Jewelry monitoring system based on characteristics

Abstract

The invention provides a jewelry monitoring system based on features. The jewelry monitoring system comprises a data module, an initial portrait module, a feature acquisition module, a portrait risk module and a hierarchical supervision module, the data module comprises feature data used for establishing a portrait, a portrait period and threshold data used for evaluating the portrait; wherein the feature data comprises a feature category, a feature value and a feature weight, and the threshold data comprises a feature difference threshold and a portrait change trend threshold; the initial portrait module can establish an initial portrait as an evaluation standard through the initial feature data; the feature acquisition module can acquire features according to a set portrait period and establish a monitoring portrait through the acquired features; the portrait risk module can obtain risk features of the monitored portrait; and the supervision module carries out hierarchical supervision on the generated risk characteristics. The data screening efficiency is improved, and the model accuracy is improved.

Description

Technical Field

[0001] The present application relates to the field of jewelry monitoring, and in particular to a feature-based jewelry monitoring system. Background Art

[0002] Jewelry feature portrait is an intelligent module that combines deep learning, computer vision and big data analysis technology. It aims to build a comprehensive and accurate feature portrait for jewelry products through multi-dimensional data analysis and feature extraction. This system provides data support for jewelry merchants to track the status of products at any time, including inventory management before sales and maintenance management after sales. In the existing technology, deep learning and computer vision technology can be used to extract rich features from multi-source information such as pictures and videos of jewelry products, including shape, color, material, craftsmanship, etc. Through big data analysis, these features are quantified and classified to provide a data basis for subsequent portrait construction. However, due to the rich variety of features, a large amount of feature data is prone to redundancy. When the monitoring system is monitoring, collection deviations often occur due to external reasons such as environment and light. When monitoring data, a large number of false alarms are often generated, resulting in system data redundancy, slow processing speed, and large model deviation. At the same time, it is also very cumbersome to manually verify this information. Therefore, the current monitoring method uses qualitative models to achieve monitoring, which has a strong dependence and poor versatility. Therefore, a targeted method is needed to quickly identify risk features, classify them quickly and handle them accordingly to eliminate false alarms. Summary of the invention

[0003] The following description is used to disclose the present invention so that those skilled in the art can implement the present invention. The preferred embodiments described below are only examples, and those skilled in the art can think of other obvious variations. The basic principles of the present invention defined in the following description can be applied to other embodiments, variations, improvements, equivalents, and other technical solutions that do not deviate from the spirit and scope of the present invention.

[0004] In order to solve the technical problem, the present application provides a feature-based jewelry monitoring system, including a data module, an initial portrait module, a feature acquisition module, a portrait risk module and a hierarchical supervision module;

[0005] The data module includes feature data for establishing a portrait, a portrait cycle, and threshold data for evaluating the portrait; wherein the feature data includes feature category, feature value, and feature weight, and the threshold data includes feature difference threshold and portrait change trend threshold;

[0006] The initial portrait module can establish an initial portrait as an evaluation standard through initial feature data;

[0007] The feature acquisition module can acquire features at a set portrait cycle and establish a monitoring portrait through the acquired features;

[0008] The portrait risk module can use the initial portrait as an evaluation standard to obtain the risk characteristics of the monitoring portrait;

[0009] The supervision module evaluates the monitoring profiles that generate risk characteristics and their changing trends, and performs graded supervision based on the obtained evaluation results, wherein the graded supervision includes weak risk supervision and strong risk supervision.

[0010] In the initial portrait module, it is set that jewelry has n features QA, where the i-th feature QA i The initial eigenvalue is W i0 , the feature weight is ω i , the portrait score U0 of the initial portrait can be obtained;

[0011] Among them, μ is the adjustment coefficient (0.7≤μ≤1).

[0012] Among them, in the feature acquisition module, the features are acquired with the set initial image cycle △T0, and the monitoring time T after the end of the jth image cycle j Get the jth monitoring image PA j Image score U j ;

[0013] Where W ij For Feature QA i In T j The characteristic value of the monitoring time, ω i For QA i The feature weights of .

