Hot pot bottom boiling identification method based on time sequence image analysis

Abstract

The invention discloses a hot pot bottom boiling identification method based on time sequence image analysis, and the method comprises the following steps: S1, obtaining image data, and obtaining a pot bottom region image; s2, an image classification algorithm; s3, converting a grey-scale map; s4, performing frame difference; s5, carrying out binarization; s6, performing open operation denoising; s7, counting the number of pixels; s8, performing moving average denoising, repeating the steps S1-S7, and when t is less than or equal to 1000, obtaining a category queue C = {c1, c2,..., ck} and a difference value queue N = {num1, num2,..., numk}, the upper limit of the queue C and the upper limit of the queue N being 1000; when t is greater than 1000, executing the steps S1-S7, storing the obtained results into the tail part of the queue, and meanwhile, rejecting the data at the head part; s9, counting a classification result, calculating an average value At of the queue N, counting a classification number in the queue C which is a label number count of pot bottom boiling, and obtaining a spectrogram through At and count; and S10, performing fusion judgment. According to the method, the pot bottom boiling degree is quantified by analyzing the time sequence change during pot bottom boiling in combination with a traditional image processing technology and a data analysis technology, and the result accuracy is enhanced in combination with an image classification algorithm.

Description

Technical Field

[0001] The present invention relates to the technical field related to intelligent detection and recognition, and specifically to a method for identifying boiling of a hot pot base based on time-series image analysis. Background Art

[0002] With the rapid development of intelligent detection and recognition technology, it is not widely used in the catering industry at present. In particular, there is no application or practical solution for the detection of boiling water at the bottom of the pot.

[0003] The currently used intelligent detection and recognition technology only analyzes the image at the current moment, but the boiling of the pot bottom is a continuous process, so directly applying mature intelligent detection and recognition methods does not have a good effect. Therefore, the present invention proposes a method for identifying the boiling of the hot pot bottom based on time series image analysis, which mainly combines the application of current intelligent detection technology in other directions, and optimizes it for the boiling of the pot bottom to achieve good results. Summary of the invention

[0004] The purpose of the present invention is to provide a method for identifying boiling of hot pot bottom based on time series image analysis to solve the problems raised in the above background technology.

[0005] To achieve the above object, the present invention provides the following technical solution: a method for identifying boiling hot pot bottom based on time series image analysis, comprising the following steps:

[0006] S1, obtain image data, obtain the image of the pot bottom area, and use the camera to obtain the image I at time t t , obtain the pot bottom area image at time t through the pot bottom area location information PotLocation marked in advance;

[0007] S2, image classification algorithm, scales the pot bottom area image at time t to 224×224 and transmits it to the image classification network to obtain category information C t ;

[0008] S3, convert to grayscale, and convert the pot bottom area image Roi t Convert to grayscale image Gray t ;

[0009] S4, frame difference, based on the grayscale image of the pot bottom area at the current moment and the previous moment, the difference between the two is obtained by using the frame difference method;

[0010] S5, binarization, performing a binarization operation on the obtained difference image to obtain a binarized image;

[0011] S6, opening operation denoising, performing morphological opening operation on the obtained binary image to remove noise and burrs;

[0012] S7, counting pixels, counting the number of white area pixels in the processed image;

[0013] S8, sliding average denoising, repeat the above steps S1-S7, when t≤1000, the category queue C={c1,c2,…,c k}, difference value queue N = {num1, num2, ..., num k}, where the upper limit of queue C and queue N is 1000; when t>1000, the results obtained by executing steps S1-S7 are stored at the tail of the queue, and the header data is removed;

[0014] S9, statistical classification results, calculate the average value A of queue N t , count the number of labels with the number of categories in queue C being the number of labels with the number of categories being ... t , where A t with count t The obtained spectrum diagram;

[0015] S10, fusion judgment, if count t >100 and A t >240 means the pot bottom is in boiling state. t >400 means the bottom of the pot is in a state of intense boiling.

[0016] Preferably, the image of the pot bottom area at time t in step S1 is Roi t =Crop(I t ,PotLocation).

