A seedling transplanting operation regulation method and system based on online identification of seedling survival rate

Abstract

The application discloses a root and fruit turning seedling operation regulation and control method and system based on online identification of clear fruit rate, and the method comprises the following steps: collecting a root and fruit image after turning seedling is laid every preset time through an image acquisition device; detecting the sampling clear fruit rate of the root and fruit image based on a pre-trained detection model; comparing the sampling clear fruit rate obtained through online identification with a benchmark clear fruit rate; if the sampling clear fruit rate is higher than the benchmark clear fruit rate, no regulation and control is needed; if the sampling clear fruit rate is lower than the benchmark clear fruit rate, a regulation and control program is started to adjust the rotating speed of a turning seedling tray. The application realizes online monitoring and intelligent regulation and control of the key indicator of clear fruit rate, significantly improves the automation level of root and fruit harvesting, reduces the demand for manual intervention, improves operation efficiency and consistency, and can improve the clear fruit rate through precise regulation and control to ensure the quality of turning and drying and the quality of harvested root and fruit.

Description

Technical Field

[0001] This invention relates to the field of control technology for root and fruit digging and turning harvesters, and in particular to a method and system for controlling root and fruit turning operations based on online identification of the fruit yield. Background Technology

[0002] In the harvesting of root and fruit crops such as peanuts, turning and drying are crucial steps to ensure harvest quality. Turning the vines exposes the roots and fruits buried in the soil to the surface, spreading them evenly so they are fully exposed to sunlight and air. This accelerates moisture evaporation, promotes uniform ripening, and reduces the risk of mold and rot, thereby improving subsequent harvesting and processing efficiency. Good turning and drying quality directly affects the yield and quality of the roots and fruits, significantly impacting agricultural production efficiency.

[0003] The applicant's earlier application, CN 117413673 A, discloses a digging and drying machine for root and fruit crops. It mainly includes a tractor-drawn frame, a digging mechanism, a conveying and lifting mechanism, and a turning and laying mechanism. It also includes a transmission system, which comprises a transmission box that connects to the power output shaft of the tractor's PTO (Power Take-Off) mechanism. The output shaft of the transmission box transmits power to the turning trays of the turning and laying mechanism via a chain transmission mechanism. In this technical solution, the drying quality lacks real-time monitoring. The operating speed of the turning trays dynamically changes with the machine's forward speed, leading to uneven drying and affecting the drying effect and subsequent harvest quality of the root and fruit crops. Summary of the Invention

[0004] Purpose of the invention: In order to overcome the shortcomings of the existing technology, the present invention provides a method and system for controlling the root and fruit turning operation based on online identification of the fruit turning rate, which can monitor the quality of turning and laying the seedlings and adjust the quality of turning and laying the seedlings in a timely manner.

[0005] Technical solution: To achieve the above objectives, the present invention provides a root and fruit turning operation control method based on online identification of fruit yield, the method comprising: An image acquisition device installed above the tail of the root and fruit digging and turning machine is used to capture images of the roots and fruits after turning and laying the rice at preset time intervals. The sampling fruit exposure rate of the root and fruit image is detected based on a pre-trained detection model; the fruit exposure rate refers to the proportion of the number of roots and fruits exposed within the sampling range of the image acquisition device after the seedlings are turned and laid out, out of the total number of roots and fruits. The sampling fruit rate calculated online is compared with the benchmark fruit rate. If the sampling fruit rate is higher than the benchmark fruit rate, no adjustment is needed. If the sampling fruit rate is lower than the benchmark fruit rate, the adjustment program is activated to adjust the rotation speed of the rice turning tray. The benchmark fruit rate is calculated in advance.

[0006] Furthermore, the detection model is a lightweight detection model for the accuracy rate based on deep learning YOLOv5s.

[0007] Furthermore, the control procedure described herein includes: If the seedling turning and laying mechanism is adjusted to reduce the rotation speed of the seedling turning tray by 5%, it is determined whether the seedling turning rate is lower than the baseline seedling turning rate in two consecutive samplings. If the sampling rate of clear fruit is lower than the benchmark rate after the speed is reduced twice consecutively after the rice turning tray is turned off, that is, if the sampling rate of clear fruit is lower than the benchmark rate after four consecutive tests, an audible and visual alarm will be activated to remind the driver to stop the operation and check.