[0014] Among them, in the portrait risk module, according to the feature QA i In T j The characteristic value W of the monitoring time ij and its initial eigenvalue W i0 , get feature QA i The characteristic difference △W generated in the jth image cycle ij =W ij -W i0 ;

[0015] According to the characteristic difference threshold △W i0 And the image change trend threshold E0, for feature QA i To conduct an assessment:

[0016] When △W ij <△W i0 When judging feature QAi normal;

[0017] When △W ij ≥△W i0 When the feature QA i As the risk feature of the j-th image period, its feature value W ij Its monitoring time T j Recorded in risk feature list AF.

[0018] Among them, in the hierarchical supervision module, the risk feature QA in the risk feature list AF i , according to the jth monitoring image PA j Image score U j The portrait score U0 compared with the initial portrait;

[0019] When U j When ≥U0, weak risk supervision is implemented;

[0020] When U j When <U0, QA according to risk characteristics i The characteristic difference △W in the jth image period ij , get the image change trend in the jth image period

[0021] When E1≥E0, strong risk supervision is implemented;

[0022] When E1<E0, weak risk supervision is implemented.

[0023] Among them, the supervision mode of weak risk supervision is:

[0024] At the j+1th monitoring time T j+1 Get the j+1th image PA of jewelry A j+1 , T j+1 =T j +△T0;

[0025] Among them, we get the i-th feature QA i At the j+1th monitoring time T j+1 The eigenvalue W i(j+1) , and the j+1th image PA of jewelry A j+1 Image score

[0026] Get the characteristic difference △W in the j+1th image period i(j+1) =W i(j+1) -W ij ;

[0027] When △W i(j+1) <△W ij When the

[0028] When △W i(j+1) ≥△W ij When the time comes, it will switch to strong risk supervision.

[0029] Among them, the supervision mode of strong risk supervision is:

[0030] Get risk feature list AF QA i The number of occurrences num, get QA i The risk distribution θ i =num / j;

[0031] Adjust the image period to △T1, where:

[0032] At the j+1th monitoring time T j+1 Get the j+1th image PA j+1 , T j+1 =T j +△T1;

[0033] Among them, QA i At the j+1th monitoring time T j+1 The eigenvalue W i(j+1) , and the j+1th image PA j+1 Image score

[0034] Get the characteristic difference △W in the j+1th image period i(j+1) =W i(j+1) -W ij ;

[0035] When △W i(j+1) <△W ij When the risk is too high, it will be switched to weak risk supervision;

[0036] When △W i(j+1) ≥△W ij While maintaining strong risk supervision.

[0037] Set the risk distribution threshold θ max , when QA i The risk distribution θ i ≥θ max When QA i Issue early warnings and push risk-related information to servers and terminals.

[0038] The beneficial effects achieved by this application are as follows:

[0039] In the process of monitoring jewelry by establishing a portrait based on features, this application can screen out a large amount of error data caused by collection deviation, light influence, etc. through risk classification supervision, thereby improving the processing speed of calculation, correcting the accuracy of the model, and saving computing power and manpower. At the same time, it can quickly check risk features in a targeted manner and quickly classify and handle them accordingly. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present application. For those skilled in the art, other drawings can also be obtained based on these drawings.

[0041] Figure 1 This is a flow chart of the monitoring steps of the feature-based jewelry monitoring system of this application. DETAILED DESCRIPTION

[0042] Below, in conjunction with the drawings in the embodiments of the present application, the technical solutions in the embodiments of the present application are clearly and completely described. Obviously, the described embodiments are part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without making creative work are within the scope of protection of the present application. It should be noted that the terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described here can be implemented in an order other than those illustrated or described here. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0043] The present application provides a feature-based jewelry monitoring system, including a data module, an initial portrait module, a feature acquisition module, a portrait risk module and a graded supervision module; the data module includes feature data for establishing a portrait, a portrait cycle and threshold data for evaluating the portrait; wherein the feature data includes feature categories, feature values ​​and feature weights, and the threshold data includes feature difference thresholds and portrait change trend thresholds; the initial portrait module can establish an initial portrait as an evaluation standard through the initial feature data; the feature acquisition module can acquire features in a set portrait cycle and establish a monitoring portrait through the acquired features; the portrait risk module can use the initial portrait as an evaluation standard to obtain the risk features of the monitoring portrait; the supervision module evaluates the monitoring portrait that generates risk features and its change trend, and performs graded supervision based on the obtained evaluation results, wherein the graded supervision includes weak risk supervision and strong risk supervision.

