Vending machine system

Abstract

The present application relates to the technical field of image processing, in particular to a vending machine system, comprising an image acquisition module for acquiring images of goods in the vending machine; an image transmission module for transmitting the images acquired by the image acquisition module to a management control module; a posture sensor for monitoring and calculating the opening and closing door angle of the vending machine and transmitting data to the management control module; a management control module for receiving and analyzing image data captured by the camera and opening and closing door angle information provided by the posture sensor to monitor the working state and goods display of the vending machine; a database for storing image data. The present application uses a camera control all-in-one machine to realize real-time monitoring and remote control of the vending machine state, the adjustable camera angle function meets different shooting needs, improves the flexibility and applicability of the system; the posture sensor combines with the accelerometer and gyroscope to ensure accurate monitoring and control of the vending machine state.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of image processing, and particularly to a vending machine system. BACKGROUND

[0002] A vending machine is an automatic retail device that utilizes advanced technology to complete the sales process without human intervention. This machine usually operates through a touch screen or buttons, allowing customers to browse product lists, select products, and make payments. Payment methods may include cash, credit cards, debit cards, or mobile payment platforms such as Alipay and WeChat Pay. Internet of Things technology enables vending machines to monitor inventory levels in real time, automatically restock, and maintain through remote management systems. Second, big data analysis allows businesses to collect consumer purchase data, gain insights into consumer habits, and optimize product mix and pricing strategies. In addition, the popularity of mobile payments and electronic money reduces transaction costs and improves transaction convenience. Traditional vending machine monitoring and control processes may be relatively simple, lacking remote monitoring capabilities, and users cannot immediately understand the inventory status or machine failures within the vending machine. At the same time, vending machines in different locations may face different shooting challenges, and fixed-angle cameras are difficult to adapt to all scenarios. SUMMARY

[0003] The main purpose of the present application is to provide a vending machine system to solve the problems of simple monitoring and control processes and camera limitations in the related art.

[0004] In order to achieve the above-mentioned purpose, according to one aspect of the present application, a vending machine system is provided, comprising:

[0005] An image acquisition module for acquiring images of goods in the vending machine;

[0006] An image transmission module for transmitting images acquired by the image acquisition module to the management control module;

[0007] A posture sensor for monitoring and calculating the opening and closing door angle of the vending machine and transmitting data to the management control module;

[0008] A management control module for receiving and analyzing image data captured by the camera and opening and closing door angle information provided by the posture sensor to monitor the working state and product display of the vending machine;

[0009] A database for storing image data.

[0010] Further, the database includes a real-time database and a historical database, the real-time database is used to store the images of the goods to be identified transmitted by the image transmission module; the historical database is used to store the images of the goods saved in advance.

[0011] Further, the image acquisition module is based on MIPI dual camera, the requirements of the camera of the MIPI dual camera include two kinds, the first kind is required to shoot the inside and outside of the cabinet; the second kind can only shoot the outside of the cabinet, the angle of the two lenses of the MIPI dual camera needs to meet the left side reference vertical line 15-30 degrees adjustable; the right side reference horizontal line 25-35 degrees adjustable.

[0012] Further, the attitude sensor includes an accelerometer and a gyroscope, the attitude sensor calculates the opening angle of the door of the vending machine based on a complementary filtering algorithm, and the calculation steps are as follows:

[0013] S1: obtain the tilt angle by using the accelerometer:

[0014] ;

[0015] wherein, is the tilt angle, is the y-axis component of the accelerometer; is the z-axis component of the accelerometer; is the noise term in the accelerometer data;

[0016] S2: obtain the angle change by using the gyroscope:

[0017] ;

[0018] wherein, is the angle estimate; is the angular velocity data; is the sampling time interval; is the noise term in the gyroscope data;

[0019] S3: the angle estimate after complementary filtering of the results of the accelerometer and the gyroscope:

[0020] ;

[0021] wherein, is the angle estimate after complementary filtering; is the filtering coefficient.