[0008] Furthermore, the method for calculating the benchmark clear rate includes: Multiple groups of root and fruit plants, after being fully turned over, were laid out within the sampling range of the image acquisition device. In this scheme, 200 groups of root and fruit plants were used to calculate the baseline fruit rate. For each group of root and fruit plants, the number of exposed root fruits and the total number of root fruits within the sampling range are counted. The number of exposed root fruits can be obtained by computer software. If the detection model has a statistical counting function, the statistical counting can be performed based on the detection model. The total number of root fruits is obtained manually by manually turning over the roots of each root and fruit plant to count the total number of root fruits. The total number includes root fruits that were not visible in the image originally captured by the image acquisition device. The average exposure amount and the average total number of roots and fruits were calculated based on the number of exposed roots and fruits for all root and fruit plants and the total number of roots and fruits. The baseline effectiveness rate is calculated based on the average exposure and the average total number of samples. Specifically, the baseline effectiveness rate is obtained by multiplying the ratio of the average exposure to the average total number of samples by 100%.

[0009] Furthermore, since the root and fruit plants are planted uniformly, and the planting uniformity of root and fruit plants in the same field is relatively stable, the average total number of root and fruit plants in all the above groups can be used as the total number of root and fruit plants in the sampling fruit rate detection, so as to facilitate obtaining real-time fruit rate data.

[0010] Furthermore, adjusting the rotation speed of the rice-turning tray includes: The rotational speed of the first hydraulic motor that drives the rice turning disc is adjusted by the first proportional solenoid valve; The method further includes: The rotational speed of the second hydraulic motor driving the lifting chain is adjusted by a second proportional solenoid valve, so that the ratio between the throwing speed of the rice turning tray and the conveying speed of the lifting chain meets a preset first ratio. In this scheme, the value of the first ratio is set to 0.5~0.7.

[0011] In addition, while adjusting the speed of the second hydraulic motor, it is necessary to consider the ratio between the tractor speed and the conveying speed of the lifting chain. This ratio needs to be controlled within the range of 0.8 to 1.2 to ensure that the dug-up root and fruit plants are promptly conveyed backward, preventing congestion and accumulation in the dug-up section, which could lead to fruit drop and loss. By ensuring that the ratio between the throwing speed of the turning tray and the conveying speed of the lifting chain meets a first ratio, it can be guaranteed that the crops conveyed by the lifting chain can be turned and laid out in a timely manner, avoiding congestion and fruit drop losses.

[0012] A root and fruit turning and rice transplanting operation control system based on online identification of fruit yield is applied to a root and fruit digging and turning rice transplanting machine. The machine consists of a tractor, a digging mechanism, a lifting chain, and a turning mechanism. The turning mechanism includes turning discs and guide rods arranged front and rear. The turning discs are driven by a first hydraulic motor, whose hydraulic oil flow is controlled by a first proportional solenoid valve. The lifting chain is driven by a second hydraulic motor, whose hydraulic oil flow is controlled by a second proportional solenoid valve. Two sets of turning discs are arranged in a V-shape, with multiple discs in each set fixed on the same turning shaft. Guide rods extend between adjacent turning discs. Each turning shaft is driven by a separate first hydraulic motor. The root and fruit turning and rice transplanting operation control system includes: An image acquisition device is installed above the tail of the root and fruit digging and turning rice seedling machine; The control system, connected to the image acquisition device, can change the rotation speed of the turning tray and the rotation speed of the lifting chain, and is used to implement the above-mentioned root and fruit turning operation control method based on online fruit recognition. Specifically, the control system controls the hydraulic oil flow of the first hydraulic motor and the second hydraulic motor respectively through the first proportional solenoid valve and the second proportional solenoid valve.

[0013] Beneficial Effects: The root and fruit turning operation control method and system based on online fruit recognition rate of the present invention has the following beneficial effects: (1) This invention acquires root and fruit images in real time using an image acquisition device, calculates the sampling fruit exposure rate using a pre-trained detection model, and automatically adjusts the rotation speed of the turning tray after comparing it with the benchmark fruit exposure rate. This method realizes online monitoring and intelligent control of the key indicator of fruit exposure rate, significantly improves the automation level of root and fruit harvesting, reduces the need for manual intervention, and improves operational efficiency and consistency. Through precise control, the fruit exposure rate can be improved to ensure the quality of drying and guarantee the quality of the harvested root and fruit.

[0014] (2) The control program achieves precise management of rice seedling turning operations through a multi-level feedback mechanism. When an anomaly occurs for the first time, the speed of the rice seedling turning tray is automatically adjusted to improve the quality of the operation; if the anomaly continues, the fault is identified and an alarm is triggered to avoid further losses. This graded response improves the continuity and efficiency of the operation, can detect equipment problems in a timely manner, reduce economic losses caused by faults, and enhance the intelligence of the system and the user experience.

