Highland barley stir-frying method and system

Abstract

The invention discloses a highland barley stir-frying method and system, and relates to the technical field of food processing.The method comprises the steps that a highland barley stir-frying machine is started, a roller is preheated to a first preset temperature, highland barley is put into a feeding port, a first image of the highland barley is obtained, the highland barley is transmitted to the roller, the highland barley is stir-fried, and a second image of the highland barley is obtained; constructing a recognition model, and analyzing the first image and the second image by the recognition model to obtain an analysis result; based on the analysis result, the highland barley is stir-fried, the stir-fried highland barley is conveyed to the discharging port and conveyed to the collecting device through the guide groove, and the problem that the yield of cooked highland barley is reduced due to the fact that the stir-frying state of the highland barley cannot be accurately mastered at present can be solved.

Description

Technical Field

[0001] The invention relates to the technical field of food processing, and in particular to a method and system for frying highland barley. Background Art

[0003] Highland barley frying machine is a kind of mechanical equipment specially used for frying highland barley, including shell, feed port, feeding hopper, guide pipe, roller, motor, transmission parts, heating device, temperature detection device, discharge port, guide trough and collection device, etc. Highland barley frying machine improves the efficiency of frying highland barley, but the frying time and temperature of highland barley are usually preset fixed time and fixed temperature, and it is impossible to accurately grasp the frying state of highland barley, resulting in undercooked, burnt or gelatinized highland barley, thereby reducing the qualified rate of cooked highland barley. Summary of the invention

[0004] In order to solve the problem that the frying state of highland barley cannot be accurately grasped, resulting in a lower qualified rate of cooked highland barley, the present invention provides a highland barley frying method, the method comprising: starting a highland barley frying machine, preheating a drum to a first preset temperature, putting highland barley into a feed port, acquiring a first image of the highland barley, transmitting the highland barley to the drum, frying the highland barley, and acquiring a second image of the highland barley; constructing a recognition model, the recognition model analyzes the first image and the second image to obtain an analysis result; based on the analysis result, frying the highland barley, transmitting the highland barley to a discharge port after frying, and transmitting the highland barley to a collecting device through a material guide chute.

[0005] Principle of this method: During the frying process of highland barley, due to high temperature heating, the moisture inside the highland barley evaporates rapidly, causing the highland barley to expand and burst. This bursting phenomenon causes the edge contour of the highland barley to gradually change. By constructing a recognition model, the changes and differences in the edge contour from raw highland barley to cooked highland barley are extracted, so as to more accurately distinguish the different states of highland barley frying, adjust the frying time and temperature in real time according to the state, and improve the qualified rate of cooked highland barley.

[0006] Furthermore, the specific steps of constructing the recognition model include: obtaining historical roasted barley images, classifying the historical roasted barley images, and obtaining several barley image sets; annotating the barley image sets, and obtaining several annotated image sets; performing background segmentation on the annotated image sets, and obtaining several first image sets, converting the color model of the first image sets to the HSI color model, and obtaining several second image sets; performing smoothing on the second image sets, and obtaining several third image sets; based on the third image sets, calculating the gradient direction and intensity of each pixel point, obtaining an edge threshold based on the gradient direction and the intensity, and extracting the edge contour of the third image set based on the edge threshold, and obtaining several fourth image sets; training the model based on the fourth image set to obtain the recognition model.

[0007] Furthermore, the specific steps of constructing the recognition model also include: obtaining the contour line of each of the fourth image sets based on the edge contour, dividing each of the fourth image sets into a number of first sub-regions based on the contour line, obtaining the HIS parameters of each of the first sub-regions, obtaining the HIS parameter range based on the HIS parameters, training the recognition model based on the HIS parameters and the HIS parameter range to obtain a first model, and updating the recognition model to the first model.

[0008] By obtaining the changes and differences in HIS parameters during the process of raw barley to cooked barley, the recognition model can be further trained to improve the recognition rate of the model.