[0044] Jewelry feature extraction usually refers to identifying and analyzing specific properties of jewelry from images or physical samples. Jewelry feature extraction is a complex and delicate process that requires the comprehensive use of multiple technical means and expertise. Through the comprehensive extraction and analysis of the visual, physical, technical and comprehensive features of jewelry, a comprehensive understanding and accurate evaluation of jewelry can be achieved. This process involves computer vision technology and machine learning algorithms. Jewelry features may include shape, color, texture, gloss, inlay craftsmanship and material. The feature extraction steps are: denoising, contrast enhancement and size normalization of jewelry images to obtain data foundation. At the same time, color information such as color distribution, hue and saturation are extracted using color space, and surface structure is reflected by calculating local binary patterns (LBP), gray level co-occurrence matrix (GLCM) or other texture descriptors, and geometric properties such as contours, edges, and pi ratios are measured. Convolutional neural networks (CNNs), such as ResNet and VGG, are used to learn features of jewelry images and automatically extract high-level features. Among them, jewelry feature extraction involves detailed observation and quantitative analysis of multiple aspects of jewelry, including visual feature extraction, physical feature extraction and technical feature extraction. Among them, in visual feature extraction, shape feature extraction includes the recognition of jewelry edges and quantitative description of shapes, color feature extraction includes the recognition of jewelry's main color, color saturation and brightness; texture feature extraction includes the analysis and quantification of jewelry surface patterns, texture directions and frequencies; gloss feature extraction includes the evaluation of jewelry surface reflectivity, gloss uniformity and gloss changes. Physical feature extraction includes the extraction of weight, size, density and hardness. Technical feature extraction includes material spectrum analysis and cut surface quality analysis.

[0045] Specifically, in the initial portrait module of one embodiment, an initial portrait PA0 is created for jewelry A, and the initial portrait PA0 is used as the evaluation basis of jewelry A, wherein the initial portrait PA0 includes n features QA, where QA = [QA1, QA2, QA3, ..., QA n ], where the i-th feature is QA i , QA i The feature weight is ω i , where the feature QA is normalized to obtain the quantized value of the feature of jewelry A, thereby obtaining the i-th feature QA i The initial eigenvalue W i0 , and then according to the i-th feature QA i The feature weight ω i , we can get the image score U0 of the initial image PA0. Specifically, Where, μ is the adjustment coefficient (0.6≤μ≤1)

[0046] According to the material and craftsmanship of jewelry A, the image period of jewelry A is set to △T0;

[0047] According to the texture, grade and detection cycle of jewelry A, set the feature difference threshold △W i0 , image change threshold E0;

[0048] The feature acquisition module is set at the first monitoring time T1 after a portrait cycle △T0, and obtains the first portrait PA1 of jewelry A, wherein the i-th feature QA is obtained i The characteristic value W at the first monitoring time T1 i1 , and the image score of the first image PA1 of jewelry A

[0049] The profile risk module is based on the i-th feature QA i The characteristic value W at the first monitoring time T1 i1 and the i-th feature QA i The initial eigenvalue W i0 , get the characteristic difference △W in the first image period i1 =W i1 -W i0 ;

[0050] When △W i1 ≥△W i0 When QA represents the i-th feature of jewelry A i In the first profiling cycle, a risk is generated and the i-th feature QA i As the risk feature of the first image period, its feature value W i1 and its monitoring time T1 are recorded in the risk characteristic list AF of Jewelry A;

[0051] Among them, when U1≥U0, that is, when the image score U1 of the first image PA1 can meet the standard, it means that the overall situation of jewelry A is relatively stable and the risk level is relatively weak, and jewelry A is subject to weak risk supervision;

[0052] Among them, when U1<U0, that is, when the image score U1 of the first image PA1 does not meet the standard, according to the risk feature QA i The characteristic difference △W in the first image period i1 , get the image change trend in the first image cycle

