Integrated method for cactus fruit thorn removal and sorting based on image processing technology

Abstract

The present invention discloses an integrated method for cactus fruit thorn removal and sorting based on image processing technology. First, a thorn removal machine is used to clean and remove the cactus fruit; then the thornless cactus fruit is transported to a quality sorting device, and a rot detection algorithm based on deep learning technology is used to detect the rot of the cactus fruit; and according to the detection results, a stepper motor is used to control the tray to rotate in different directions to screen the cactus fruit; the selected cactus fruit with good quality is transported to a conveyor belt, and a camera is used to collect photos of the cactus fruit on the conveyor belt and transmit them to a computer in real time; first, the color image is preprocessed to obtain a grayscale image; then, the grayscale image is feature extracted to extract the target area features required for image segmentation; finally, the image segmentation technology is used to obtain the counting result. Compared with the traditional cactus fruit thorn removal and counting method, the present invention has the characteristics of strong public practicability, high timeliness, low cost, high automation, and high production efficiency.

Description

Technical Field

[0001] The invention relates to the technical field of digital image processing, agricultural product processing and sorting and counting devices, and in particular to a system and method for counting thornless cactus fruits based on image processing technology. Background Art

[0002] Cactus fruit is the fruit of the cactus plant. It is a high-protein, low-fat, low-sugar, high-dietary-fiber fruit. In addition, many studies have also proven that the polysaccharides extracted from cactus fruit have a good effect in preventing and treating high blood pressure, high blood sugar, high blood lipids and anti-tumor.

[0003] However, since there are many burrs on the surface of cactus fruit, and the burrs are very small, if they are not treated, when people eat cactus fruit at home, it is easy for the burrs to pierce the inside of the hands, and it is difficult to remove them after piercing. Therefore, the burrs of cactus fruit must be thoroughly cleaned before they are sold. Before selling, cactus fruit generally goes through two processes. The first process is to remove the thorns, which is currently done manually. On the one hand, the burrs will pierce the hands, and on the other hand, it will cause the cactus fruit to be damaged and difficult to preserve; the second process is sorting and counting, which is currently done manually. The workload of removing the thorns and sorting and counting of cactus fruit is large, and the manual operation time is long, resulting in the long retention time of cactus fruit in the hands of fruit farmers, and it cannot be put on the market in time, affecting the economic benefits of fruit farmers. At the same time, manual operation also has problems such as high labor intensity, low efficiency, poor thorn removal effect, few sorting grades, poor consistency, and inaccurate counting results. In view of the problems existing in the processing of cactus fruit, it is an urgent need for the development of the cactus fruit industry to develop a cactus fruit processing equipment with automatic thorn removal and intelligent sorting and counting. Summary of the invention

[0004] The present invention overcomes the shortcomings of the prior art and provides a system and method for counting thornless cactus fruits based on image processing technology. One of the technical problems to be solved is that the existing cactus fruit picking process or after thorn removal requires manual counting.

[0005] In view of the above problems in the prior art, according to one aspect of the present invention, in order to solve the above technical problems, the present invention adopts the following technical solutions:

[0006] An integrated method for removing thorns and sorting cactus fruits based on image processing technology, comprising:

[0007] S1: transporting the cleaned and thornless cactus fruits to a fruit quality sorting device, weighing the cactus fruits using a weighing device in the fruit quality sorting device, transporting the cactus fruits with a weight less than a preset value from a discharge port to a fruit collection box, and subjecting the cactus fruits with a weight greater than or equal to the preset value to further quality judgment;

[0008] S2: Use image acquisition equipment to acquire images of cactus fruits that meet the weight requirements, send the acquired images to a computer, and then use a fruit rot detection algorithm based on deep learning technology built into the computer to perform morphological judgment on the cactus fruits; transport the cactus fruits that are determined to be rotten by the algorithm to a fruit collection box through a discharge port; and transport the cactus fruits that are determined to be of good quality by the algorithm to a conveyor belt;

[0009] S3: Use an image acquisition device to acquire images of good quality cactus fruits on the conveyor belt, and transmit the acquired images to a computer in real time. Then use a cactus fruit counting algorithm built into the computer to count the good quality cactus fruits, and then output the counting results.

[0010] In order to better realize the present invention, a further technical solution is:

[0011] Furthermore, the step S1 includes:

[0012] After the weighing tray has finished weighing the cactus fruit, the weight is sent to the computer in real time;

[0013] There are two sorting levers at the bottom of the weighing tray, which are connected to the stepper motor. When the weight of the cactus fruit is less than the preset value, the computer returns a signal to the stepper motor, and the stepper motor controls the tray to rotate 45° to the left, so that the cactus fruit with a weight less than the preset value is transported from the discharge port to the fruit collection box.

