An intelligent detection device and method for the trash content of graded machine-harvested sugarcane

Abstract

The present invention discloses an intelligent detection device and method for the trash content of graded machine-harvested sugarcane, the device comprises a sugarcane segment conveying mechanism, a scraping mechanism and a visual detection mechanism; the sugarcane segment conveying mechanism comprises a conveyor belt and a conveying drive mechanism; the scraping mechanism comprises a scraping installation plate and a scraping roller, the scraping installation plate is arranged above the conveyor belt, the scraping roller is rotatably connected to the bottom of the scraping installation plate; a limiting channel for the passage of sugarcane segments is formed between the scraping roller and the lowered conveyor belt; the visual detection mechanism comprises a detection box, a camera and an image processing module, the camera is arranged in the detection box and located above the conveyor belt, and the camera is electrically connected to the image processing module. The present invention can automatically detect the trash content of sugarcane segments, and has the advantages of high working efficiency and high detection accuracy.

Description

Technical Field

[0001] The present invention relates to a device and method for detecting the trash content of sugarcane, and in particular to an intelligent device and method for detecting the trash content of graded machine-harvested sugarcane. Background Art

[0002] Sugarcane is a perennial grass plant that grows mainly in tropical and subtropical regions. It is an important economic pillar industry in many countries and regions. Sugarcane sugar extraction is the process of using sugarcane as raw material to make white sugar, raw sugar and other products through processes such as juice extraction, purification, evaporation, crystallization, honey separation and drying. The characteristics of sugarcane, the chemical composition and properties of sugarcane juice have a great influence on the sugar-making process and are the basis for selecting production methods and process conditions. Among them, before the formal sugar extraction work begins, the impurity content of the sugarcane segment must be measured to determine whether it meets the quality standards.

[0003] The existing method of detecting the trash content of sugarcane segments is mainly manual sampling, which involves manually removing impurities and weighing to determine the trash content. Although this method can measure the trash content, it will result in a high false positive rate if the trash content is judged manually, and the test results are not very reliable. In addition, manual testing is slow and inefficient. Summary of the invention

[0004] The purpose of the present invention is to overcome the above-mentioned problems and provide an intelligent detection device and method for the trash content of graded mechanically harvested sugarcane. The intelligent detection device and method can automatically detect the trash content of sugarcane segments and have the advantages of high working efficiency and high detection accuracy.

[0005] The purpose of the present invention is achieved through the following technical solutions:

[0006] An intelligent detection device for trash content of graded mechanically harvested sugarcane, comprising a sugarcane segment conveying mechanism, a scraping mechanism and a visual detection mechanism;

[0007] The sugarcane segment conveying mechanism comprises a conveyor belt and a conveying drive mechanism;

[0008] The scraping mechanism comprises a scraping installation plate and a scraping roller, wherein the scraping installation plate is arranged above the conveyor belt, and the scraping roller is rotatably connected to the bottom of the scraping installation plate; a limiting channel for the sugarcane segment to pass through is formed between the scraping roller and the conveyor belt lowered;

[0009] The visual inspection mechanism includes an inspection box, a camera and an image processing module. The camera is arranged in the inspection box and located above the conveyor belt. The camera is electrically connected to the image processing module.

[0010] The working principle of the above-mentioned intelligent detection device for the trash content of graded machine-harvested sugarcane is as follows:

[0011] During operation, the sugarcane segments received by the machine are poured onto the conveyor belt, and the conveyor belt is driven by the conveying drive mechanism to transport the sugarcane segments downward; when the sugarcane segments pass under the scraping roller, the scraping roller will scrape off the sugarcane segments that are stacked too high and spread the sugarcane segments flat to improve the accuracy and precision of visual inspection.

[0012] The scraped and leveled sugarcane segments are conveyed downward to the bottom of the camera, and the camera acquires the image on the conveyor belt and transmits the image to the image processing module, which processes the image to calculate the impurity content of the sugarcane segments.