[0022] Further, the management control module monitors the working state of the vending machine, and the steps are as follows:

[0023] S4: set the angle threshold of the door opening and closing;

[0024] S5: according to the calculated angle information and the set threshold, judge whether the current door state is open or closed;

[0025] S6: according to the door state judgment result, realize the corresponding control logic;

[0026] S7: Trigger the opening or closing action of the vending machine door according to the control logic.

[0027] Furthermore, the management and control module includes an image processing module, an image recognition module, and a monitoring and analysis module;

[0028] The image processing module is used to preprocess the transmitted images;

[0029] The image recognition module is used to identify and judge the products in the image;

[0030] The monitoring and analysis module is used to monitor and analyze the working status of the vending machine and the product display.

[0031] Furthermore, the image recognition module uses the SIFT algorithm to recognize and judge the input image, and the steps are as follows:

[0032] S8: Extract SIFT features from the image, including detecting keypoints in the image, calculating gradient information of the regions around these keypoints, and generating a 128-dimensional SIFT descriptor for each keypoint;

[0033] S9: For all images to be matched, extract SIFT features and store these features to build a feature library containing SIFT features of all images;

[0034] S10: Compare each feature descriptor of the query image with all descriptors in the feature library;

[0035] S11: Finally, the matching results are visualized and displayed, showing the correspondence between the query image and the matching image, thereby achieving the purpose of image recognition and matching.

[0036] Furthermore, in S8, keypoint detection is based on the Difference-of-Gaussian (DOB) pyramid. The DOB pyramid finds extreme points in the image by performing Gaussian blurring and differencing operations on the image at different scales. The expression is:

[0037] ;

[0038] ;

[0039] in, These are the key points after Gaussian blur and difference operations; These are the layers of the Gaussian pyramid; The input is the image of the product to be detected; k is the scale factor; It is a scale parameter; is a Gaussian function; * indicates a convolution operation.

[0040] Furthermore, in S8, the gradient information of the region surrounding these keypoints is calculated. The gradient information includes the horizontal gradient, vertical gradient, gradient magnitude, and gradient direction, expressed as:

[0041] ;

[0042] ;

[0043] ;

[0044] ;

[0045] in, For horizontal gradient; The vertical gradient; As the key point; This represents the gradient magnitude of that pixel. This represents the gradient direction of that pixel. These are the key points of the input image of the product to be detected.

[0046] Furthermore, in S10, each feature descriptor of the query image is compared with all descriptors in the feature library based on Euclidean distance, expressed as:

[0047] Descriptor is ,in, This describes the i-th element in the subvector;

[0048] The distance between two descriptors is expressed as:

[0049] ;

[0050] in, These represent the descriptors of the input product image to be detected; This represents the corresponding feature descriptor in the feature library; This represents the i-th component in the input vector of the product image to be detected; The i-th component in the corresponding vector in the feature library; This represents the weight of the i-th component;

[0051] A threshold for the distance between descriptors and a threshold for the number of successful matches are set in advance. If the distance between a pair of feature descriptors is less than the set threshold, the two feature points are successfully matched. The number of successfully matched feature point pairs is counted. If the number exceeds a certain proportion or a certain number threshold, the two images are determined to be the same.

[0052] Compared with the prior art, the present invention has the following beneficial effects:

[0053] This invention utilizes an integrated camera and control unit, supporting remote monitoring and management. It enables real-time monitoring and remote control of vending machine status. The adjustable camera angle meets diverse shooting needs, enhancing system flexibility and applicability. An attitude sensor, combined with an accelerometer and gyroscope, employs a complementary filtering algorithm to calculate the door opening and closing angle, ensuring accurate monitoring and control of the vending machine's status. Furthermore, the SIFT algorithm is used for image recognition and matching, extracting key features and performing effective matching to achieve automatic product identification and judgment. This invention boasts enhanced intelligence, automation, and accurate recognition capabilities, effectively monitoring vending machine status, identifying products, improving service efficiency, and providing users with a more convenient shopping experience. Attached Figure Description