[0014] Furthermore, in step S2, the method of using the fruit decay detection algorithm based on deep learning technology built into the computer to perform morphological judgment on the cactus fruit includes:

[0015] The computer processes the collected graphics to create a data set;

[0016] Use the dataset to train the network model;

[0017] Test the trained network model;

[0018] The trained network model is used to detect the corruption of cactus fruit and screen and classify them.

[0019] Furthermore, in step S2, after the algorithm determines that the cactus fruit is of good quality, the computer sends a signal to the stepper motor. If there is no rotten area in the fruit, the stepper motor controls the tray to rotate 45° to the right, so that the cactus fruit of good quality is transported to the conveyor belt.

[0020] Furthermore, in the step S3, the image of the cactus fruit with good quality on the collection conveyor belt is a color image M.

[0021] Furthermore, in step S3, the method of using the cactus fruit counting algorithm built into the computer to count the cactus fruits with good quality includes:

[0022] Perform general grayscale processing on the color image M to obtain a grayscale image G1;

[0023] Perform selective grayscale processing on the color image M to obtain a grayscale image G2;

[0024] Denoise and filter the grayscale images G1 and G2 respectively to obtain images Q1 and Q2;

[0025] Perform edge detection on the grayscale image Q1 to obtain image P1;

[0026] Mix the grayscale image Q2 with the image P1 to obtain a mixed image P2;

[0027] Perform morphological corrosion and reconstruction on the mixed image P2 to obtain image P3;

[0028] Mark the foreground and background of image P3 to obtain image P4;

[0029] Image P5 is obtained by performing isolated pixel and hole processing on image P4;

[0030] Image P5 is segmented and counted to obtain the number of cactus fruits in the image.

[0031] Furthermore, the image acquisition device in step S2 and / or S3 is a camera.

[0032] Furthermore, the image acquisition device in step S2 and / or S3 is arranged above the conveyor belt.

[0033] Furthermore, the weighing device in step S1 is a weighing tray.

[0034] Compared with the prior art, one of the beneficial effects of the present invention is:

[0035] The integrated cactus fruit thorn removal and sorting method based on image processing technology of the present invention is a solution for cactus fruit thorn removal and counting. First, the cactus fruits picked from the farm are thorned, and then the cactus fruits are counted using image processing technology. Compared with traditional cactus fruit thorn removal and counting methods, the present invention has the characteristics of strong public practicability, high timeliness, low cost, high degree of automation and high production efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] In order to more clearly illustrate the embodiments of the present application document or the technical solutions in the prior art, the drawings required for use in the description of the embodiments or the prior art are briefly introduced below. Obviously, the drawings described below are only references to some embodiments in the present application document. For those skilled in the art, other drawings can be obtained based on these drawings without creative work.

[0037] Figure 1 Schematic diagram of a cactus fruit thorn removal and quality sorting and counting device according to an embodiment of the present invention.

[0038] Figure 2 Schematic diagram of a quality sorting device according to an embodiment of the present invention.

[0039] Figure 3 The figure is a flowchart of a corruption detection algorithm based on deep learning according to an embodiment of the present invention.

[0040] Figure 4 The figure is a flow chart of a cactus fruit counting algorithm based on image processing technology according to an embodiment of the present invention.

[0041] The names of the drawings corresponding to the reference numerals are:

[0042] 1-feeding port, 2-thorn remover, 3-weighing tray, 4-fruit collecting box, 5-first camera, 6-second camera, 7-sorting lever, 8-stepping motor. DETAILED DESCRIPTION

[0043] The present invention is further described in detail below in conjunction with examples, but the embodiments of the present invention are not limited thereto.

[0044] See also Figure 3 As shown, an integrated method for removing thorns and sorting cactus fruits based on image processing technology comprises:

[0045] S1: The cleaned and dethorned cactus fruits are transported to a fruit quality sorting device, and the cactus fruits are weighed using a weighing device in the fruit quality sorting device. The cactus fruits with a weight less than a preset value are transported from a discharge port to a fruit collection box, and the cactus fruits with a weight greater than or equal to the preset value are further judged for quality.

[0046] Among them, combined Figure 1 and Figure 2 As shown, the cactus fruit enters the thorn remover 2 from the feed port 1 and Figure 1 Clean the A shown in the figure. Figure 1The thorns are removed at B as shown, and after the thorns are removed, the fruit enters the quality sorting device through the discharge port and is weighed by a weighing device, which can be a weighing tray 3. After weighing, the unqualified ones are discharged into the fruit collection box 4.