[0015] (3) In the calculation method of the benchmark fruit rate, the total number of roots and fruits is manually counted, the exposure is counted by the detection model, and the average exposure and the average total number are calculated based on multiple sets of samples to obtain the benchmark fruit rate. This method ensures the objectivity, comprehensiveness and accuracy of the benchmark value, reduces random errors and deviations, provides a reliable basis for the comparison of real-time fruit rates, and thus improves the credibility and stability of the entire control system. At the same time, the average total number is used as a constant to simplify the real-time detection process, reduce the computational complexity and improve the work efficiency. Attached Figure Description

[0016] Figure 1 Side view of the structure of the root and fruit digging and turning machine; Figure 2 A 3D structural diagram of a root and fruit digging and turning machine; Figure 3 This is a flowchart illustrating the control method for root and fruit turning operations based on online identification of fruit yield.

[0017] In the diagram: 1-Tractor; 2-Digging mechanism; 3-Lifting rod chain; 4-Turning rice seedling mechanism; 41-Turning rice seedling tray; 42-Guide rod; 51-First hydraulic motor; 52-Second hydraulic motor; 53-First proportional solenoid valve; 54-Second proportional solenoid valve; 6-Image acquisition device. Detailed Implementation

[0018] The invention will now be further described with reference to the accompanying drawings.

[0019] The root and fruit turning operation control method based on online fruit recognition rate of the present invention is implemented by a root and fruit turning operation control system and is applied to, for example... Figure 1The root and fruit digging and turning rice seedling machine shown consists of a tractor 1, a digging mechanism 2, a lifting rod chain 3, and a turning mechanism 4. The turning mechanism 4 includes a turning disc 41 arranged in a front-to-back configuration and guide rods 42. The turning disc 41 is driven by a first hydraulic motor 51, and the hydraulic oil flow of the first hydraulic motor 51 is controlled by a first proportional solenoid valve 53. The lifting rod chain 3 is driven by a second hydraulic motor 52, and the hydraulic oil flow of the second hydraulic motor 52 is controlled by a second proportional solenoid valve 54. The two sets of turning discs 41 are arranged in a V-shape. Multiple turning discs 41 in each set are fixed on the same turning shaft. There are guide rods 42 extending between adjacent turning discs 41. Each turning shaft is driven by a separate first hydraulic motor 51.

[0020] The root and fruit turning and vine turning operation control system includes: Image acquisition device 6 is installed above the tail of the root and fruit digging and turning machine; The control system, connected to the image acquisition device, can change the rotation speed of the turning tray 41 and the rotation speed of the lifting rod chain 3, and is used to implement the above-mentioned root and fruit turning operation control method based on online fruit recognition. Specifically, the control system controls the hydraulic oil flow of the first hydraulic motor 51 and the second hydraulic motor 52 through the first proportional solenoid valve 53 and the second proportional solenoid valve 54, respectively.

[0021] Based on this, such as Figure 2 The method for regulating root and fruit turning operations based on online identification of fruit yield, as shown, includes the following steps S101-S103: Step S101: The image acquisition device 6 installed above the tail of the root and fruit digging and turning machine is used to acquire an image of the root and fruit after turning and laying the seedlings every preset time (e.g., 20 seconds). Step S102: Detect the sampling fruit exposure rate of the root and fruit image based on a pre-trained detection model; the fruit exposure rate refers to the proportion of the number of roots and fruits exposed within the sampling range of the image acquisition device 6 after the seedlings are turned and laid out, relative to the total number of roots and fruits. The formula for calculating the clear effect rate is as follows:

[0022] in, For the sake of clear results; The number of roots and fruits exposed after being turned over and laid out within the sampling range of image acquisition device 6, expressed in units of individual plants. The total number of roots and fruits after turning and laying the seedlings within the sampling range of image acquisition device 6, in units of individuals.

[0023] Step S103: Compare the sampled fruit rate obtained by online identification calculation with the benchmark fruit rate. If the sampled fruit rate is higher than the benchmark fruit rate, no adjustment is required. If the sampled fruit rate is lower than the benchmark fruit rate, start the adjustment program to adjust the rotation speed of the rice turning tray 41. The benchmark fruit rate is calculated in advance.