[0009] Furthermore, the specific steps of constructing the recognition model also include: obtaining the number of contour lines of each of the fourth image sets based on the contour lines, obtaining the contour line variation range based on the number of contour lines, training the recognition model based on the contour line variation range to obtain a second model, and updating the recognition model to the second model.

[0010] Due to the bursting phenomenon during the frying process of highland barley, the edge contours of highland barley become more during the maturity process, and the number of contour lines gradually increases. By obtaining the changes and differences in the number of contour lines in the process from raw highland barley to cooked highland barley, the recognition model can be further trained to improve the recognition rate of the model.

[0011] Furthermore, the specific steps of constructing the recognition model also include: obtaining the number of areas of the first sub-area in each of the fourth image sets, and obtaining several quantity change ratio ranges based on the number of areas; obtaining the area area of ​​the first sub-area, obtaining the area range of each of the fourth image sets based on the area area, and obtaining several area change ranges and several area change ratio ranges based on the area range; training the recognition model based on the quantity change ratio range, the area change range and the area change ratio range to obtain a third model, and updating the recognition model to the third model.

[0012] Similarly, due to the bursting phenomenon during the frying process of highland barley, the barley will burst into pieces, and the number of its regions will increase, and the area of ​​the bursting regions will also change. By obtaining the changes and differences in the number and area of ​​regions from raw highland barley to cooked highland barley, the recognition model can be further trained to improve the recognition rate of the model.

[0013] Furthermore, the specific steps of constructing the recognition model also include: marking the first image set based on the contour line to obtain several fifth image sets; obtaining the first area of ​​each of the fifth image sets, dividing each of the fifth image sets to obtain several second sub-areas, obtaining the pixel value of each of the second sub-areas, obtaining the second area based on the pixel value and the preset pixel range, and obtaining the color change ratio range based on the first area and the second area; obtaining the pixel contour line of each of the fifth image sets based on the contour line and the preset pixel range, obtaining the number of pixel contour lines of the pixel contour line, and obtaining the pixel contour line variation range based on the number of pixel contour lines and the number of contour lines; training the recognition model based on the color change ratio range and the pixel contour line variation range to obtain a fourth model, and updating the recognition model to the fourth model.

[0014] The color of raw barley is usually green or yellowish brown, while the color of cooked barley will turn into golden yellow or dark yellow after cooking. In the process of popping barley, the moisture and starch inside the barley will change, resulting in white areas in the popped part, and white contour lines will appear between the white areas. By obtaining the changes and differences in pixel values ​​and the number of white contour lines in the process of raw barley to cooked barley, the recognition model can be further trained to improve the recognition rate of the model.

[0015] Furthermore, the specific steps of obtaining the analysis results include: constructing a highland barley maturity table; the recognition model analyzes the first image and the second image to obtain an analysis edge contour, obtains highland barley maturity based on the analysis edge contour, obtains frying data based on the highland barley maturity and the highland barley maturity table, and obtains the analysis results based on the frying data, wherein the frying data includes frying time and frying temperature.

[0016] The current maturity of the barley can be judged by the barley maturity table and the edge contour of the barley, so that the frying time and temperature can be adjusted in real time, which can improve the qualified rate of cooked barley.

[0017] Furthermore, the specific steps of obtaining the analysis results also include: the recognition model analyzes the first image and the second image to obtain change data, and the change data includes HIS parameter range, contour line change range, quantity change ratio range, area change range, area change ratio range, color change ratio range and pixel contour line change range; based on the change data, the second barley maturity is obtained, based on the preset weight, the first barley maturity and the second barley maturity, the third barley maturity is obtained, and the first barley maturity is updated to the third barley maturity.

[0018] By obtaining the changes and differences of various parameters in the process of raw barley to cooked barley, multi-dimensional analysis can be conducted to more accurately judge the current maturity of the barley, thereby adjusting the frying time and temperature in real time, which can improve the qualified rate of cooked barley.

[0019] Furthermore, the feed port and the drum of the highland barley frying machine are equipped with a collection device, and a temperature sensor is installed inside the drum to collect real-time images of raw highland barley and highland barley during frying, which is convenient for subsequent analysis.

