Automatic sorting system and sorting method for yellow leaves and green leaves of sugarcane tails

Abstract

The invention discloses an automatic sorting system and method for yellow leaves and green leaves of sugarcane tails. The automatic sorting system comprises a feeding conveying device, an image collecting device, a visual processing unit and a sorting execution device which are sequentially arranged. The system obtains a sugarcane leaf material image through an image acquisition device, and a visual processing unit comprehensively analyzes color, shape and texture features based on a multi-feature fusion recognition algorithm, accurately distinguishes sugarcane tips, green leaves and yellow leaves, and generates a sorting instruction; the sorting execution device sorts different types of materials to corresponding discharging channels at the accurate moment through the angle and air pressure adjustable high-pressure air flow nozzle array according to the instruction, and finally the materials enter the independent smashing and packaging unit. According to the sugarcane leaf sorting device, high-precision and high-efficiency automatic sorting of sugarcane leaves is achieved, the problem of value loss caused by mixing treatment is effectively solved, and the economic value and the resource utilization level of sugarcane by-products are remarkably improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of agricultural machinery, and relates to an automatic sugarcane tail yellow leaf and green leaf sorting system and method. BACKGROUND

[0002] In the process of sugarcane harvesting and cutting, leaf stripping, especially step-by-step leaf stripping, will produce a large amount of sugarcane leaves (including cane tops, green leaves and yellow leaves). The economic values of these by-products are significantly different: green leaves are rich in nutrients and suitable for high-quality feed; yellow leaves are high in cellulose content and suitable for fuel or base material; and cane tops may have other uses such as silage, pigment extraction, polyphenol extraction, etc. Currently, these materials are usually mixed together for crushing and packaging or even discarded, resulting in mixed quality of the final product and failure to achieve "high quality and high price", causing waste of resources and economic loss.

[0003] Realizing sorting is the key to value enhancement. However, manual sorting is extremely inefficient, costly and cannot guarantee accuracy. Existing sorting equipment based on simple color sensors cannot cope with the irregular shape of sugarcane leaves, the gradual change of color (such as yellow-green transition), and the complex changes of color and texture under different varieties and humidity conditions, and the sorting effect is poor. Therefore, there is an urgent need for an intelligent solution that can adapt to complex working conditions and realize high-precision and automatic sorting of cane tops, green leaves and yellow leaves. SUMMARY

[0004] The present application aims to provide an automatic sugarcane tail yellow leaf and green leaf sorting system and method, which has high sorting accuracy, strong adaptability and can significantly improve the economic value of sugarcane leaves.

[0005] According to the purpose of the present application, the present application provides an automatic sugarcane tail yellow leaf and green leaf sorting system, comprising: a feeding conveying device for uniformly conveying mixed sugarcane leaf materials; an image acquisition device arranged above the conveying path of the feeding conveying device for acquiring image information of the sugarcane leaf materials; a visual processing unit in communication connection with the image acquisition device for processing the image information and distinguishing cane tops, green leaves and yellow leaves in the materials based on a multi-feature fusion recognition algorithm, and further generating sorting instructions; a sorting execution device in communication connection with the visual processing unit, receiving the sorting instructions and sorting materials of different categories to different discharge channels.

[0006] Further, the visual processing unit comprises a feature extraction module and a classification and recognition module; the feature extraction module is used to extract color features, shape features and texture features from images; and the classification and recognition module adopts a target detection or image classification model based on deep learning.

[0007] Further, the visual processing unit further comprises a self-learning module for collecting sample images and labels of samples with recognition confidence lower than a threshold or sorting errors confirmed by manual review in the sorting process, and using the data to incrementally train and optimize the recognition model.

[0008] Further, the sorting execution device comprises a high-pressure gas source, a main gas path connected with the high-pressure gas source, and a nozzle array arranged at the end of the feed conveying device; the nozzle array comprises a plurality of independently controlled electromagnetic valve nozzles corresponding to the discharge channels of the cane tops, green leaves and yellow leaves respectively.

[0009] Further, the spray angle of the nozzle is adjustable relative to the conveying plane; and / or, the system further comprises a gas pressure adjusting device for dynamically adjusting the output gas pressure of the high-pressure gas source according to the real-time flow and category of the material.

[0010] Further, the end of the discharge channel is connected with at least two independent subsequent processing units, the subsequent processing units are crushing and packaging machines, and the crushing fineness of different subsequent processing units can be independently set.

[0011] According to another object of the present application, the present application provides a sorting method of the above-mentioned automatic sorting system for cane tops and green leaves, comprising the following steps: S1: conveying the mixed cane leaf material to the image acquisition area through the feed conveying device; S2: acquiring the image of the material through the image acquisition device; S3: the visual processing unit processes the image, and classifies the material into cane tops, green leaves and yellow leaves by fusing color, shape and texture features; S4: the sorting execution device sorts the materials of different categories into corresponding discharge channels according to the recognition result.