[0047] And the preferred solution of this step 1 is to continue as follows Figure 1 As shown, after the weighing tray 3 weighs the cactus fruit, the weight is sent to the computer in real time;

[0048] like Figure 2 As shown, there are two sorting levers 7 at the bottom of the weighing tray 3, and the sorting levers 7 are connected to the stepping motor 8. When the weight of the cactus fruit is less than the preset value, the computer returns a signal to the stepping motor 8, and the stepping motor 8 controls the weighing tray 3 to rotate 45° to the left, so that the cactus fruit with a weight less than the preset value is transported from the discharge port to the fruit collection box 4.

[0049] S2: Use image acquisition equipment to acquire images of cactus fruits that meet the weight requirements, send the acquired images to a computer, and then use a fruit rot detection algorithm based on deep learning technology built into the computer to perform morphological judgment on the cactus fruits; transport the cactus fruits that are determined to be rotten by the algorithm to a fruit collection box through a discharge port; and transport the cactus fruits that are determined to be of good quality by the algorithm to a conveyor belt;

[0050] Among them, Figure 1 As shown, the image acquisition device of this embodiment can be a first camera 5, and the first camera 5 can be arranged on the top of the fruit quality sorting device.

[0051] The above methods of using the deep learning technology-based fruit rot detection algorithm built into the computer to perform morphological judgment on cactus fruit include:

[0052] The computer processes the collected graphics to create a data set;

[0053] Use the dataset to train the network model;

[0054] Test the trained network model;

[0055] The trained network model is used to detect the corruption of cactus fruit and screen and classify them.

[0056] And if there is no rotten area on the fruit, the algorithm determines that the cactus fruit is of good quality, and the computer sends a signal to the stepper motor, and the stepper motor controls the tray to rotate 45 degrees to the right, so that the cactus fruit of good quality is transported to the conveyor belt.

[0057] S3: Use an image acquisition device to acquire images of good quality cactus fruits on the conveyor belt, and transmit the acquired images to a computer in real time. Then use a cactus fruit counting algorithm built into the computer to count the good quality cactus fruits, and then output the counting results.

[0058] Among them, Figure 1 As shown, the image acquisition device in this step may preferably be a second camera 6, and the second camera 6 may be arranged above the conveyor belt.

[0059] Generally, the cactus fruit image with good quality collected on the conveyor belt is preferably a color image M, and the specific processing method thereof can adopt the following schemes, such as Figure 4 As shown:

[0060] Perform general grayscale processing on the color image M to obtain a grayscale image G1;

[0061] Perform selective grayscale processing on the color image M to obtain a grayscale image G2;

[0062] Denoise and filter the grayscale images G1 and G2 respectively to obtain images Q1 and Q2;

[0063] Perform edge detection on the grayscale image Q1 to obtain image P1;

[0064] Mix the grayscale image Q2 with the image P1 to obtain a mixed image P2;

[0065] Perform morphological corrosion and reconstruction on the mixed image P2 to obtain image P3;

[0066] Mark the foreground and background of image P3 to obtain image P4;

[0067] Image P5 is obtained by performing isolated pixel and hole processing on image P4;

[0068] Image P5 is segmented and counted to obtain the number of cactus fruits in the image.

[0069] In summary, in view of the problems existing in the current cactus fruit thorn removal and counting process: first, the cactus fruit thorn removal and counting process cannot be carried out at the same time, thus resulting in low production efficiency; second, the use of manual counting has high labor intensity and inaccurate counting results. The present invention designs a cactus fruit thorn removal and counting method that combines image processing technology and mobile device technology, so as to achieve fast, reliable, low-cost and highly automated thorn removal and counting of cactus fruits. First, the cactus fruit is cleaned and thorned by a thorn removal machine; then the thornless cactus fruit is transported to a conveyor belt, and a camera is used to collect photos of the cactus fruit on the conveyor belt and transmit them to a computer in real time; first, the color image is preprocessed, and the color image is respectively subjected to general grayscale and selective grayscale to obtain two grayscale images, and wavelet denoising and median filtering are performed on the two grayscale images respectively to obtain two images after image preprocessing operations; then the grayscale image is feature extracted. Since the general grayscale image retains more graphic edge details, the algorithm should start from the general grayscale image to extract the edge of the target object. Then, the extracted edge is superimposed with the preprocessed selective grayscale image to obtain a grayscale image with enhanced target edge. Finally, the adhered image is further separated by morphological-based corrosion and dilation operations and image reconstruction calculations to extract the target area features required for image segmentation; the image is segmented using the watershed algorithm, and the foreground and background need to be marked first when applying the algorithm. After marking, there are some factors in the foreground and background that interfere with the results, mainly isolated pixels and a small number of holes, so these interference factors need to be cleaned up. Finally, the watershed algorithm is used to obtain the segmentation results and counting results. Compared with traditional cactus fruit thorn removal and counting methods, the present invention has the characteristics of strong public practicability, high timeliness, low cost, high automation, and high production efficiency.