[0017] Preferably, the image scaling in step S2 is Resize t =ImageResize(Roi t ), the image classification network obtains category information as C t =Classification(Resize t ).

[0018] Preferably, the grayscale image converted in step S3 is Gray t =CoverToGray(Roi t ).

[0019] Preferably, the difference between the grayscale image of the pot bottom area at the current moment and the previous moment in step S4 is D t =Diff(Gray t ,Gray t-1 ).

[0020] Preferably, the binary image in step S5 is Bt =Binary(D t ).

[0021] Preferably, in step S6, the morphological opening operation is performed on the binary image. t =MorphOpen(B t ).

[0022] Preferably, the number of pixels in the white area of ​​the image after statistical processing in step S7 is num t =Count(Mo t ).

[0023] Preferably, the white areas counted in step S7 are the main differences between the images.

[0024] Compared with the prior art, the present invention has the following beneficial effects:

[0025] 1. The algorithm used in the present invention analyzes the changes in the time series of boiling at the bottom of the pot, combines traditional image processing technology and data analysis technology to quantify the boiling degree of the bottom of the pot, and combines the image classification algorithm to enhance the accuracy of the results and distinguish the boiling degree;

[0026] 2. The present invention combines intelligent detection and recognition technology in the quantitative operation of the boiling degree of hot pot bottom in the catering industry, filling the application gap of intelligent detection and recognition technology in the catering industry. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 It is a flow chart of a hot pot bottom boiling recognition method based on time series image analysis of the present invention;

[0028] Figure 2 The present invention provides a boiling state spectrum diagram of a hot pot bottom obtained by a method for identifying boiling of a hot pot bottom based on time series image analysis. DETAILED DESCRIPTION

[0029] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0030] See also Figure 1-2 The present invention provides a technical solution: a method for identifying boiling hot pot bottom based on time series image analysis, comprising the following steps:

[0031] S1, obtain image data, obtain the image of the bottom area of ​​the pot, and use the camera to obtain the image I at time tt , obtain the pot bottom area image at time t through the pot bottom area location information PotLocation marked in advance;

[0032] S2, image classification algorithm, scales the pot bottom area image at time t to 224×224 and transmits it to the image classification network to obtain category information C t ;

[0033] S3, convert to grayscale, and convert the pot bottom area image Roi t Convert to grayscale image Gray t ;

[0034] S4, frame difference, based on the grayscale image of the pot bottom area at the current moment and the previous moment, the difference between the two is obtained by using the frame difference method. The implementation of the frame difference method is that the motion change area caused by the motion of the target pot bottom includes the common position of the moving target pot bottom boiling in the previous and next frames, the background area newly revealed in the current frame, and the background area newly covered;

[0035] S5, binarization, binarization operation is performed on the obtained difference image to obtain a binarized image. The binarization operation adopts a threshold method to analyze the shape and contour of the target pot bottom boiling image;

[0036] S6, opening operation denoising, performing morphological opening operation on the obtained binary image to remove noise and burrs. The opening operation is to corrode the obtained binary image first and then expand it. The denoising can adopt the sliding average algorithm of time series data dimensionality reduction denoising, or it can be replaced by a non-stationary time series detection algorithm;

[0037] S7, counting pixels, counting the number of white area pixels in the processed image;

[0038] S8, sliding average denoising, repeat the above steps S1-S7, when t≤1000, get the category queue C={c1,c2,…,c k}, difference value queue N = {num1, num2, ..., num k}, where the upper limit of queue C and queue N is 1000; when t>1000, the results obtained by executing steps S1-S7 are stored at the tail of the queue, and the header data is removed;

[0039] S9, statistical classification results, calculate the average value A of queue N t , count the number of labels with the number of categories in queue C being the number of labels with the number of categories being ... t , where A t with count t The obtained spectrum is as follows Figure 2 As shown;

[0040] S10, fusion judgment, if count t >100 and A t >240 means the pot bottom is in boiling state. t >400 means the bottom of the pot is in a strong boiling state;