[0054] Fig. 1 This is a system block diagram of the entire invention;

[0055] Fig. 2 This is a system block diagram of the management and control module in this invention;

[0056] Illustration:

[0057] 1. Image acquisition module; 2. Image transmission module; 3. Attitude sensor; 4. Management and control module; 41. Image processing module; 42. Image recognition module; 43. Monitoring and analysis module; 5. Database; 51. Real-time database; 52. Historical database. Detailed Implementation

[0058] To further illustrate the technical means and effects of the present invention in achieving its intended purpose, the following detailed description of the specific implementation methods, structures, features, and effects of the present invention, in conjunction with the accompanying drawings and preferred embodiments, is provided below.

[0059] Please see Figs. 1-2 This embodiment provides an unmanned vending machine system, including:

[0060] Image acquisition module 1: Used to acquire images of products in the vending machine;

[0061] Image transmission module 2: Used to transmit images acquired by image acquisition module 1 to management and control module 4;

[0062] Attitude sensor 3: Used to monitor and calculate the opening and closing angle of the vending machine and transmit the data to the management and control module 4;

[0063] Management and control module 4: Used to receive and analyze image data captured by the camera and door opening and closing angle information provided by the attitude sensor 3, in order to monitor the working status of the vending machine and the product display.

[0064] Database 5: Used to store image data.

[0065] Database 5 includes a real-time database 51 and a historical database 52. The real-time database 51 is used to store images of goods that need to be identified, transmitted from the image transmission module 2; the historical database 52 is used to store images of goods that have been saved in advance.

[0066] Image acquisition module 1 is based on MIPI dual cameras. The requirements for the MIPI dual cameras include two types: the first is to capture both the inside and outside of the cabinet; the second is to capture only the outside of the cabinet. The angles of the two lenses of the MIPI dual cameras need to be adjustable from 15 to 30 degrees to the left reference vertical line and from 25 to 35 degrees to the right reference horizontal line.

[0067] Attitude sensor 3 includes an accelerometer and a gyroscope. Attitude sensor 3 calculates the opening and closing angle of the vending machine based on a complementary filtering algorithm. The calculation steps are as follows:

[0068] S1: Obtain the tilt angle using an accelerometer:

[0069] ;

[0070] in, The tilt angle, The y-axis component of the accelerometer; The z-axis component of the accelerometer; This refers to the noise component in the accelerometer data.

[0071] S2: Obtain angle changes using a gyroscope:

[0072] ;

[0073] in, This is an estimated angle value; This is angular velocity data; The sampling time interval; This refers to the noise component in the gyroscope data.

[0074] S3: Angle estimate after complementary filtering of the accelerometer and gyroscope results:

[0075] ;

[0076] in, This is the angle estimate after complementary filtering; These are the filter coefficients.

[0077] The steps for the management control module 4 to monitor the working status of the vending machine are as follows:

[0078] S4: Set the threshold angles for door opening and closing;

[0079] S5: Based on the calculated angle information and the set threshold, determine whether the current door is open or closed;

[0080] S6: Based on the state of the door, determine the result and implement the corresponding control logic;

[0081] S7: Trigger the opening or closing action of the vending machine door according to the control logic.

[0082] The management and control module 4 includes an image processing module 41, an image recognition module 42, and a monitoring and analysis module 43;

[0083] Image processing module 41 is used to preprocess the transmitted images;

[0084] Image recognition module 42 is used to identify and judge the goods in the image;

[0085] The monitoring and analysis module 43 is used to monitor and analyze the working status and product display of the vending machine.