[0012] Further, in step S3, a self-learning step is further included: when the recognition confidence of the model for the current material is lower than a preset threshold, the material image and the correct label of subsequent manual review are stored in the training database, and the model is retrained periodically.

[0013] Further, in step S4, the system calculates the required time for the material to move to the sorting point according to the position coordinates of the material in the image and the conveying belt speed, and triggers the sorting action at the accurate time; and / or, the system dynamically adjusts the working parameters of the sorting execution device according to the real-time material flow.

[0014] The beneficial effects of the present application are: This invention integrates modules for feeding and conveying, image acquisition, vision processing, and sorting execution to construct a complete automated sorting system. This system completely replaces the inefficient, high-cost, and inaccurate manual sorting with automated operations, resulting in an order-of-magnitude improvement in sorting efficiency. The core vision processing unit employs a multi-feature fusion intelligent recognition algorithm, capable of accurately distinguishing between sugarcane tops, green leaves, and yellow leaves that appear similar but have vastly different values. The sorting accuracy far surpasses human judgment and traditional photoelectric sorting equipment, ensuring the high purity of the final product. This high-precision sorting directly realizes the optimal use of the three types of materials, ensuring high quality and high price, transforming low-value byproducts resulting from mixed processing into high-value classified commodities, greatly improving resource utilization and economic benefits, and providing reliable equipment support for cost reduction, efficiency improvement, and intelligent upgrading in the sugarcane industry. Attached Figure Description

[0015] Figure 1 This is a schematic diagram of the structural layout of the system according to an embodiment of the present invention; Figure 2 This is a flowchart illustrating the sorting method according to an embodiment of the present invention.

[0016] In the diagram: 100, feeding conveyor; 200, image acquisition device; 300, vision processing unit; 400, sorting execution device; 410, first channel; 420, second channel; 430, third channel; 500, post-processing unit; 510, first crushing and baling machine; 520, second crushing and baling machine; 530, third crushing and baling machine. Detailed Implementation

[0017] The specific embodiments of the present invention will be further described below. It should be noted that these descriptions are for the purpose of aiding understanding the present invention, but do not constitute a limitation thereof. Furthermore, the technical features involved in the various embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0018] Example 1 like Figure 1 As shown, an automated sorting system for yellow and green sugarcane leaves includes: The feeding conveyor 100 is used to uniformly convey the mixed sugarcane leaf material. The image acquisition device 200 is installed above the conveying path of the feeding conveyor 100 and is used to acquire image information of sugarcane leaf material. The vision processing unit 300 is communicatively connected to the image acquisition device 200. It is used to process the image information and distinguish sugarcane tops, green leaves and yellow leaves in the material based on a multi-feature fusion recognition algorithm, thereby generating sorting instructions. The sorting execution device 400 is in communication connection with the visual processing unit 300, receives the sorting instruction, and sorts the materials of different categories into different discharge channels; At least two independent subsequent processing units 500 are respectively connected to the discharge channels for separately processing the sorted materials of different categories.

[0019] Specifically, the feeding conveying device 100 is a belt conveyor, the conveying speed of which is adjustable, and the surface of the conveying belt is provided with a material uniformizing mechanism for single-layer paving of the materials.

[0020] The image acquisition device 200 includes a high-speed industrial camera, the type of which is CCD or CMOS, and is equipped with a high-frequency flash or a constant-illumination LED light source to eliminate motion blur.

[0021] The visual processing unit 300 includes a pre-processing module, a feature extraction module, and a classification and recognition module; the feature extraction module is used to extract color features, shape features, and texture features from the image. The classification and recognition module adopts a target detection or image classification model based on deep learning, which is trained by a data set containing images of sugarcane shoots, green leaves, and yellow leaves collected under different varieties and different weather conditions.

[0022] The visual processing unit 300 further includes a self-learning module for collecting sample images and labels of samples with recognition confidence lower than a threshold value or sorting errors confirmed by manual review in the sorting process, and using these data to incrementally train and optimize the recognition model.

[0023] The sorting execution device 400 includes a high-pressure gas source, a main gas path connected to the high-pressure gas source, and a nozzle array arranged at the end of the feeding conveying device 100; the nozzle array includes a plurality of independently controlled electromagnetic valve nozzles corresponding to the discharge channels of the sugarcane shoots, green leaves, and yellow leaves. The jet angle of the nozzles is adjustable relative to the conveying plane to adapt to the requirements of the trajectories of different materials. The system further includes a gas pressure adjusting device for dynamically adjusting the output gas pressure of the high-pressure gas source according to the real-time flow and category of the materials.