[0041] The image of the pot bottom area at time t in step S1 is Roi t =Crop(I t ,PotLocation);

[0042] In step S2, the image is scaled to Resize t =ImageResize(Roi t ), the image classification network obtains category information as C t =Classification(Resize t );

[0043] The grayscale image converted to Gray in step S3 t =CoverToGray(Roi t );

[0044] The difference between the grayscale image of the pot bottom area at the current moment and the previous moment in step S4 is D t =Diff(Gray t ,Gray t-1 );

[0045] The binary image in step S5 is B t =Binary(D t );

[0046] In step S6, the morphological opening operation is performed on the binary image to obtain Mo t =MorphOpen(B t );

[0047] The number of pixels in the white area of ​​the image after statistical processing in step S7 is num t =Count(Mo t );

[0048] The white areas counted in step S7 are the main differences between the images.

[0049] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0050] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for identifying boiling hot pot bottom based on time series image analysis, characterized in that: The following steps are involved: S1, obtain image data, obtain the image of the bottom area of ​​the pot, and use the camera to obtain the image I at time t t , obtain the pot bottom area image at time t through the pot bottom area location information PotLocation marked in advance; S2, image classification algorithm, scales the pot bottom area image at time t to 224×224 and transmits it to the image classification network to obtain category information C t ; S3, convert to grayscale, and convert the pot bottom area image Roi t Convert to grayscale image Gray t ; S4, frame difference, based on the grayscale image of the pot bottom area at the current moment and the previous moment, the difference between the two is obtained by using the frame difference method; S5, binarization, performing a binarization operation on the obtained difference image to obtain a binarized image; S6, opening operation denoising, performing morphological opening operation on the obtained binary image to remove noise and burrs; S7, counting pixels, counting the number of white area pixels in the processed image; S8, sliding average denoising, repeat the above steps S1-S7, when t≤1000, the category queue C={c1,c2,…,c k }, difference value queue N = {num1, num2, ..., num k }, where the upper limit of queue C and queue N is 1000; when t>1000, the results obtained by executing steps S1-S7 are stored at the tail of the queue, and the header data is removed; S9, statistical classification results, calculate the average value A of queue N t , count the number of labels with the number of categories in queue C being the number of labels with the number of categories being ... t , where A t with count t The obtained spectrum diagram; S10, fusion judgment, if count t >100 and A t >240 means the pot bottom is in boiling state. t >400 means the bottom of the pot is in a state of intense boiling.

2. The method for identifying boiling point of hot pot base based on time series image analysis according to claim 1, characterized in that: In step S1, the image of the pot bottom area at time t is Roi t =Crop(I t ,PotLocation).

3. The method for identifying boiling point of hot pot base based on time series image analysis according to claim 1, characterized in that: In step S2, the image is scaled to Resize t =ImageResize(Roi t ), the image classification network obtains category information as C t =Classification(Resize t ).

4. The method for identifying boiling point of hot pot base based on time series image analysis according to claim 1, characterized in that: The grayscale image converted into Gray in step S3 is t =CoverToGray(Roi t ).

5. The method for identifying boiling point of hot pot base based on time series image analysis according to claim 1, characterized in that: The difference between the grayscale image of the pot bottom area at the current moment and the previous moment in step S4 is D t =Diff(Gray t ,Gray t-1 ).

6. The method for identifying boiling hot pot bottom based on time series image analysis according to claim 1, characterized in that: The binary image in step S5 is B t =Binary(D t ).

7. The method for identifying boiling hot pot bottom based on time series image analysis according to claim 1, characterized in that: In step S6, the morphological opening operation is performed on the binary image. t =MorphOpen(B t ).

8. The method for identifying boiling point of hot pot base based on time series image analysis according to claim 1, characterized in that: The number of pixels in the white area of ​​the image after statistical processing in step S7 is num t =Count(Mo t ).

9. The method for identifying boiling hot pot bottom based on time series image analysis according to claim 1, characterized in that: The white areas counted in step S7 are the main differences between images.

Images