Monitoring image processing method for refrigerator

Image enhancement processing is performed through the Hal wavelet transformation and bilateral filtering improved Retinex algorithm, which solves the problem of poor imaging effect of the refrigerator monitoring system, and realizes high-quality image display, avoiding the increase in the cost of hardware upgrades.

CN118921549APending Publication Date: 2024-11-08AUCMA
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411028142.9
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-30
Publication Date
2024-11-08

AI Technical Summary

Technical Problem

Due to the high humidity, low temperature and poor light environment, the existing refrigerator monitoring system has poor imaging effects, blurred pictures, and is limited by cost and space, making it difficult to effectively monitor the interior of the refrigerator.

Method used

The Hal wavelet transformation is used to separate the image information into low-frequency and high-frequency information, and the bilateral filtering is improved Retinex algorithm for processing, combined with wavelet reconstruction, image enhancement processing is performed, and displayed on the smart screen.

Benefits of technology

The image quality of internal monitoring of refrigerators is improved without hardware upgrades. It only needs to upgrade software algorithms, effectively control costs and achieve clearer image display.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118921549A_ABST
    Figure CN118921549A_ABST
Patent Text Reader

Abstract

The invention discloses a monitoring image processing method for a refrigerator, which is characterized in that a camera module is arranged in a refrigerator compartment, a smart screen is arranged outside the refrigerator, the camera module is in signal connection with the smart screen, the camera module performs image information conversion on shot sceneries, and then the converted image information is transmitted to the smart screen for image enhancement processing; the method specifically comprises the following steps: s1, separating image information based on Haar wavelet transform to obtain low-frequency information and high-frequency information; s2, processing the low-frequency information by using a bilateral filtering improved Retinex algorithm; s3, performing fuzzy enhancement processing on the high-frequency information; s4, performing wavelet reconstruction on the two parts of information processed in the steps s2 and s3 to obtain a complete image; and s5, displaying the image through a screen. According to the method, image enhancement processing is carried out by combining Haar wavelet transform and utilizing a bilateral filtering improved Retniex method, the image quality is good, in addition, hardware upgrading is not needed, only a software algorithm is upgraded, and the cost is effectively controlled.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of refrigerators, and in particular to a monitoring image processing method for refrigerators. Background Art

[0002] At present, there are monitoring refrigerators based on the Android operating system on the market. The camera module is installed in the refrigerator compartment and the display screen is embedded in the refrigerator door to monitor the interior of the refrigerator in real time. Since the humidity inside the refrigerator is high, the temperature is low, and the light is poor, and monitoring the real-time image is not the main function of the refrigerator, the camera module will be limited by factors such as cost and placement space, resulting in poor imaging and blurred images.

[0003] In view of this, it is necessary to improve the existing technology. Summary of the invention

[0004] The object of the present invention is to provide a monitoring image processing method for a refrigerator, aiming to solve at least one of the technical problems existing in the above-mentioned prior art. To achieve the above-mentioned object, the technical solution adopted is as follows:

[0005] A monitoring image processing method for a refrigerator, wherein a camera module is arranged inside the refrigerator compartment, and a smart screen is arranged outside the refrigerator, the camera module is connected to the smart screen signal, the camera module converts the image information of the captured scene, and then transmits it to the smart screen for image enhancement processing, specifically comprising the following steps:

[0006] s1. Separate the image information into low-frequency information and high-frequency information based on Haar wavelet transform;

[0007] s2. Low-frequency information is processed using the Retinex algorithm improved by bilateral filtering;

[0008] s3. Fuzzy enhancement processing of high-frequency information;

[0009] s4. Reconstruct the two parts of information processed in steps s2 and s3 by wavelet to obtain a complete image;

[0010] s5. Display images on the screen.

[0011] Preferably, the camera module includes a lens, an image sensor and a digital signal processing chip.

[0012] Preferably, in step s1, the two-dimensional image is divided into four sub-images after being processed by Haar wavelet transform:

[0013] L i f(x,y)= <f(x,y),φ j,n (x)φ j,m (y)>

[0014]

[0015]

[0016]

[0017] Where, ψ: wavelet function; : scaling function; L i f(x,y): low-frequency information; High-frequency information of horizontal edges; High-frequency information of vertical edges; High frequency information in the diagonal direction.