[0086] Image recognition module 42 uses the SIFT algorithm to recognize and judge the input image. The steps are as follows:

[0087] S8: Extract SIFT features from the image, including detecting keypoints in the image, calculating gradient information of the regions around these keypoints, and generating a 128-dimensional SIFT descriptor for each keypoint;

[0088] S9: For all images to be matched, extract SIFT features and store these features to build a feature library containing SIFT features of all images;

[0089] S10: Compare each feature descriptor of the query image with all descriptors in the feature library;

[0090] S11: Finally, the matching results are visualized and displayed, showing the correspondence between the query image and the matching image, thereby achieving the purpose of image recognition and matching.

[0091] In S8, keypoint detection is based on the Difference-of-Gaussian (DOB) pyramid. The DOB pyramid finds extreme points in an image by performing Gaussian blurring and differencing operations on the image at different scales. The expression is:

[0092] ;

[0093] ;

[0094] in, These are the key points after Gaussian blur and difference operations; These are the layers of the Gaussian pyramid; The input is the image of the product to be detected; k is the scale factor; It is a scale parameter; is a Gaussian function; * indicates a convolution operation.

[0095] In S8, the gradient information of the region surrounding these keypoints is calculated. The gradient information includes the horizontal gradient, vertical gradient, gradient magnitude, and gradient direction, expressed as:

[0096] ;

[0097] ;

[0098] ;

[0099] ;

[0100] in, For horizontal gradient; The vertical gradient; As the key point; This represents the gradient magnitude of that pixel. This represents the gradient direction of that pixel. These are the key points of the input image of the product to be detected.

[0101] In S10, each feature descriptor of the query image is compared with all descriptors in the feature library based on Euclidean distance, expressed as:

[0102] Descriptor is ,in, This describes the i-th element in the subvector;

[0103] The distance between two descriptors is expressed as:

[0104] ;

[0105] in, These represent the descriptors of the input product image to be detected; This represents the corresponding feature descriptor in the feature library; This represents the i-th component in the input vector of the product image to be detected; The i-th component in the corresponding vector in the feature library; This represents the weight of the i-th component;

[0106] A threshold for the distance between descriptors and a threshold for the number of successful matches are set in advance. If the distance between a pair of feature descriptors is less than the set threshold, the two feature points are successfully matched. The number of successfully matched feature point pairs is counted. If the number exceeds a certain proportion or a certain number threshold, the two images are determined to be the same.

[0107] In this embodiment, the image acquisition module 1 acquires images of the products in the vending machine, and the image transmission module 2 transmits the acquired images to the image processing module 41. Simultaneously, the attitude sensor 3 detects and calculates the opening and closing angle of the vending machine's door and transmits the data to the monitoring and analysis module 43. The image processing module 41 preprocesses the transmitted images and stores the processed data in the real-time database 51. The image recognition module 42 compares the images in the real-time database 51 with the images in the historical database 52 to determine whether the products displayed for sale match the manufacturer's requirements. The final result is then transmitted to the monitoring and analysis module 43 to monitor the vending machine's operating status and ensure consistency with product sales. The specific method is as follows:

[0108] Attitude sensor 3 includes an accelerometer and a gyroscope. The selected attitude sensor is installed on the vending machine door to ensure it accurately detects the door's movement and angle changes. Using the data provided by the sensor, the opening and closing angle of the vending machine door is calculated through a complementary filtering algorithm. The complementary filtering algorithm is a commonly used method for fusing accelerometer and gyroscope data for attitude estimation. By fusing the accelerometer and gyroscope data, a more accurate and stable attitude estimation result can be obtained. The calculation steps are as follows:

[0109] First, the tilt angle is obtained using an accelerometer:

[0110] ;

[0111] in, The tilt angle, The y-axis component of the accelerometer; The z-axis component of the accelerometer; This refers to the noise component in the accelerometer data.

[0112] Then, the angle change is obtained using a gyroscope:

[0113] ;

[0114] in, This is an estimated angle value; This is angular velocity data; The sampling time interval; This refers to the noise component in the gyroscope data.