[0053] When E1≥E0, it means that the average change trend of the portrait in the first period is large, the risk level is high, and Jewelry A is subject to strong risk supervision steps;

[0054] When E1<E0, it means that the average change trend of the portrait in the first period is small, the risk level is low, and Jewelry A is subject to weak risk supervision steps;

[0055] Among them, the steps of weak risk supervision are as follows:

[0056] After a first monitoring time T1 and a portrait cycle △T0, a second monitoring time T2 is obtained to obtain a second portrait PA2 of jewelry A, wherein the i-th feature QA is obtained. i The characteristic value W at the second monitoring time T2 i2 , and the image score of the second image PA2 of jewelry A

[0057] Get the characteristic difference △W in the second image period i2 =W i2 -W i1 ;

[0058] When △W i2 <△W i1 When QA represents the i-th feature of jewelry A i In the second profiling cycle, the risk has been reduced, and supervision will continue with weak risk supervision steps;

[0059] When △W i2 ≥△W i1 When QA represents the i-th feature of jewelry A i The second profiling cycle increases the risk level on the basis of the first profiling cycle and supervises with strong risk supervision steps;

[0060] Among them, the process steps of strong risk supervision are:

[0061] Adjust the image period of jewelry A to △T1, where:

[0062] Where θ is the risk feature list QA i The proportion of occurrence times to the total monitoring times;

[0063] After a first monitoring time T1 and a portrait cycle △T1, a second monitoring time T2 is obtained to obtain a second portrait PA2 of jewelry A, wherein the i-th feature QA is obtained. i The characteristic value W at the second monitoring time T2 i2 , and the image score of the second image PA2 of jewelry A

[0064] Get the characteristic difference △W in the second image period i2 =W i2 -W i1 ;

[0065] When △W i2 <△W i1 When QA represents the i-th feature of jewelry A i In the second profiling cycle, the risk is reduced and supervision is switched to weak risk supervision steps;

[0066] When △W i2 ≥△W i1 When QA represents the i-th feature of jewelry A i In the second profiling cycle, the risk level is increased on the basis of the first profiling cycle, and supervision continues with strong risk supervision steps;

[0067] Set the risk distribution threshold θ max , when QA i The risk distribution θ i ≥θ max When QA i Provide early warning and push risk-related information to servers and terminals;

[0068] Specifically, when the characteristic deterioration trend is high, the system will give corresponding maintenance reference suggestions and send push content to relevant personnel.

[0069] Furthermore, the present invention also provides a computer-readable storage medium, wherein the computer-readable storage medium includes a stored program, wherein the program executes the method described in the above method embodiment when it is run.

[0070] Furthermore, the present invention also provides an electronic device, including a memory and a processor, wherein the memory stores a computer program, and the processor is configured to execute the method described in the above method embodiment through the computer program.

[0071] Further, it should be understood that since the setting of each module is only for illustrating the functional units of the device of the present invention, the physical devices corresponding to these modules may be the processor itself, or a part of the software in the processor, a part of the hardware, or a part of the combination of software and hardware. Therefore, the number of each module in the figure is only schematic.

[0072] It is understood by those skilled in the art that all or part of the processes in the method for implementing the above-mentioned embodiments of the present invention can also be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a computer-readable storage medium, and when the computer program is executed by a processor, the steps of each of the above-mentioned method embodiments can be implemented. Among them, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form, etc. The computer-readable storage medium may include: any entity or device, medium, U disk, mobile hard disk, disk, optical disk, computer memory, read-only memory, random access memory, electric carrier signal, telecommunication signal and software distribution medium, etc., which can carry the computer program code.

[0073] The computer device may be a terminal, and the computer device includes a processor, a memory, a communication interface, a display screen and an input device connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface network of the computer device, NFC (near field communication) or other technical implementations. The display screen of the computer device may be a liquid crystal display or an electronic ink display screen, and the input device of the computer device may be a touch layer covering the display screen, or a button, trackball or touchpad provided on the housing of the computer device, or an external keyboard, touchpad or mouse, etc.

[0074] Those skilled in the art will appreciate that the various modules in the device may be adaptively split or merged, and such splitting or merging of specific modules will not cause the technical solution to deviate from the principles of the present invention. Therefore, the technical solutions after splitting or merging will fall within the protection scope of the present invention.