[0070] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the various embodiments can be referenced to each other.

[0071] References to "one embodiment", "another embodiment", "embodiment", etc. in this specification refer to specific features, structures or characteristics described in conjunction with the embodiment as included in at least one embodiment generally described in this application. The appearance of the same expression in multiple places in the specification does not necessarily refer to the same embodiment. Further, when a specific feature, structure or characteristic is described in conjunction with any embodiment, it is claimed that the realization of such feature, structure or characteristic in conjunction with other embodiments also falls within the scope of the present invention.

[0072] Although the present invention is described herein with reference to a number of illustrative embodiments of the present invention, it will be appreciated that those skilled in the art may devise many other modifications and implementations that fall within the scope and spirit of the principles disclosed herein. More specifically, within the scope of the present disclosure, drawings, and claims, a variety of variations and improvements may be made to the components and / or layout of the subject combination layout. In addition to variations and improvements made to the components and / or layout, other uses will also be apparent to those skilled in the art.

Claims

1. An integrated method for cactus fruit thorn removal and sorting based on image processing technology, Features include: S1: transporting the cleaned and thornless cactus fruits to a fruit quality sorting device, weighing the cactus fruits using a weighing device in the fruit quality sorting device, transporting the cactus fruits with a weight less than a preset value from a discharge port to a fruit collection box, and subjecting the cactus fruits with a weight greater than or equal to the preset value to further quality judgment; S2: Use image acquisition equipment to acquire images of cactus fruits that meet the weight requirements, send the acquired images to a computer, and then use a fruit rot detection algorithm based on deep learning technology built into the computer to perform morphological judgment on the cactus fruits; transport the cactus fruits that are determined to be rotten by the algorithm to a fruit collection box through a discharge port; and transport the cactus fruits that are determined to be of good quality by the algorithm to a conveyor belt; S3: using an image acquisition device to acquire images of good quality cactus fruits on the conveyor belt, wherein the acquired images of good quality cactus fruits on the conveyor belt are color images M, and the acquired images are transmitted to a computer in real time, and then the cactus fruit counting algorithm built into the computer is used to count the good quality cactus fruits, and the counting method includes: Perform general grayscale processing on the color image M to obtain a grayscale image G1; Perform selective grayscale processing on the color image M to obtain a grayscale image G2; Denoise and filter the grayscale images G1 and G2 respectively to obtain images Q1 and Q2; Perform edge detection on the grayscale image Q1 to obtain image P1; Mix the grayscale image Q2 with the image P1 to obtain a mixed image P2; Perform morphological corrosion and reconstruction on the mixed image P2 to obtain image P3; Mark the foreground and background of image P3 to obtain image P4; Image P5 is obtained by performing isolated pixel and hole processing on image P4; Image P5 is segmented and counted to obtain the number of cactus fruits in the image; and the counting result is then output.

2. The integrated method for removing thorns and sorting cactus fruits based on image processing technology according to claim 1, Features The step S1 includes: After the weighing tray has finished weighing the cactus fruit, the weight is sent to the computer in real time; There are two sorting levers at the bottom of the weighing tray, which are connected to the stepper motor. When the weight of the cactus fruit is less than the preset value, the computer returns a signal to the stepper motor, and the stepper motor controls the tray to rotate 45° to the left, so that the cactus fruit with a weight less than the preset value is transported from the discharge port to the fruit collection box.

3. The integrated method for removing thorns and sorting cactus fruits based on image processing technology according to claim 1, Features In step S2, the method of using a fruit decay detection algorithm based on deep learning technology built into a computer to perform morphological judgment on the cactus fruit includes: The computer processes the collected graphics to create a data set; Use the dataset to train the network model; Test the trained network model; The trained network model is used to detect the corruption of cactus fruit and screen and classify them.

4. The integrated method for removing thorns and sorting cactus fruits based on image processing technology according to claim 2, Features In step S2, after the algorithm determines that the cactus fruit is of good quality, the computer sends a signal to the stepper motor. If there is no rotten area in the fruit, the stepper motor controls the tray to rotate 45° to the right, so that the cactus fruit of good quality is transported to the conveyor belt.

5. The integrated method for removing thorns and sorting cactus fruits based on image processing technology according to claim 1, Features The image acquisition device in step S2 and / or S3 is a camera.

6. The integrated method for removing thorns and sorting cactus fruits based on image processing technology according to claim 1, Features The image acquisition device in step S2 and / or S3 is arranged above the conveyor belt.

7. The integrated method for removing thorns and sorting cactus fruits based on image processing technology according to claim 1, Features The weighing device in step S1 is a weighing tray.

Images