[0115] Finally, the angle estimate is obtained by performing complementary filtering on the results from the accelerometer and gyroscope:

[0116] ;

[0117] in, This is the angle estimate after complementary filtering; These are the filter coefficients. The filter coefficients are used to balance the weights of accelerometer and gyroscope data in attitude estimation. By continuously updating the value of θ and combining the accelerometer and gyroscope data, a more stable and accurate attitude estimation result can be obtained.

[0118] In practical applications, sensor data can be affected by noise. The noise term in accelerometer data is used to simulate the uncertainty and noise in accelerometer measurements. Accelerometers are typically used to measure gravitational acceleration and the acceleration of an object in space; therefore, they may be affected by vibration, shock, or other environmental factors, leading to a certain degree of noise in the measurements. Similarly, the noise term in gyroscope data is used to simulate the uncertainty and noise in gyroscope measurements. Gyroscopes are typically used to measure angular velocity, the speed at which an object rotates around its axes; gyroscope data may be affected by mechanical vibration, temperature changes, and other factors, resulting in a certain degree of noise in the measurements. Introducing a noise term into the complementary filtering algorithm can more effectively handle noise in sensor data, improving the accuracy and stability of attitude estimation.

[0119] Furthermore, the steps for monitoring the vending machine's operating status are as follows: First, pre-set the door opening and closing angle thresholds; based on the calculated angle information and the set thresholds, determine whether the current door state is open or closed; based on the door state determination result, implement the corresponding control logic; for example: if the door angle exceeds the opening threshold and the current door state is closed, trigger the door opening action; if the door angle is below the closing threshold and the current door state is open, trigger the door closing action. According to the control logic, trigger the vending machine door's opening or closing action; continuously monitor the vending machine door's angle information, and update the door state and execute corresponding control actions as needed.

[0120] The above steps enable real-time monitoring and control of the vending machine door status, ensuring that the door opens and closes as expected during operation, thereby improving system stability, security, and efficiency. Through refined door control logic, the system can promptly detect and handle anomalies, guaranteeing the normal operation of the vending machine and enhancing user experience and management efficiency.

[0121] Image acquisition module 1 is based on a MIPI dual-camera system. The MIPI dual-camera system has two camera requirements: first, it needs to capture both the inside and outside of the cabinet; second, it only needs to capture the outside. The angles of the two lenses in the MIPI dual-camera system need to be adjustable from 15-30 degrees to the left relative to the vertical line and from 25-35 degrees to the right relative to the horizontal line. In the MIPI dual-camera system, the raw image data captured by the cameras is directly transmitted to the main control CPU for processing, enabling rapid data processing without the need for additional processing chips. By using two cameras, coverage of different areas inside and outside the cabinet can be achieved. One camera focuses on capturing the items inside the cabinet, while the other captures the environment outside. This setup ensures information is acquired from multiple perspectives, thereby improving the overall monitoring capability and accuracy of the system. Simultaneously, the installation angles of the two cameras can be adjusted as needed to ensure that each camera can effectively capture a specific area. Adjusting the angle of the left camera relative to the vertical line ensures clear imaging of the items inside the cabinet, while adjusting the angle of the right camera relative to the horizontal line helps capture activities outside the cabinet.

[0122] After receiving the collected data, the image recognition module 42 performs recognition and judgment on the input image based on the SIFT algorithm. The steps are as follows: extract SIFT features from the image, including detecting key points in the image, calculating the gradient information of the region around these key points, and generating a 128-dimensional SIFT descriptor for each key point; extract SIFT features for all images to be matched and store these features to establish a feature library containing SIFT features of all images; compare each feature descriptor of the query image with all descriptors in the feature library; finally, visualize the matching results, showing the correspondence between the query image and the matching image, thereby achieving the purpose of image recognition and matching.