[0020] The present invention also provides a highland barley frying system, the system comprising:

[0021] A starting unit is used to start the highland barley frying machine, preheat the drum to a first preset temperature, put highland barley into a feed port, obtain a first image of the highland barley, and transmit the highland barley to the drum;

[0022] A frying unit: used for frying highland barley and obtaining a second image of the highland barley;

[0023] An analysis unit is used to construct a recognition model, wherein the recognition model analyzes the first image and the second image to obtain an analysis result;

[0024] Result unit: used for frying highland barley based on the analysis result, and after frying, transferring the highland barley to a discharge port, and then transferring the highland barley to a collecting device through a guide trough.

[0025] The principle and effect of this system are similar to those of this method, and no corresponding description will be given for this system.

[0026] One or more technical solutions provided by the present invention have at least the following technical effects or advantages:

[0027] 1. Build a recognition model to extract the changes and differences in the edge contour from raw barley to cooked barley, so as to more accurately distinguish the different states of barley frying, adjust the frying time and temperature in real time according to the state, and improve the qualified rate of cooked barley.

[0028] 2. By obtaining the changes and differences in HIS parameters, the number of contour lines, the ratio range of quantity change and the ratio range of area change, the ratio range of color change and the range of pixel contour line change in the process of raw barley to cooked barley, multi-dimensional analysis is performed to further train the recognition model, which can improve the recognition rate of the model. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] The drawings described herein are used to provide a further understanding of the embodiments of the present invention, constitute a part of the present invention, and do not constitute a limitation on the embodiments of the present invention;

[0030] Figure 1 It is a schematic flow chart of constructing an identification model for a method for frying highland barley in the present invention;

[0031] Figure 2 The invention discloses a schematic flow chart of an update recognition model of a highland barley frying method. DETAILED DESCRIPTION

[0032] In order to more clearly understand the above-mentioned purpose, features and advantages of the present invention, the present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be noted that the embodiments of the present invention and the features in the embodiments can be combined with each other without conflict.

[0033] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those within the scope of this description. Therefore, the protection scope of the present invention is not limited to the specific embodiments disclosed below.

[0034] Embodiment 1

[0035] refer to Figure 1 This embodiment provides a method for frying highland barley, the method comprising:

[0036] Starting a highland barley frying machine, preheating a drum to a first preset temperature, feeding highland barley into a feed port, acquiring a first image of the highland barley, transmitting the highland barley to the drum, frying the highland barley, and acquiring a second image of the highland barley;

[0037] Wherein, the feed inlet and the drum of the highland barley frying machine are equipped with a collection device, and a temperature sensor is installed inside the drum. In this embodiment, the collection device can be a high-temperature resistant camera, and the temperature sensor can be a high-precision thermistor temperature sensor that is resistant to high temperatures.

[0038] Constructing a recognition model, wherein the recognition model analyzes the first image and the second image to obtain an analysis result;

[0039] Based on the analysis results, the highland barley is fried, and after frying, the highland barley is transferred to a discharge port and then transferred to a collecting device through a guide chute.

[0040] Among them, the specific steps of constructing the recognition model include: obtaining historical roasted barley images, classifying the historical roasted barley images, and obtaining several barley image sets, such as classifying the images according to the maturity of the barley to obtain raw barley images, half-ripe barley images, ripe barley images, and over-ripe barley images, etc.; annotating the barley image set to obtain several annotated image sets; using MATLAB tools to perform background segmentation on the annotated image set to obtain several first image sets, converting the color model of the first image set to the HSI color model, and also performing grayscale processing, bilateral filtering, and image threshold segmentation, etc. Preprocessing makes the image clearer and reduces noise to obtain several second image sets; using convolution of bilateral filtering, that is, combining bilateral filtering and convolution, the second image set is smoothed to obtain several third image sets; based on the third image set, the gradient direction and intensity of each pixel are calculated, non-maximum suppression is used to eliminate stray responses, and the maximum inter-class variance method and other methods are used to obtain the edge threshold based on the gradient direction and the intensity, and the edge contour of the third image set is extracted based on the edge threshold to obtain several fourth image sets; based on the fourth image set training model, the recognition model is obtained.