[0013] A preferred embodiment of the present invention, wherein the conveying drive mechanism comprises a conveying drive motor and conveying rollers, wherein two conveying rollers are provided and rotatably connected to a workbench, wherein the conveying drive motor is connected to one of the conveying rollers; and the conveyor belt is connected between the two conveying rollers.

[0014] A preferred solution of the present invention is that the scraping mounting plate is arranged above the conveyor belt through a mounting frame, and a vertical adjustment structure is arranged between the scraping mounting plate and the mounting frame, and the vertical adjustment structure includes a vertical adjustment rod, a guide rod and a guide hole, the vertical adjustment rod is arranged vertically and connected to the mounting frame through a threaded structure; the guide hole is arranged on the side of the mounting frame; one end of the guide rod is fixed to the scraping mounting plate, and the other end of the guide rod extends horizontally into the guide hole. Through the above structure, the height of the scraping roller can be adjusted to adapt to sugarcane segments of different sizes.

[0015] In a preferred embodiment of the present invention, the scraping mechanism is provided with at least two groups and arranged along the conveying direction of the conveyor belt;

[0016] Along the conveying direction, the height of the limiting channel at the rear is less than that of the limiting channel at the front. In this way, the sugarcane sections stacked too high can be scraped flat step by step, improving the passability of the sugarcane and preventing blockage.

[0017] In a preferred embodiment of the present invention, a fill light is provided in the detection box, and the fill light is located above the conveyor belt.

[0018] An intelligent detection method for trash content of graded mechanically harvested sugarcane comprises the following steps:

[0019] Pour the sugarcane segments received by the machine onto the conveyor belt, and drive the conveyor belt through the conveying drive mechanism to transport the sugarcane segments downward;

[0020] When the sugarcane segments pass under the scraping roller, the scraping roller will scrape off the sugarcane segments that are stacked too high and spread them out, thereby improving the accuracy and precision of visual inspection;

[0021] The scraped and leveled sugarcane segments are conveyed downward to the bottom of the camera, and the camera acquires the image on the conveyor belt and transmits the image to the image processing module, which processes the image and calculates the impurity content of the sugarcane segments.

[0022] In a preferred embodiment of the present invention, the specific operation of the image processing module for image processing is:

[0023] Before testing the impurity content of mechanically harvested sugarcane, collect impurities and measure their density in advance: use a high-precision electronic scale to weigh the mass M of the impurity, use the liquid displacement method to measure its volume V, calculate the density value according to the density formula, take the average value of multiple measurements, and establish a quantitative model of impurity content based on the pixel area of ​​impurities in the image;

[0024] Use Python-OpenCV image processing technology to process the image to obtain each component and obtain the pixel area of ​​different components;

[0025] Randomly select n images from the collected sugarcane images as samples, use VisualStudio software configured with OpenCV to intercept the internal areas of sugarcane and impurities in the sample images, statistically analyze the R, G, B, H, and S parameters of the pixels in the area, and obtain the color characteristic parameter value range of sugarcane and impurities;

[0026] Multiple noise reduction methods are used to reduce image noise and smooth impurity areas. The gradient vector between pixels of the machine-harvested sugarcane image is solved to obtain the gradient amplitude. The local gradient of the machine-harvested sugarcane image is scanned to find the maximum amplitude in the window as the edge point. The image edge information is extracted using the improved Canny operator. The extracted image is then superimposed with the original image to achieve edge enhancement. The closed area is traversed to identify the various components in the sugarcane image. Morphological processing is performed in combination with the geometric features of impurities to obtain the impurity segmentation result. The precision and recall rates are used to evaluate the image segmentation results and test the performance of the image segmentation algorithm.