[0123] Among them, the keypoint detection is based on the Difference of Gaussian (DG) pyramid. The DG pyramid finds the extreme points in the image by performing Gaussian blurring and differencing operations on the image at different scales. The expression is:

[0124] ;

[0125] ;

[0126] in, Keypoints after Gaussian blur and difference operations are represented by feature response images obtained after Gaussian blur and difference operations on the original image at different scales. They are used to find extreme points in the image, i.e., potential keypoints. The layers of the Gaussian pyramid represent smoothed image layers obtained by performing Gaussian blur on the original image at a specific scale. Using different layers of the Gaussian pyramid to calculate the difference, instead of performing the difference operation directly on the original image, can improve the stability and accuracy of key points, enabling the SIFT algorithm to better detect key points in the image at different scales. The input is the image of the product to be detected; k is the scale factor, used to adjust the distance between different layers in the scale space, affecting the hierarchical structure of the pyramid; It is a scaling parameter used to control the standard deviation of the Gaussian function, affecting the degree of Gaussian blur and the feature scale; `<Gaussian>` is a Gaussian function used to perform a Gaussian blur operation on the image of the product to be detected; `*` indicates a convolution operation.

[0127] The gradient information includes the horizontal gradient, vertical gradient, gradient magnitude, and gradient direction, expressed as:

[0128] ;

[0129] ;

[0130] ;

[0131] ;

[0132] in, For horizontal gradient; The vertical gradient; As the key point; This represents the gradient magnitude of that pixel. This represents the gradient direction of that pixel. These are the key points of the input image of the product to be detected.

[0133] A SIFT descriptor is typically a 128-element vector used to describe the features of the region surrounding a keypoint. These elements are calculated based on gradient information around the keypoint. Each element represents the contribution of gradient information in a specific direction to the descriptor. Each feature descriptor of the query image is compared with all descriptors in the feature library based on Euclidean distance, expressed as:

[0134] Descriptor is ,in, This describes the i-th element in the subvector;

[0135] The distance between two descriptors is expressed as:

[0136] ;

[0137] in, These represent the descriptors of the input product image to be detected; This represents the corresponding feature descriptor in the feature library; This represents the i-th component in the input vector of the product image to be detected; The i-th component in the corresponding vector in the feature library; This represents the weight of the i-th component.

[0138] A threshold for the distance between descriptors and a threshold for the number of successful matches are set in advance. If the distance between a pair of feature descriptors is less than the set threshold, the two feature points are successfully matched. The number of successfully matched feature point pairs is then counted. If it exceeds a certain proportion or number threshold, the two images are determined to be the same, indicating that the items for sale in the cabinet are consistent with the items required by the cabinet manufacturer.

[0139] Simultaneously, by combining data from the attitude sensor 3, the inventory level of goods can be determined. For example, if the frequency of a certain product appearing in the image decreases, it may mean that the product is being consumed and needs to be replenished. Secondly, by analyzing the product images, it is possible to understand whether the product placement meets the optimal display effect, such as whether it is arranged in an orderly manner according to category or brand, and whether there are any issues such as obstruction, thereby guiding adjustments to the product display. Finally, the long-term collected image and attitude data can be used to analyze sales trends, predict which products are more popular, and which products may need to be promoted or removed from the shelves.

[0140] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some modifications or alterations to the above-disclosed technical content to create equivalent embodiments without departing from the scope of the present invention. Any indirect modifications, equivalent changes, and alterations made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the scope of the present invention.