[0041] The specific steps of obtaining the analysis results include:

[0042] Construct a highland barley maturity table; as shown in Table 1:

[0043] Table 1

[0044] Maturity Stir-frying time (min) Frying temperature (°) 0.2 40 270 0.5 20 250 0.6 10 220

[0045] The recognition model analyzes the first image and the second image to obtain an analysis edge contour, obtains the maturity of highland barley based on the analysis edge contour, such as obtaining the maturity of highland barley based on the difference in edge contours of highland barley at different stages, obtains frying data based on the highland barley maturity and the highland barley maturity table, obtains the analysis result based on the frying data, and the frying data includes frying time and frying temperature.

[0046] Embodiment 2

[0047] refer to Figure 2 Based on the first embodiment, in this embodiment, the specific steps of constructing the recognition model further include:

[0048] Based on the edge contour, the contour line of each of the fourth image sets is obtained, such as obtaining the shortest path between two points as a contour line, and dividing each of the fourth image sets into a number of first sub-areas based on the contour line, and obtaining the HIS parameters of each of the first sub-areas, and obtaining the HIS parameter range based on the HIS parameters. For example, the HIS parameter ranges of raw barley images, half-ripe barley images, ripe barley images and over-ripe barley images are different, and the recognition model is trained based on the HIS parameters and the HIS parameter range to obtain a first model, and the recognition model is updated to the first model.

[0049] Embodiment 3

[0050] refer to Figure 2 Based on the above embodiment, in this embodiment, the specific steps of constructing the recognition model further include:

[0051] Based on the contour lines, the number of contour lines of each of the fourth image sets is obtained, and the contour line variation range is obtained based on the contour line number. If the number of contour lines in each stage changes due to the popping of highland barley, the range of the number of contour lines in each stage is taken, and the recognition model is trained based on the contour line variation range to obtain a second model, and the recognition model is updated to the second model. If the range of the number of contour lines of the raw highland barley image and the half-ripe highland barley image is 200-300 and 500-600 respectively, the contour line variation range is 500-300-600-200, that is, 200-400.

[0052] In this embodiment, the contour line change ratio range can also be obtained based on the number of contour lines, and the recognition model can be trained based on the contour line change ratio range. From raw barley to half-ripe barley, the contour line change ratio range is 500 / 300 to 600 / 200, that is, 1.67 to 3.

[0053] Embodiment 4

[0054] refer to Figure 2 Based on the above embodiment, in this embodiment, the specific steps of constructing the recognition model further include:

[0055] The number of areas of the first sub-area in each of the fourth image sets is obtained, and several quantity change ratio ranges are obtained based on the number of areas; for example, the number of areas of the raw barley image and the half-ripe barley image are 200-300 and 500-600 respectively, then the quantity change ratio range is 500 / 300-600 / 200, i.e. 1.67-3.

[0056] Using algorithms such as the Monte Carlo algorithm or the watershed algorithm, the area of ​​the first sub-region is obtained, the area range of each of the fourth image sets is obtained based on the area, and several area change ranges and several area change ratio ranges are obtained based on the area range; for example, the area ranges of the first sub-region in the raw barley image and the half-ripe barley image are 0.01-0.04 and 0.005-0.045 respectively, and the area change range is 0.045-0.04-0.045-0.01, i.e. 0.005-0.035; the area change ratio range is 0.045 / 0.04-0.01-0.005, i.e. 1.125-2.

[0057] In this embodiment, the reduced area range and the expanded area range can also be obtained based on the area range, such as the reduced area range and the expanded area range are 0-0.01-0.005 and 0-0.045-0.04, that is, both are 0-0.005. Similarly, the reduced area ratio range and the expanded area ratio range can be obtained based on the area range.

[0058] Based on the quantity change ratio range, the area change range and the area change ratio range, the recognition model is trained to obtain a third model, and the recognition model is updated to the third model.