[0024] This invention uses an image acquisition device 6 to acquire root and fruit images in real time, and calculates the sampling fruit transparency rate using a pre-trained detection model. After comparing the result with a benchmark fruit transparency rate, the rotation speed of the turning tray 41 is automatically adjusted. This method achieves online monitoring and intelligent control of the key indicator of fruit transparency rate, significantly improving the automation level of root and fruit harvesting, reducing the need for manual intervention, and increasing operational efficiency and consistency. Precise control can improve the fruit transparency rate, ensure the quality of drying, and guarantee the quality of the harvested root and fruit.

[0025] Preferably, the detection model is a lightweight fruit exposure rate detection model based on deep learning YOLOv5s. This detection model is pre-trained and can efficiently and accurately identify the number of exposed fruits and the total number of fruits in the root-fruit image. The model requires minimal computational resources, ensuring real-time performance and practicality. Through deep learning, the model adapts to various environmental and lighting conditions, improving the accuracy and robustness of fruit exposure rate detection.

[0026] Preferably, the control procedure in step S103 includes steps S201-S202: Step S201: Determine whether the fruit yield rate is lower than the baseline fruit yield rate for two consecutive samplings. If so, adjust the hydraulic motor of the rice seedling turning and laying mechanism to reduce the rotation speed of the rice seedling turning tray 41 by 5%. Step S202: Determine whether the sampling fruit rate is lower than the benchmark fruit rate after slowing down the rice turning tray 41 twice in a row, that is, whether the sampling fruit rate obtained after four consecutive detections is lower than the benchmark fruit rate. If so, activate the audible and visual alarm to remind the driver to stop the operation and check.

[0027] The control program achieves precise management of rice seedling turning operations through a multi-level feedback mechanism. Upon the first anomaly, the speed of the turning tray 41 is automatically fine-tuned to improve operational quality; continued anomalies trigger a fault diagnosis and alarm to prevent further losses. This tiered response improves operational continuity and efficiency, promptly identifies equipment problems, reduces economic losses due to malfunctions, and enhances system intelligence and user experience.

[0028] Preferably, the method for calculating the benchmark clear rate includes the following steps S301-S304: Step S301: Lay out multiple groups of root and fruit plants after thorough turning within the sampling range of the image acquisition device 6. In this embodiment, 200 groups of root and fruit plants were used for the calculation of the baseline fruit rate. Step S302: For each group of root and fruit plants, count the number of exposed roots and fruits and the total number of roots and fruits within the sampling range. The number of exposed roots and fruits can be obtained by computer software. If the detection model has a statistical counting function, the statistical counting can be performed based on the detection model. The total number of roots and fruits is obtained manually. The roots of each root and fruit plant are manually turned over to count the total number of roots and fruits. The total number includes roots and fruits that were not visible in the image originally captured by the image acquisition device 6. Step S303: Calculate the average exposure amount and the average total number of roots and fruits based on the number of exposed roots and fruits corresponding to all root and fruit plants and the total number of roots and fruits. Step S304: Calculate the baseline effectiveness rate based on the average exposure and the average total number of samples. That is, multiply the ratio of the average exposure to the average total number of samples by 100% to obtain the baseline effectiveness rate.

[0029] Furthermore, since the root and fruit plants are planted uniformly, and the planting uniformity of root and fruit plants in the same field is relatively stable, the average total number of root and fruit plants in all the above groups can be used as the total number of root and fruit plants in the sampling fruit rate detection, so as to facilitate obtaining real-time fruit rate data.

[0030] The method for calculating the baseline fruit detection rate involves manually counting the total number of roots and fruits, combining this with the exposure count from the detection model, and then calculating the average exposure and the average total number based on multiple sample groups to arrive at the baseline fruit detection rate. This method ensures the objectivity, comprehensiveness, and accuracy of the baseline value, reduces random errors and biases, and provides a reliable basis for comparing real-time fruit detection rates, thereby improving the credibility and stability of the entire control system. At the same time, using the average total number as a constant simplifies the real-time detection process, reduces computational complexity, and improves operational efficiency.

[0031] Preferably, adjusting the rotation speed of the rice-turning tray 41 in step S103 above includes: The rotational speed of the first hydraulic motor 51, which drives the rice turning disc 41, is adjusted by the first proportional solenoid valve 53. The method further includes: The rotational speed of the second hydraulic motor 52, which drives the lifting rod chain 3, is adjusted by the second proportional solenoid valve 54 so that the ratio of the throwing speed of the rice turning tray 41 to the conveying speed of the lifting rod chain 3 satisfies a preset first ratio. In this embodiment, the value of the first ratio is 1-1.4. Since the diameter of the rice turning tray 41 is twice the diameter of the corresponding drive sprocket of the lifting rod chain 3, the ratio of the first hydraulic motor 51 to the second hydraulic motor 52 is set to 0.5-0.7.