Claims

1. A vending machine system, characterized by, Comprise: The image acquisition module (1) is based on MIPI dual camera, the requirements of the camera of MIPI dual camera include two, the two lens angles of MIPI dual camera need to meet left reference vertical line 15-30 ° adjustable, right reference horizontal line 25-35 ° adjustable, for collecting the image of the goods in the vending machine; Image transmission module (2): for transmitting the image collected by image acquisition module (1) to management control module (4); Attitude sensor (3) includes accelerometer and gyroscope, for monitoring and calculating the opening and closing angle of vending machine door and transmitting data to management control module (4), the attitude sensor (3) calculates the opening and closing angle of vending machine door based on complementary filtering algorithm, the calculation steps are as follows: S1: the inclination angle obtained by using accelerometer: ； wherein, is the tilt angle, is the y-axis component of the accelerometer; is the z-axis component of the accelerometer; is a noise term in the accelerometer data; S2: the angle change obtained by using gyroscope: ； wherein is an angle estimate value; is an angle estimate value of a previous time instant; is an angular velocity data; is a sampled time interval; is a noise term in the gyroscope data; S3: the angle estimation value after complementary filtering of the results of accelerometer and gyroscope: ； wherein is the angle estimate after complementary filtering; is the filter coefficient; Management control module (4) is used for receiving and analyzing image data captured by camera and opening and closing angle information provided by attitude sensor (3) to monitor the working state and commodity display of vending machine;And management control module (4) includes image processing module (41), image recognition module (42) and monitoring analysis module (43); Image processing module (41) is used for pre-processing the transmitted image; Image recognition module (42) is used for identifying and judging the goods in the image, wherein the image recognition module (42) identifies and judges the input image based on SIFT algorithm, and the steps are as follows: S8: extract SIFT features in the image, including detecting key points in the image, calculating gradient information of the region around the key points, and generating 128-dimensional SIFT descriptor for each key point; S9: for all images to be matched, extract SIFT features and store the features, and establish a feature library containing SIFT features of all images; S10: compare each feature descriptor of the query image with all descriptors in the feature library, the expression is: descriptor is wherein, denotes the element of the descriptor vector; The distance between two descriptors is expressed as: ； wherein, respectively represent the descriptors of the input image of the goods to be detected; represent the corresponding feature descriptors in the feature library; represent the first component in the input image vector of the goods to be detected; the first component in the corresponding vector in the feature library; represent the weight of the first component; S11: finally, the matching result is visualized and displayed, showing the corresponding relationship between the query image and the matching image, so as to realize the purpose of image recognition and matching; Monitoring analysis module (43) is used for monitoring and analyzing the working state and commodity display of vending machine, Database (5) includes real-time database (51) and historical database (52), real-time database (51) is used for storing the image of the goods to be identified transmitted by image transmission module (2);The historical database (52) is used for storing the image of the goods saved in advance.

2. The vending machine system of claim 1, wherein, Attitude sensor (3).

3. The vending machine system of claim 1, wherein, The steps of management control module (4) monitoring the working state of vending machine are as follows: S4: set the angle threshold of door opening and closing; S5: according to the calculated angle information and the set threshold, judge whether the current door is opened or closed; S6: according to the state judgment result of the door, realize the corresponding control logic; S7: according to the control logic, trigger the opening or closing action of vending machine door.

4. The vending machine system of claim 1, wherein, In S8, the key points are detected based on a Gaussian difference pyramid, which finds extreme points in the image by Gaussian blurring and difference operation on the image at different scales, and the expression is as follows: ； ； wherein, is the key point after Gaussian blur and difference operation; is the layer of the Gaussian pyramid; is the input image of the goods to be detected; k is the scale factor; is the scale parameter; is the Gaussian function; * denotes the convolution operation.

5. The vending machine system of claim 1, wherein, In S8, the gradient information of the region around the key points is calculated, and the gradient information includes horizontal gradient, vertical gradient, gradient amplitude and gradient direction, and the expression is as follows: ； ； ； ； wherein, is a horizontal gradient; is a vertical gradient; is a key point; is a gradient magnitude of a pixel point; is a gradient direction of a pixel point; is a key point of an input image of a commodity to be detected.

6. The vending machine system of claim 1, wherein, In S10, the threshold value of the distance between the descriptors and the threshold value of the number of successful matches are set in advance, the distance between a pair of feature descriptors is less than the set threshold value, then the two feature points are successfully matched, the number of successfully matched feature point pairs is counted, and if the number exceeds a certain proportion or quantity threshold value, it is determined that the two images are the same.

Images