[0059] Embodiment 5

[0060] refer to Figure 2 Based on the above embodiment, in this embodiment, the specific steps of constructing the recognition model further include:

[0061] Marking the first image set based on the contour lines to obtain a plurality of fifth image sets;

[0062] Obtain the first area of ​​each of the fifth image sets, use contour lines to divide each of the fifth image sets to obtain a number of second sub-areas, obtain the pixel value of each of the second sub-areas, obtain the second area based on the pixel value and the preset pixel range, and obtain the color change ratio range based on the first area and the second area; if the appearance colors of raw barley and cooked barley are different, set different pixel value ranges to obtain the part that has changed color from raw barley to cooked barley, calculate its second area, and then obtain the color change ratio range by dividing the second area by the first area. Or because the moisture and starch inside the barley will change during the popping process, resulting in a white area in the popping part, then obtain the area corresponding to the white color, thereby calculating its color change ratio range.

[0063] Based on the contour line and the preset pixel range, the pixel contour line of each of the fifth image sets is obtained, the number of pixel contour lines of the pixel contour line is obtained, and the pixel contour line variation range is obtained based on the number of pixel contour lines and the number of contour lines; similarly, the contour line of the pixel area that has changed is obtained, and its pixel contour line variation range is calculated.

[0064] Based on the color change ratio range and the pixel contour line change range, the recognition model is trained to obtain a fourth model, and the recognition model is updated to the fourth model.

[0065] Embodiment 6

[0066] Based on the above embodiment, in this embodiment, the specific steps of obtaining the analysis result further include:

[0067] The recognition model analyzes the first image and the second image to obtain change data, wherein the change data includes an HIS parameter range, a contour line change range, a quantity change ratio range, an area change range, an area change ratio range, a color change ratio range, and a pixel contour line change range;

[0068] A second barley maturity is obtained based on the change data, a third barley maturity is obtained based on a preset weight, the first barley maturity and the second barley maturity, and the first barley maturity is updated to the third barley maturity.

[0069] Embodiment 7

[0070] This embodiment also provides a highland barley frying system, the system comprising:

[0071] A starting unit is used to start the highland barley frying machine, preheat the drum to a first preset temperature, put highland barley into a feed port, obtain a first image of the highland barley, and transmit the highland barley to the drum;

[0072] A frying unit: used for frying highland barley and obtaining a second image of the highland barley;

[0073] An analysis unit is used to construct a recognition model, wherein the recognition model analyzes the first image and the second image to obtain an analysis result;

[0074] Result unit: used for frying highland barley based on the analysis result, and after frying, transferring the highland barley to a discharge port, and then transferring the highland barley to a collecting device through a guide trough.

[0075] Although the preferred embodiments of the present invention have been described, those skilled in the art may make other changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present invention.

[0076] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. A method for frying highland barley, characterized in that: The method comprises: Starting a highland barley frying machine, preheating a drum to a first preset temperature, feeding highland barley into a feed port, acquiring a first image of the highland barley, transmitting the highland barley to the drum, frying the highland barley, and acquiring a second image of the highland barley; Constructing a recognition model, wherein the recognition model analyzes the first image and the second image to obtain an analysis result; Based on the analysis results, the highland barley is fried, and after frying, the highland barley is transferred to a discharge port and then transferred to a collecting device through a guide chute.

2. A method for frying highland barley according to claim 1, characterized in that: The specific steps to build a recognition model include: Acquire historical roasted barley images, classify the historical roasted barley images, and obtain a plurality of barley image sets; annotate the barley image sets, and obtain a plurality of annotated image sets; Performing background segmentation on the annotated image set to obtain a plurality of first image sets, converting the color model of the first image set to an HSI color model to obtain a plurality of second image sets; performing smoothing processing on the second image set to obtain a plurality of third image sets; Based on the third image set, the gradient direction and intensity of each pixel are calculated, an edge threshold is obtained based on the gradient direction and the intensity, and the edge contour of the third image set is extracted based on the edge threshold to obtain several fourth image sets; the recognition model is obtained based on the fourth image set training model.