[0018] Preferably, step s2 includes the following specific steps:

[0019] s2.1. Image i f9x,y) is decomposed into two different images: the reflected image R x,y and the incident image L x,y, which can be expressed by the following formula: i f(x,y)=S(x,y)=R(x,y)·L(x,y);

[0020] Group the colors contained in the image according to the three colors of red, green and blue, calculate the pixel value of each group of images, and convert them to the logarithmic domain;

[0021] s2.2.rgb The three channels are processed once respectively to get the r(x,y) of the three channels. The formula used is:

[0022]

[0023] Where, *: convolution symbol; λ: weight sum of each pixel value in the filter window, used for weight normalization;

[0024] s2.3. The information of the three RGB channels R(x,y) is merged into an RGB low-frequency image R(x,y), and the logarithmic domain R(x,y) is converted into the real domain R(x,y) image information through the EXP function;

[0025] s2.4. Linearly stretch the output image R(x,y), and display the output in a corresponding format to finally obtain the processed low-frequency information.

[0026] Preferably, in step s2.2, it is found through experiments that when the filter radius = 1, the global variance = 1.6, and the local variance = 0.01, the effect after bilateral filtering is good.

[0027] Preferably, step s3 includes the following specific steps:

[0028] s3.1. Calculate the local contrast using the formula:

[0029] Where, D(x,y): local contrast; A max : Maximum gray value; A min : minimum gray value;

[0030] s3.2. Calculation of membership η i (x,y),η i (i=1,2,3, η1: low detail region membership; η2: medium detail region membership η3: high detail region membership, calculate the size of the membership of the three regions respectively, using the formula:

[0031] η i (x,y)=4sigm{α t [D(x,y)-β t ]}×sigm{-α t [D(x,y)-β t ];

[0032] Where, t=1,2,3; D(x,y)∈[0,1];

[0033] s3.3. Calculate the enhancement coefficient. Use the membership degree obtained in step s3.2. as the weight to calculate the enhancement coefficient of each point. The formula used is:

[0034]

[0035] s3.4. Calculate the enhanced high frequency subband map using the formula:

[0036]

[0037] Enhanced wavelet coefficients; f(x,y): original wavelet coefficients;

[0038] After the above steps, the processed subband graphs of each high frequency component are obtained.

[0039] Preferably, in step s3.3, it is found through experiments that when λ low =λ mid =1.5,λ high When =3, image enhancement can achieve an ideal effect.

[0040] Preferably, in step s4, wavelet reconstruction is an inverse process of wavelet transform.

[0041] Preferably, in step s5, an imaging app is pre-installed in the smart screen, and the optimized image is displayed through the imaging app.

[0042] Compared with the prior art, the present invention has the following beneficial effects:

[0043] The present invention discloses a monitoring image processing method for a refrigerator. The method combines Haar wavelet transform with a Retniex method improved by bilateral filtering to perform image enhancement processing. The imaging quality is good. In addition, no hardware upgrade is required, and only the software algorithm is upgraded, thereby effectively controlling the cost. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] In order to more clearly illustrate the technical solution of the present invention, the drawings required for use in the implementation mode will be briefly introduced below. Obviously, the drawings described below are only some implementation modes of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0045] Figure 1 It is a flow chart of the image formation process.

[0046] Figure 2 It is a flow chart of the image enhancement process.

[0047] Figure 3 It is the Retinex schematic diagram.

[0048] Figure 4 This is a flow chart of the Retinex processing process. DETAILED DESCRIPTION

[0049] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the accompanying drawings in the embodiments of the present invention.

[0050] A preferred embodiment of the present invention provides a monitoring image processing method for a refrigerator.

[0051] like Figure 1 As shown, a camera module is installed in the refrigerator compartment, and the camera module includes a lens, an image sensor and a digital signal processing chip. A smart screen is installed outside the refrigerator, and the camera module is connected to the smart screen signal. The scene generates an optical image through the lens, and is projected onto the surface of the image sensor, converted into an electrical signal, and converted into a digital image signal after A / D (digital-to-analog conversion), and then processed in a digital signal processing chip (DSP), and then transmitted to the smart screen for image enhancement processing.

[0052] Specifically, Figure 2 As shown, the image enhancement processing method includes the following steps:

[0053] s1. Based on Haar wavelet transform, the image information is separated into low-frequency information and high-frequency information.

[0054] There are two main reasons for choosing Haar wavelet: first, the amount of calculation required for Haar wavelet transform is relatively small, and the fast algorithm is relatively easy to implement; second, Haar wavelet transform is reversible, which can ensure that the compressed data can be accurately restored to the original data.

[0055] Specifically, the two-dimensional image is divided into four sub-images after being processed by Haar wavelet transform:

[0056] L i f(x,y)= <f(x,y),φ j,n (x)φ j,m (y)>

[0057]

[0058]

[0059]

[0060] Where, ψ: wavelet function; : scaling function; L i f(x,y): low-frequency information; High-frequency information of horizontal edges; High-frequency information of vertical edges; High frequency information in the diagonal direction.