[0024] The sorting instruction generated by the visual processing unit 300 includes the category of the target material and its position coordinates and movement speed on the conveying belt; the system calculates the time delay according to the position coordinates and movement speed, and triggers the action of the electromagnetic valve nozzle of the corresponding category at the accurate time.

[0025] Specifically, the discharge channels include a first channel 410 for collecting sugarcane shoots, a second channel 420 for collecting green leaves, and a third channel 430 for collecting yellow leaves.

[0026] The subsequent processing unit 500 includes a first pulverizing and packaging machine 510 connected with the first channel 410, a second pulverizing and packaging machine 520 connected with the second channel 420, and a third pulverizing and packaging machine 530 connected with the third channel 430. The pulverizing fineness of the first, second, and third pulverizing and packaging machines 510, 520, and 530 can be independently set according to the characteristics of the materials processed thereby.

[0027] As shown in Figure 2 The sorting method of the automatic sorting system for sugarcane tail yellow leaves and green leaves includes the following steps: S1: uniformly and single-layeredly conveying the mixed sugarcane leaf materials to an image acquisition area through a feeding conveying device 100; S2: acquiring images of the materials on the conveying belt in real time through an image acquisition device 200; S3: processing the images by a visual processing unit 300, extracting color, shape, and texture features, and classifying the materials into cane tops, green leaves, and yellow leaves by using a recognition model; S4: generating sorting instructions with delay calculation by the visual processing unit 300 according to the recognition results and the positions of the materials; S5: executing actions at accurate time by a sorting execution device 400 to sort the materials of different categories into corresponding discharge channels; S6: conveying the sorted cane tops, green leaves, and yellow leaves into independent subsequent processing units 500 for pulverization and packaging, respectively.

[0028] In step S3, the recognition model is a convolutional neural network model which comprehensively judges by fusing HSV components in the color space, contour moments in the shape features, and LBP features in the texture features.

[0029] In step S3, a self-learning step is further included: when the recognition confidence of the model for the current material is lower than a preset threshold, the material image and the correct label after subsequent manual review are stored in a training database, and the model is retrained periodically using the updated training database.

[0030] In step S5, the required time T for the material to move from the shooting point to the center point of the nozzle array is calculated according to the position coordinates (x, y) of the material in the image and the conveying belt speed V, and the corresponding nozzle is triggered after T time.

[0031] In step S5, the system monitors the throughput of materials of different categories per unit time in real time, and dynamically adjusts the air pressure of the high-pressure gas source according to the flow rate, i.e., the air pressure is increased when the flow rate is high and the air pressure is decreased when the flow rate is low.

[0032] Example 2 As shown in Figure 1As shown in the figure, this embodiment of an automated sorting system for yellow and green sugarcane leaves includes a feeding conveyor 100, an image acquisition device 200, a vision processing unit 300, a sorting execution device 400, and at least two independent post-processing units 500; wherein: The feeding conveyor 100 adopts a variable frequency speed control belt conveyor, and the belt surface is equipped with scrapers to ensure that the sugarcane leaf material passes through evenly and in a single layer.

[0033] The image acquisition device 200 uses a 2-megapixel CMOS high-speed industrial camera, equipped with a ring LED high-frequency flash, and is installed directly above the belt to capture high-definition, motion-free material images.

[0034] The vision processing unit 300 is a high-performance industrial computer (ICC). Its internal software first preprocesses the images (e.g., Gaussian filtering, contrast stretching). Then, a pre-trained YOLOv5 model is used for object detection and classification. During training, the input data for this model not only includes RGB images but also converts them to the HSV color space and calculates LBP texture feature maps. These features are used together as network input to improve the model's ability to distinguish between unsaturated yellow leaves and uniquely shaped sugarcane shoots. A self-learning service is also deployed within the ICC, which automatically records all images with a recognition confidence score below 0.85 and awaits operator confirmation via a UI during quality inspection. Confirmed data is automatically added to the training set, and incremental training of the model is initiated weekly.

[0035] The sorting actuator 400 includes an air compressor (providing an adjustable air pressure of 0.4-0.8 MPa), a set of solenoid valves, and three nozzles. The three nozzles are respectively aligned with the collection channels for sugarcane tops, green leaves, and yellow leaves (410, 420, 430). The nozzles are mounted via universal joints, and their pitch angle is adjustable from 15° to 60°. The industrial control computer calculates the precise time it takes for the material to reach directly below the nozzle based on the identified material type and its position on the conveyor belt, and controls the corresponding solenoid valve to open 10-30 ms.

[0036] The subsequent processing unit 500 includes three independent crushing and baling units (510, 520, 530). For example, the second crushing and baling unit 520 for green leaves can be set to a finer particle size to facilitate feed production; while the third crushing and baling unit 530 for yellow leaves can be set to a coarser particle size to meet combustion requirements.