[0027] Based on the Visual Studio environment, the deep learning framework Pytorch is used to build the network model, the open source library Open CV is used for image processing, and the PyQT5 open source library is used to write the interface program; the camera is controlled by the lower computer to collect images in real time, and the image information is encoded and transmitted to the upper computer server through the Raspberry Pi. The server first decodes the image information and performs impurity segmentation on the image. The segmented image is classified and identified by the YOLOv4 model, and data is processed based on the impurity quantification model to obtain the real-time impurity rate and average impurity rate based on quality, and finally the results are output together with the segmented image.

[0028] Compared with the prior art, the present invention has the following beneficial effects:

[0029] 1. The present invention can automatically detect the impurity content of sugarcane segments and has the advantages of high working efficiency and high detection accuracy.

[0030] 2. Use the scraping roller to scrape off the sugarcane segments that are stacked too high and spread them out flat, which can improve the accuracy and precision of visual inspection. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] Figure 1 It is a schematic diagram of the three-dimensional structure of the intelligent detection device for trash content of graded mechanically harvested sugarcane of the present invention.

[0032] Figure 2 It is a three-dimensional structural schematic diagram of the scraping mechanism of the present invention.

[0033] Figure 3 It is a front view of the visual detection mechanism of the present invention. DETAILED DESCRIPTION

[0034] In order to enable those skilled in the art to better understand the technical solution of the present invention, the present invention is further described below in conjunction with embodiments and drawings, but the embodiments of the present invention are not limited thereto.

[0035] See also Figure 1 The intelligent detection device for impurity content of graded mechanically harvested sugarcane of this embodiment includes a sugarcane segment conveying mechanism, a scraping mechanism and a visual detection mechanism; the sugarcane segment conveying mechanism includes a conveyor belt 1 and a conveying drive mechanism; wherein the conveying drive mechanism includes a conveying drive motor 2 and a conveying roller, wherein two conveying rollers are provided and rotatably connected to the workbench, and the conveying drive motor 2 is connected to one of the conveying rollers; the conveyor belt 1 is connected between the two conveying rollers.

[0036] See also Figure 1-2 The scraping mechanism includes a scraping installation plate 3 and a scraping roller 4. The scraping installation plate 3 is arranged above the conveyor belt 1, and the scraping roller 4 is rotatably connected to the bottom of the scraping installation plate 3; a limiting channel for the sugarcane segment to pass through is formed between the scraping roller 4 and the lowered conveyor belt 1;

[0037] See also Figure 1-2 The scraping installation plate 3 is arranged above the conveyor belt 1 through the mounting frame 5. A vertical adjustment structure is arranged between the scraping installation plate 3 and the mounting frame 5. The vertical adjustment structure includes a vertical adjustment rod 6, a guide rod 7 and a guide hole 5-1. The vertical adjustment rod 6 is arranged vertically and connected to the mounting frame 5 through a threaded structure; the guide hole 5-1 is arranged on the side of the mounting frame 5; one end of the guide rod 7 is fixed on the scraping installation plate 3, and the other end of the guide rod 7 extends horizontally into the guide hole 5-1. Through the above structure, the height of the scraping roller 4 can be adjusted to adapt to sugarcane segments of different sizes.

[0038] See also Figure 1-2 At least two groups of the scraping mechanisms are arranged along the conveying direction of the conveyor belt 1; along the conveying direction, the height of the limiting channel at the rear is less than the height of the limiting channel at the front. In this way, the sugarcane segments stacked too high can be scraped and leveled step by step, improving the passability of the sugarcane and preventing blockage.

[0039] See also Figure 1 and Figure 3 The visual inspection mechanism includes a detection box 8, a camera 9 and an image processing module. The camera 9 is arranged in the detection box 8 and located above the conveyor belt 1. The camera 9 is electrically connected to the image processing module.

[0040] Wherein, a fill light 10 is provided in the detection box 8 , and the fill light 10 is located above the conveyor belt 1 .

[0041] See also Figure 1-3 The intelligent detection method for trash content of graded mechanically harvested sugarcane of this embodiment comprises the following steps:

[0042] The sugarcane segments received by the machine are poured onto the conveyor belt 1, and the conveyor belt 1 is driven by the conveying driving mechanism to convey the sugarcane segments downward.