[0061] s2. The low-frequency information is processed using the Retinex algorithm improved by bilateral filtering. The low-frequency information is processed by using the bilateral filtering combined with the Retinex algorithm. Bilateral filtering is used to replace the original Gaussian filtering, thereby retaining the detail information to a large extent.

[0062] The basic theory of Retinex theory can be understood as the color of objects seen by the naked eye is determined by the object's ability to reflect three wavelengths of light (i.e., red, green, and blue). It has almost nothing to do with the intensity of the light reflected by the object. The color of the target object is consistent and is not affected by lighting unevenness. That is, Retinex is based on color consistency.

[0063] s2.1. Image i f(x,y) is decomposed into two different images: the reflected image R x,y and the incident image L x,y, as Figure 3 As shown, the original image can be regarded as composed of a reflected image and an incident image. After the incident light hits the surface of the object, the reflected light can be reflected to the human eye (i.e., the reflected image). Finally, the object will be imaged on the human retina. The final image seen by the human eye can be expressed by the following formula: i f(x,y)=S(x,y(=R(x,y)·L(x,y)).

[0064] The colors contained in the image are grouped according to the three colors of red, green and blue, the pixel values ​​of each group of images are calculated, and converted to the logarithmic domain.

[0065] s2.2.rgb The three channels are processed once respectively to get the r(x,y) of the three channels. The formula used is:

[0066]

[0067] Where, *: convolution symbol; λ: weight sum of each pixel value in the filter window, used for weight normalization.

[0068] Through experiments, it is found that when the filter radius = 1, the global variance = 1.6, and the local variance = 0.01, the effect of bilateral filtering is good.

[0069] The information of the three RGB channels R(x,y) is merged into an RGB low-frequency image R(x,y), and the logarithmic domain R(x,y) is converted into real domain R(x,y) image information through the EXP function.

[0070] The basic idea of ​​Retinex theory is to eliminate or reduce the impact of the incident image on the original image by some method, so as to retain the image of the object's essential reflective properties as much as possible. From a mathematical point of view, solving R(x,y) is a singular problem and can only be calculated by approximate mathematical methods. The processing process is as follows: Figure 4 shown.

[0071] s2.4. Linearly stretch the output image R(x,y), and display the output in a corresponding format to finally obtain the processed low-frequency information.

[0072] In the above method, the Gaussian filter in the traditional Retinex algorithm is replaced by a bilateral filter. The bilateral filter is a nonlinear filter that uses a weighted average method. Its principle is to multiply two Gaussian functions, one of which is related to the spatial distance, and the other is related to the grayscale distance. It has the advantages of maintaining edges, reducing noise, and smoothing. The intensity of a pixel is represented by the weighted average of the pixels around it. The most important point is that the bilateral filter takes into account the Euclidean distance of the pixel, while the ordinary Gaussian low-pass filter only considers the influence of the position on the central pixel, and also takes into account the radiation difference. These two weights are considered when calculating the central pixel. The closer the pixel values ​​are, the greater the weight.

[0073] s3. The high-frequency information is subjected to fuzzy enhancement processing, which specifically includes the following steps:

[0074] s3.1. Calculate the local contrast using the formula:

[0075] Where, D(x,y): local contrast; A max : Maximum gray value; A min : Minimum grayscale value.

[0076] s3.2. Calculation of membership η i (x,y),η i (i=1,2,3, η1: low detail region membership; η2: medium detail region membership η3: high detail region membership, calculate the size of the membership of the three regions respectively, using the formula:

[0077] η i (x,y)=4sigm{α t [D(x,y)-β t ]}×sigm{-α t [D(x,y)-β t ];

[0078] Where, t=1,2,3; D(x,y)∈[0,1];

[0079] s3.3. Calculate the enhancement coefficient. Use the membership degree obtained in step s3.2. as the weight to calculate the enhancement coefficient of each point. The formula used is:

[0080]

[0081] Through experiments, it is found that when λ low =λ mid =1.5,λ high When =3, image enhancement can achieve an ideal effect.

[0082] s3.4. Calculate the enhanced high frequency subband map using the formula:

[0083]

[0084] Enhanced wavelet coefficients; f(x,u): original wavelet coefficients;

[0085] After the above steps, the processed subband graphs of each high frequency component are obtained.

[0086] s3. The two parts of information processed in steps s2 and s3 are reconstructed by wavelet to obtain a complete image.

[0087] Image reconstruction is to fully re-present the image. Since the image decomposition adopts wavelet transform for decomposition, the wavelet reconstruction method is used here to reconstruct the image. Wavelet reconstruction is the inverse process of wavelet transform. The low-frequency information processed by the improved Retinex algorithm and the various high-frequency information after processing are reconstructed by wavelet transform to obtain the final dehazed and enhanced image.

