting a plurality of frames of to-be-detected pictures about seeding conditions of a work field in succession into a preset detection network, detecting a plurality of bounding boxes in each frame of to-be-detected pictures and confidence of each bounding box, the detection network being trained to output, in a case of taking a picture as input, a bounding box in the picture and confidence of the bounding box, the bounding box representing a position of a bare seed; determining a unique target box of each bare seed from the bounding boxes according to the confidence, counting a number of target boxes of each frame of to-be-detected pictures, and obtaining a number of bare seeds of each frame of to-be-detected pictures; obtaining a seed bare rate of each frame of to-be-detected pictures according to the number of bare seeds, and obtaining a seeding quality according to the seed bare rates of a plurality of frames of to-be-detected pictures in succession.

[0177] Further limitations of the electronic device 290 and the computer readable storage medium are described in the detailed description of the seeding quality obtaining method above, and will not be repeated here.

[0178] In the several embodiments provided in this application, it should be understood that the disclosed apparatus and methods can also be implemented in other ways. The apparatus embodiments described above are merely illustrative; for example, the flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functions, and operations of possible implementations of apparatus, methods, and computer program products according to various embodiments of the present invention. In this regard, each block in a flowchart or block diagram may represent a module, program segment, or part of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions marked in the blocks may occur in a different order than those marked in the drawings. For example, two consecutive blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in a block diagram and / or flowchart, and combinations of blocks in block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or action, or using a combination of dedicated hardware and computer instructions.

[0179] In addition, the functional modules in the various embodiments of the present invention can be integrated together to form an independent part, or each module can exist independently, or two or more modules can be integrated to form an independent part.

[0180] If the aforementioned functions are implemented as software functional modules and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0181] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for obtaining seeding quality, characterized by, The method comprises: inputting continuous multiple frames of to-be-detected pictures about seeding conditions of a work field into a preset detection network, detecting multiple bounding boxes in each frame of the to-be-detected pictures and confidence of each bounding box, the detection network being trained to output the bounding box in the picture and the confidence of the bounding box when the picture is input, the bounding box representing the position of a bare seed; determining the bounding box with the maximum confidence among all the bounding boxes of the same bare seed as a comparison box of the bare seed; calculating the intersection-over-union value of the comparison box of each bare seed and the remaining bounding boxes of the bare seed in sequence by using an intersection-over-union loss function; for each bare seed, comparing all the intersection-over-union values with a threshold value, if none of the intersection-over-union values is greater than the threshold value, taking the coordinate information of the comparison box as the target coordinate information of the bare seed, and determining a target box according to the target coordinate information; otherwise, eliminating the comparison box, and determining a new comparison box among the remaining bounding boxes of the bare seed until the target coordinate information of the bare seed is obtained; counting the number of the target boxes of each frame of the to-be-detected pictures to obtain the number of bare seeds of each frame of the to-be-detected pictures; obtaining the seed bare rate of each frame of the to-be-detected pictures according to the number of bare seeds, and obtaining the seeding quality according to the seed bare rates of the to-be-detected pictures of continuous preset frames.

2. The sowing quality acquisition method according to claim 1, characterized by, The step of obtaining the seeding quality according to the seed bare rates of the to-be-detected pictures of continuous preset frames comprises: if the seed bare rates of the to-be-detected pictures of continuous preset frames are all greater than a preset value, determining that the seeding quality is poor, otherwise determining that the seeding quality is good.

3. The sowing quality acquisition method according to claim 2, characterized in that, After the step of obtaining the seeding quality, the method further comprises: if the seeding quality is poor, issuing a warning signal.

4. The sowing quality acquisition method according to claim 1, characterized by, The step of obtaining the seed bare rate of each frame of the to-be-detected pictures according to the number of bare seeds comprises: obtaining the total amount of seeds involved in each frame of the to-be-detected pictures according to the work field coverage area of the to-be-detected pictures and the seeding specification; obtaining the seed bare rate of each frame of the to-be-detected pictures according to the number of bare seeds and the total amount of seeds of each frame of the to-be-detected pictures.

5. The sowing quality acquisition method according to claim 1, characterized by, The detection network comprises a convolution layer, a pooling layer, a full connection layer and an output layer, and the output layer comprises a linear activation function.

6. A sowing quality acquisition system characterized by comprising: The method comprises: the seeding machine and the unmanned aerial vehicle, the camera is carried on the unmanned aerial vehicle, the seeding machine is in communication connection with the unmanned aerial vehicle, the unmanned aerial vehicle is provided with a flight control subsystem and an algorithm processing subsystem; the seeding machine is used for seeding the work field and generating flight following information in real time, and the flight following information is sent to the unmanned aerial vehicle; the flight control subsystem is used for controlling the unmanned aerial vehicle to follow the seeding machine to fly by using a preset dynamic route generation algorithm based on the flight following information; the camera is used for shooting the to-be-detected pictures of the seeded work field in real time, and sending the to-be-detected pictures to the algorithm processing subsystem; the algorithm processing subsystem is used for obtaining the seeding quality of the seeding machine by using the seeding quality acquisition method according to the to-be-detected pictures.

7. The sowing quality acquisition system according to claim 6, characterized in that, The seeding quality acquisition system further comprises a cloud platform, and the seeding machine is provided with an alarm subsystem, and the cloud platform is in communication connection with the seeding machine; The cloud platform is configured to set a work plan of the seeding machine and control the seeding machine to perform seeding work according to the work plan; The algorithm processing subsystem is further configured to send a pre-warning signal to the alarm subsystem in the case that the seeding quality is substandard; The alarm subsystem is configured to send an alarm signal to the cloud platform according to the pre-warning signal to prompt the cloud platform to perform alarm.

8. A seeding quality acquisition system characterized by comprising: The seeding machine is provided with a camera and an algorithm processing subsystem, and the camera is in communication connection with the algorithm processing subsystem; The seeding machine is configured to perform seeding on a work field; The camera is configured to capture a to-be-tested picture of the seeded work field in real time and send the to-be-tested picture to the algorithm processing subsystem; The algorithm processing subsystem is configured to acquire the seeding quality of the seeding machine by using the seeding quality acquisition method according to any one of claims 1 to 5 according to the to-be-tested picture.

9. A seeding quality acquisition system characterized by comprising: The seeding machine and the unmanned aerial vehicle are in communication connection, the seeding machine is provided with an algorithm processing subsystem, and the unmanned aerial vehicle is provided with a camera and a flight control subsystem; The seeding machine is configured to perform seeding on a work field and generate flight following information in real time, and send the flight following information to the unmanned aerial vehicle; The flight control subsystem is configured to control the unmanned aerial vehicle to follow the seeding machine to fly by using a preset dynamic route generation algorithm based on the flight following information; The camera is configured to capture a to-be-tested picture of the seeded work field in real time and send the to-be-tested picture to the algorithm processing subsystem; The algorithm processing subsystem is configured to acquire the seeding quality of the seeding machine by using the seeding quality acquisition method according to any one of claims 1 to 5 according to the to-be-tested picture.

10. An electronic device, comprising: The computer program is executed by the processor to implement the seeding quality acquisition method according to any one of claims 1 to 5.

11. A computer readable storage medium having stored thereon a computer program, characterized in that, The computer program is executed by the processor to implement the seeding quality acquisition method according to any one of claims 1 to 5.

Images