[0043] When the sugarcane segments pass under the scraping roller 4, the scraping roller 4 will scrape off the sugarcane segments that are stacked too high and spread the sugarcane segments flat, thereby improving the accuracy and precision of visual inspection.

[0044] The scraped sugarcane segments are conveyed downward to the bottom of the camera 9, and the camera 9 obtains the image on the conveyor belt 1 and transmits the image to the image processing module, which processes the image and calculates the impurity content of the sugarcane segments.

[0045] Furthermore, the specific operations of the image processing module for image processing are:

[0046] The Python-OpenCV image processing technology is used to process the image to obtain each component and obtain the pixel area of ​​different components.

[0047] Since different varieties of sugarcane and their impurity densities are different and their shapes are irregular, they need to be measured according to the growth conditions of the sugarcane that year. Before the detection of impurity content in machine-harvested sugarcane, the impurities are collected in advance and their density is measured by classification. The specific calculation formula is as follows: for each impurity density ρ, the mass M of the impurity is first weighed with a high-precision electronic scale, and then its volume V is measured using the drainage method. The average value is taken after multiple measurements to establish a quantitative impurity content model based on the impurity pixel area of ​​the image.

[0048] N images were randomly selected from the collected sugarcane images as samples, and the internal areas of sugarcane and impurities in the sample images were intercepted using VisualStudio software configured with OpenCV. The R, G, B, H, and S parameters of the pixels in the area were statistically analyzed to obtain the color characteristic parameter value ranges of sugarcane and impurities.

[0049] Multiple denoising methods are used to reduce image noise and smooth impurity areas. The gradient vector between pixels of the machine-harvested sugarcane image is solved to obtain the gradient amplitude. The local gradient of the machine-harvested sugarcane image is scanned to find the maximum amplitude in the window as the edge point. The image edge information is extracted using the improved Canny operator, and the extracted image is superimposed with the original image to achieve edge enhancement. The closed area is traversed to identify the various components in the sugarcane image. Morphological processing is performed based on the impurity geometric features to obtain the impurity segmentation result. The precision and recall rates are used to evaluate the image segmentation results and test the performance of the image segmentation algorithm.

[0050] Based on the Visual Studio environment, the deep learning framework Pytorch is used to build the network model, the open source library Open CV is used for image processing, and the PyQT5 open source library is used to write the interface program; the lower computer controls the camera 9 to collect images in real time, and the image information is encoded and transmitted to the upper computer server through the Raspberry Pi. The server first decodes the image information and performs impurity segmentation on the image. The segmented image is classified and identified by the YOLO v4 model, and data is processed based on the impurity quantification model to obtain the real-time impurity rate and average impurity rate based on quality, and finally output the results together with the segmented image.

[0051] The above is a preferred embodiment of the present invention, but the embodiment of the present invention is not limited to the above content. Any other changes, modifications, substitutions, combinations, and simplifications made without departing from the spirit and principles of the present invention should be equivalent replacement methods and are included in the protection scope of the present invention.