3. A method for frying highland barley according to claim 2, characterized in that: The specific steps of building a recognition model also include: Based on the edge contour, a contour line of each of the fourth image sets is obtained; based on the contour line, each of the fourth image sets is divided into a plurality of first sub-regions; HIS parameters of each of the first sub-regions are obtained; based on the HIS parameters, a HIS parameter range is obtained; based on the HIS parameters and the HIS parameter range, the recognition model is trained to obtain a first model; and the recognition model is updated to the first model.

4. A method for frying highland barley according to claim 3, characterized in that: The specific steps of building a recognition model also include: The number of contour lines of each of the fourth image sets is obtained based on the contour lines, a contour line variation range is obtained based on the contour line number, the recognition model is trained based on the contour line variation range to obtain a second model, and the recognition model is updated to the second model.

5. A method for frying highland barley according to claim 4, characterized in that: The specific steps of building a recognition model also include: Acquire the number of regions of the first sub-region in each of the fourth image sets, and obtain a plurality of quantity change ratio ranges based on the number of regions; Acquire the area of ​​the first sub-region, obtain the area range of each of the fourth image sets based on the area, and obtain a plurality of area change ranges and a plurality of area change ratio ranges based on the area range; Based on the quantity change ratio range, the area change range and the area change ratio range, the recognition model is trained to obtain a third model, and the recognition model is updated to the third model.

6. A method for frying highland barley according to claim 5, characterized in that: The specific steps of building a recognition model also include: Marking the first image set based on the contour lines to obtain a plurality of fifth image sets; Acquire a first area of ​​each of the fifth image sets, divide each of the fifth image sets to obtain a plurality of second sub-areas, acquire a pixel value of each of the second sub-areas, acquire a second area based on the pixel value and a preset pixel range, and acquire a color change ratio range based on the first area and the second area; Obtaining a pixel contour line of each of the fifth image sets based on the contour line and the preset pixel range, obtaining the number of pixel contour lines of the pixel contour line, and obtaining a pixel contour line variation range based on the number of pixel contour lines and the number of contour lines; Based on the color change ratio range and the pixel contour line change range, the recognition model is trained to obtain a fourth model, and the recognition model is updated to the fourth model.

7. A method for frying highland barley according to claim 6, characterized in that: The specific steps to obtain the analysis results include: Construct a table of highland barley maturity; The recognition model analyzes the first image and the second image to obtain an analysis edge contour, obtains highland barley maturity based on the analysis edge contour, obtains frying data based on the highland barley maturity and the highland barley maturity table, and obtains the analysis result based on the frying data, wherein the frying data includes frying time and frying temperature.

8. A method for frying highland barley according to claim 7, characterized in that: The specific steps to obtain the analysis results also include: The recognition model analyzes the first image and the second image to obtain change data, wherein the change data includes an HIS parameter range, a contour line change range, a quantity change ratio range, an area change range, an area change ratio range, a color change ratio range, and a pixel contour line change range; A second barley maturity is obtained based on the change data, a third barley maturity is obtained based on a preset weight, the first barley maturity and the second barley maturity, and the first barley maturity is updated to the third barley maturity.

9. A method for frying highland barley according to claim 1, characterized in that: The feeding port and the drum of the highland barley frying machine are equipped with a collecting device, and a temperature sensor is installed inside the drum.

10. A highland barley frying system, characterized in that: The system comprises: A starting unit is used to start the highland barley frying machine, preheat the drum to a first preset temperature, put highland barley into a feed port, obtain a first image of the highland barley, and transmit the highland barley to the drum; A frying unit: used for frying highland barley and obtaining a second image of the highland barley; An analysis unit is used to construct a recognition model, wherein the recognition model analyzes the first image and the second image to obtain an analysis result; Result unit: used for frying highland barley based on the analysis result, and after frying, transferring the highland barley to a discharge port, and then transferring the highland barley to a collecting device through a guide trough.

Images