[0088] s5. Displaying images through the screen. Specifically: Huiping is pre-installed with an imaging app, and the optimized image is displayed through the imaging app and presented to the user.

[0089] In summary, the monitoring image processing method for a refrigerator described in an embodiment of the present invention combines Haar wavelet transform and the Retniex method improved by bilateral filtering to perform image enhancement processing, and has good imaging quality. In addition, there is no need to upgrade hardware, only the software algorithm is upgraded, which effectively controls the cost.

[0090] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A monitoring image processing method for a refrigerator, wherein a camera module is provided inside the refrigerator compartment, a smart screen is provided outside the refrigerator, and the camera module is connected to the smart screen signal, characterized in that: The camera module converts the image information of the captured scene and then transmits it to the smart screen for image enhancement processing, which specifically includes the following steps: s1. Separate the image information into low-frequency information and high-frequency information based on Haar wavelet transform; s2. Low-frequency information is processed using the Retinex algorithm improved by bilateral filtering; s3. Fuzzy enhancement processing of high-frequency information; s4. Reconstruct the two parts of information processed in steps s2 and s3 by wavelet to obtain a complete image; s5. Display images on the screen.

2. A monitoring image processing method for a refrigerator according to claim 1, characterized in that: The camera module includes a lens, an image sensor and a digital signal processing chip.

3. The monitoring image processing method for a refrigerator according to claim 1, characterized in that: In step s1, the two-dimensional image is divided into four sub-images after being processed by Haar wavelet transform: L i f(x,y)=<f(x,y),φ j,n (x)φ j,m (y)>; Where, ψ: wavelet function; φ: scaling function; L i f(x,y): low-frequency information; High-frequency information of horizontal edges; High-frequency information of vertical edges; High frequency information in the diagonal direction.

4. A monitoring image processing method for a refrigerator according to claim 3, characterized in that: Step s2 includes the following specific steps: s2.

1. Image L i f(x,y) is decomposed into two different images: the reflected image R(x,y) and the incident image L(x,y), which can be expressed as follows: i f(x,y)=S(x,y)=R(x,y)·L(x,y); Group the colors contained in the image according to the three colors of red, green and blue, calculate the pixel value of each group of images, and convert them to the logarithmic domain; s2.2.rgb The three channels are processed once respectively to get the r(x,y) of the three channels. The formula used is: Where, *: convolution symbol; λ: weight sum of each pixel value in the filter window, used for weight normalization; s2.

3. The information of the three RGB channels R(x,y) is merged into an RGB low-frequency image R(x,y), and the logarithmic domain R(x,y) is converted into the real domain R(x,y) image information through the EXP function; s2.

4. Linearly stretch the output image R(x,y), and display the output in a corresponding format to finally obtain the processed low-frequency information.

5. A monitoring image processing method for a refrigerator according to claim 4, characterized in that: In step s2.2, it is found through experiments that when the filter radius = 1, the global variance = 1.6, and the local variance = 0.01, the effect after bilateral filtering is good.

6. The monitoring image processing method for a refrigerator according to claim 3, characterized in that: Step s3 includes the following specific steps: s3.

1. Calculate the local contrast using the formula: Where, D(x,y): local contrast; A max : Maximum gray value; A min : minimum gray value; s3.

2. Calculation of membership η i (x,y),η i (i=1,2,3), η1: low detail region membership; η2: medium detail region membership η3: high detail region membership, calculate the size of the membership of the three regions respectively, using the formula: or i (x,y)=4sigm{α t [D(x,y)-β t ]}×sigm{-α t [D(x,y)-β t ]; Where, t=1,2,3; D(x,y)∈[0,1]; s3.

3. Calculate the enhancement coefficient. Use the membership degree obtained in step s3.

2. as the weight to calculate the enhancement coefficient of each point. The formula used is: s3.

4. Calculate the enhanced high frequency subband map using the formula: Enhanced wavelet coefficients; f(x,y): original wavelet coefficients; After the above steps, the processed subband graphs of each high frequency component are obtained.

7. A monitoring image processing method for a refrigerator according to claim 6, characterized in that: In step s3.3, it is found through experiments that when λ low =λ mid =1.5,λ high When =3, image enhancement can achieve an ideal effect.

8. The monitoring image processing method for a refrigerator according to claim 1, characterized in that: In step s4, wavelet reconstruction is the inverse process of wavelet transform.

9. The monitoring image processing method for a refrigerator according to claim 1, characterized in that: In step s5, the smart screen is pre-installed with an imaging app, and the optimized image is displayed through the imaging app.