[0032] In addition, while adjusting the speed of the second hydraulic motor 52, it is necessary to consider the ratio between the speed of the tractor 1 and the conveying speed of the lifting chain 3. This ratio needs to be controlled within the range of 0.8 to 1.2 to ensure that the dug-up root and fruit plants are promptly conveyed backward, preventing congestion and accumulation in the dug-up section, which could lead to fruit drop and loss. By ensuring that the ratio between the throwing speed of the turning tray 41 and the conveying speed of the lifting chain 3 meets a first ratio, it can be guaranteed that the crops conveyed by the lifting chain 3 can be turned and laid out in a timely manner, avoiding congestion and fruit drop losses.

[0033] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A method for regulating root and fruit turning operations based on online identification of fruit yield, characterized in that, The methods include: The image acquisition device (6) acquires images of the roots and fruits after the seedlings have been turned over and laid out at preset intervals. The sampling fruit exposure rate of the root and fruit image is detected based on the pre-trained detection model; the fruit exposure rate refers to the proportion of the number of roots and fruits exposed outside the sampling range of the image acquisition device (6) after the seedlings are turned over and laid out, to the total number of roots and fruits. The sampling fruit rate obtained by online identification calculation is compared with the benchmark fruit rate. If the sampling fruit rate is higher than the benchmark fruit rate, no adjustment is required. If the sampling fruit rate is lower than the benchmark fruit rate, the adjustment program is started to adjust the rotation speed of the rice turning tray (41). The benchmark fruit rate is calculated in advance.

2. The root-fruit turning and rice-planting operation control method based on online fruit identification rate according to claim 1, characterized in that, The detection model is a lightweight detection model for the accuracy rate based on deep learning YOLOv5s.

3. The method for regulating root and fruit turning operations based on online identification of fruit yield according to claim 1, characterized in that, The control procedure includes: If the sampling rate is lower than the baseline rate for two consecutive samplings, adjust the hydraulic motor of the rice seedling turning and laying mechanism to reduce the rotation speed of the rice seedling turning tray (41) by 5%. If the sampling rate of the seedling turning tray (41) is lower than the benchmark rate after two consecutive decelerations, an audible and visual alarm will be activated to remind the driver to stop the operation and check.

4. The root-fruit turning and rice-planting operation control method based on online fruit identification rate according to claim 1, characterized in that, The method for calculating the benchmark clear rate includes: Multiple groups of root and fruit plants that have been thoroughly turned over are laid out within the sampling range of the image acquisition device (6); For each group of root and fruit plants, count the number of exposed root fruits and the total number of root fruits within the sampling range. The average exposure amount and the average total number of roots and fruits were calculated based on the number of exposed roots and fruits for all root and fruit plants and the total number of roots and fruits. The baseline efficacy rate is calculated based on the average exposure and the average total number.

5. The root-fruit turning and rice-planting operation control method based on online fruit identification rate according to claim 1, characterized in that, The adjustment of the rotation speed of the rice turning tray (41) includes: The rotational speed of the first hydraulic motor (51) that drives the rice turning disc (41) to rotate is adjusted by the first proportional solenoid valve (53); The method further includes: The rotational speed of the second hydraulic motor (52) that drives the lifting rod chain (3) is adjusted by the second proportional solenoid valve (54) so ​​that the ratio between the throwing speed of the rice turning tray (41) and the conveying speed of the lifting rod chain (3) satisfies the preset first ratio.

6. A root and fruit turning and rice turning operation control system based on online identification of fruit yield, applied to a root and fruit digging and turning rice turning machine, the root and fruit digging and turning rice turning machine is composed of a tractor (1), a digging mechanism (2), a lifting rod chain (3) and a turning mechanism (4), the turning mechanism (4) including a turning disc (41) arranged in front and behind and a guide rod (42); characterized in that, The root and fruit turning and vine turning operation control system includes: Image acquisition device (6) is installed above the tail of the root and fruit digging and turning machine; The control system, which is connected to the image acquisition device, is capable of changing the rotation speed of the rice turning tray (41) and is used to implement the root and fruit turning operation control method based on online identification of fruit rate as described in any one of claims 1-5.

Images