Claims

1. An intelligent detection method for trash content in graded machine-harvested sugarcane, characterized in that: The following steps are involved: Pour the sugarcane segments received by the machine onto the conveyor belt, and drive the conveyor belt through the conveying drive mechanism to transport the sugarcane segments downward; When the sugarcane segments pass under the scraping roller, the scraping roller will scrape off the sugarcane segments that are stacked too high and spread them out, thereby improving the accuracy and precision of visual inspection; The scraped sugarcane segments are conveyed downward to the bottom of the camera, and the camera obtains the image on the conveyor belt and transmits the image to the image processing module, which processes the image and calculates the impurity content of the sugarcane segments. The image processing module performs the following operations on the image: Before testing the impurity content of mechanically harvested sugarcane, collect impurities and measure their density in advance: use a high-precision electronic scale to weigh the mass M of the impurity, use the liquid displacement method to measure its volume V, calculate the density value according to the density formula, take the average value of multiple measurements, and establish a quantitative model of impurity content based on the pixel area of ​​impurities in the image; Use Python-OpenCV image processing technology to process the image to obtain each component and obtain the pixel area of ​​different components; Randomly select n images from the collected sugarcane images as samples, use Visual Studio software configured with Open CV to intercept the internal areas of sugarcane and impurities in the sample images, statistically analyze the R, G, B, H, and S parameters of the pixels in the area, and obtain the color characteristic parameter value range of sugarcane and impurities; Multiple noise reduction methods are used to reduce image noise and smooth impurity areas. The gradient vector between pixels of the machine-harvested sugarcane image is solved to obtain the gradient amplitude. The local gradient of the machine-harvested sugarcane image is scanned to find the maximum amplitude in the window as the edge point. The image edge information is extracted using the improved Canny operator. The extracted image is then superimposed with the original image to achieve edge enhancement. The closed area is traversed to identify the various components in the sugarcane image. Morphological processing is performed in combination with the geometric features of impurities to obtain the impurity segmentation result. The precision and recall rates are used to evaluate the image segmentation results and test the performance of the image segmentation algorithm. Developed based on Visual Studio environment, using deep learning framework Pytorch to build network models, using open source library Open CV for image processing, and using PyQT5 open source library to write interface programs; The lower computer controls the camera to collect images in real time, and the Raspberry Pi encodes the image information and transmits it to the upper computer server. The server first decodes the image information and performs impurity segmentation on the image. The segmented image is classified and identified using the YOLO v4 model, and data is processed based on the impurity quantification model to obtain the real-time impurity rate and average impurity rate based on quality, and finally outputs the result together with the segmented image. The intelligent detection method for trash content of graded mechanically harvested sugarcane is applied to an intelligent detection device, which includes a sugarcane segment conveying mechanism, a scraping mechanism, and a visual detection mechanism; The sugarcane segment conveying mechanism comprises a conveyor belt and a conveying drive mechanism; The scraping mechanism comprises a scraping installation plate and a scraping roller, wherein the scraping installation plate is arranged above the conveyor belt, and the scraping roller is rotatably connected to the bottom of the scraping installation plate; a limiting channel for the sugarcane segment to pass through is formed between the scraping roller and the conveyor belt below; The visual inspection mechanism includes an inspection box, a camera and an image processing module. The camera is arranged in the inspection box and located above the conveyor belt. The camera is electrically connected to the image processing module.

2. The intelligent detection method for trash content of graded machine-harvested sugarcane according to claim 1 is characterized in that: The conveying drive mechanism comprises a conveying drive motor and a conveying roller. Two conveying rollers are provided and are rotatably connected to the workbench. The conveying drive motor is connected to one of the conveying rollers. The conveying belt is connected between the two conveying rollers.

3. The intelligent detection method for trash content in graded machine-harvested sugarcane according to claim 1, characterized in that: The scraping mounting plate is arranged above the conveyor belt through the mounting frame, and a vertical adjustment structure is provided between the scraping mounting plate and the mounting frame, and the vertical adjustment structure includes a vertical adjustment rod, a guide rod and a guide hole, the vertical adjustment rod is vertically arranged and connected to the mounting frame through a threaded structure; the guide hole is arranged on the side of the mounting frame; one end of the guide rod is fixed on the scraping mounting plate, and the other end of the guide rod extends horizontally into the guide hole.

4. The intelligent detection method for trash content in graded machine-harvested sugarcane according to claim 1, characterized in that: The scraping mechanism is provided with at least two groups and arranged along the conveying direction of the conveyor belt; Along the conveying direction, the height of the limiting channel located at the rear is smaller than the height of the limiting channel located at the front.

5. The intelligent detection method for trash content in graded machine-harvested sugarcane according to claim 1, characterized in that: A fill light is arranged in the detection box, and the fill light is located above the conveyor belt.

Images