[0037] like Figure 2 As shown, the workflow of this system is as follows: After the material is loaded and image is collected, the vision system extracts multiple features and intelligently identifies them as "cane top", "green leaf" or "yellow leaf", and then controls the corresponding nozzle to spray air at the precise moment according to the motion tracking results, completing the sorting. The sorted materials enter independent crushing and packaging processes.

[0038] The present application solves the problem of fine sorting of sugarcane leaves by combining software and hardware, and provides effective technical equipment for improving the added value of agricultural products.

[0039] The present application can effectively distinguish between cane tops, green leaves and yellow leaves with similar appearances by fusing multi-dimensional features such as color, shape and texture, and combining a powerful deep learning model. The sorting accuracy is much higher than traditional methods. The unique self-learning function of the present application enables the system to adapt to changes in crop varieties and environment, ensuring long-term stability and accuracy.

[0040] The adjustable air flow nozzle system of the present application ensures the precision and efficiency of the sorting action, and can reduce physical damage to fragile materials (such as green leaves) through parameter optimization. The present application realizes high-purity separation of the three types of materials, so that each type of material can be utilized according to its highest value channel, significantly improving economic efficiency. The present application is fully automated, greatly reducing labor costs and labor intensity, and embodies the technical development direction of intelligent agriculture.

[0041] Although embodiments of the present application have been shown and described, it will be understood by those having ordinary skill in the art that various changes, modifications, substitutions and alterations can be made without departing from the principles and spirit of the present application, and the scope of the present application is defined by the appended claims and their equivalents.

Claims

1. An automatic sugarcane yellow leaf and green leaf sorting system, characterized by, The system comprises: a feeding conveyor for uniformly conveying the mixed sugarcane leaf material; an image acquisition device arranged above the conveying path of the feeding conveyor for acquiring image information of the sugarcane leaf material; a visual processing unit in communication with the image acquisition device for processing the image information and distinguishing sugarcane tops, green leaves and yellow leaves in the material based on a multi-feature fusion recognition algorithm, and then generating a sorting instruction; a sorting execution device in communication with the visual processing unit for receiving the sorting instruction and sorting the materials of different categories into different discharge channels.

2. The sugarcane automatic pale and green leaves sorting system according to claim 1, characterized by, The visual processing unit comprises a feature extraction module and a classification recognition module; the feature extraction module is used to extract color features, shape features and texture features from the image; and the classification recognition module adopts a target detection or image classification model based on deep learning.

3. The sugarcane automatic pale and green leaves sorting system according to claim 2, wherein The visual processing unit further comprises a self-learning module for collecting sample images and labels of the materials whose recognition confidence is lower than a threshold or whose sorting error is confirmed by manual review during the sorting process, and using these data to incrementally train and optimize the recognition model.

4. The sugarcane automatic pale and green leaves sorting system according to claim 1, characterized by the fact that The sorting execution device comprises a high-pressure gas source, a main gas path connected with the high-pressure gas source, and a nozzle array arranged at the end of the feeding conveyor; the nozzle array comprises a plurality of independently controlled electromagnetic valve nozzles corresponding to the discharge channels of the sugarcane tops, green leaves and yellow leaves, respectively.

5. The sugarcane automatic pale and green leaves sorting system according to claim 4, wherein, The spray angle of the nozzle is adjustable relative to the conveying plane; and / or, the system further comprises a gas pressure adjusting device for dynamically adjusting the output gas pressure of the high-pressure gas source according to the real-time flow and category of the material.

6. The sugarcane automatic pale and green leaves sorting system according to claim 1, wherein The end of the discharge channel is connected with at least two independent subsequent processing units, the subsequent processing units are crushing packers, and the crushing fineness of different subsequent processing units can be independently set.

7. The sorting method of the automated sorting system for yellow and green sugarcane leaves according to any one of claims 1-6, characterized in that, The system comprises the following steps: S1: conveying the mixed sugarcane leaf material to the image acquisition area through the feeding conveyor; S2: acquiring the image of the material through the image acquisition device; S3: the visual processing unit processes the image, and classifies the material into sugarcane tops, green leaves and yellow leaves by fusing color, shape and texture features; S4: the sorting execution device sorts the materials of different categories into corresponding discharge channels according to the recognition result.

8. The sorting method of claim 7, wherein, In step S3, a self-learning step is further included: when the recognition confidence of the model for the current material is lower than a preset threshold, the material image and the correct label after manual review are stored in a training database, and the model is retrained periodically.

9. The sorting method of claim 7, wherein, In step S4, the system calculates the required time for the material to move to the sorting point according to its position coordinates in the image and the conveying belt speed, and triggers the sorting action at the accurate time; and / or, the system dynamically adjusts the working parameters of the sorting execution device according to the real-time material flow.

Images