[0075] Furthermore, the present invention also provides a computer-readable storage medium, wherein the computer-readable storage medium includes a stored program, wherein the program executes the method described in the above method embodiment when it is run.

[0076] Furthermore, the present invention also provides an electronic device, including a memory and a processor, wherein the memory stores a computer program, and the processor is configured to execute the method described in the above method embodiment through the computer program.

[0077] Further, it should be understood that since the setting of each module is only for illustrating the functional units of the device of the present invention, the physical devices corresponding to these modules may be the processor itself, or a part of the software in the processor, a part of the hardware, or a part of the combination of software and hardware. Therefore, the number of each module in the figure is only schematic.

[0078] It is understood by those skilled in the art that all or part of the processes in the method for implementing the above-mentioned embodiments of the present invention can also be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a computer-readable storage medium, and when the computer program is executed by a processor, the steps of each of the above-mentioned method embodiments can be implemented. Among them, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form, etc. The computer-readable storage medium may include: any entity or device, medium, U disk, mobile hard disk, disk, optical disk, computer memory, read-only memory, random access memory, electric carrier signal, telecommunication signal and software distribution medium, etc., which can carry the computer program code.

[0079] The computer device may be a terminal, and the computer device includes a processor, a memory, a communication interface, a display screen and an input device connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface network of the computer device, NFC (near field communication) or other technical implementations. The display screen of the computer device may be a liquid crystal display or an electronic ink display screen, and the input device of the computer device may be a touch layer covering the display screen, or a button, trackball or touchpad provided on the housing of the computer device, or an external keyboard, touchpad or mouse, etc.

[0080] Those skilled in the art will appreciate that the various modules in the device may be adaptively split or merged, and such splitting or merging of specific modules will not cause the technical solution to deviate from the principles of the present invention. Therefore, the technical solutions after splitting or merging will fall within the protection scope of the present invention.

Claims

1. A feature-based jewelry monitoring system, characterized in that: It includes data module, initial portrait module, feature acquisition module, portrait risk module and hierarchical supervision module; The data module includes feature data for establishing a portrait, a portrait cycle, and threshold data for evaluating the portrait; wherein the feature data includes feature category, feature value, and feature weight, and the threshold data includes feature difference threshold and portrait change trend threshold; The initial portrait module can establish an initial portrait as an evaluation standard through initial feature data; The feature acquisition module can acquire features at a set portrait cycle and establish a monitoring portrait through the acquired features; The portrait risk module can use the initial portrait as an evaluation standard to obtain the risk characteristics of the monitoring portrait; The supervision module evaluates the monitoring profiles that generate risk characteristics and their changing trends, and performs graded supervision based on the obtained evaluation results, wherein the graded supervision includes weak risk supervision and strong risk supervision.

2. The feature-based jewelry monitoring system of claim 1, wherein: In the initial portrait module, it is assumed that jewelry has n features QA, where the i-th feature QA i The initial eigenvalue is W i0 , the feature weight is ω i , the portrait score U0 of the initial portrait can be obtained; Among them, μ is the adjustment coefficient (0.7≤μ≤1).

3. The feature-based jewelry monitoring system of claim 2, wherein: In the feature acquisition module, the feature is acquired with the set initial image cycle △T0, and the monitoring time T after the end of the jth image cycle j Get the jth monitoring image PA j Image score U j ; Where W ij For Feature QA i In T j The characteristic value of the monitoring time, ω i For QA i The feature weights of .

4. The feature-based jewelry monitoring system of claim 3, wherein: In the portrait risk module, according to the feature QA i In T j The characteristic value W of the monitoring time ij and its initial eigenvalue W i0 , get feature QA i The characteristic difference △W generated in the jth image cycle ij =W ij -W i0 ; According to the characteristic difference threshold △W i0 And the image change trend threshold E0, for feature QA i To conduct an assessment: When △W ij <△W i0 When judging feature QA i normal; When △W ij ≥△W i0 When the feature QA i As the risk feature of the j-th image period, its feature value W ij Its monitoring time T j Recorded in risk feature list AF.

5. The feature-based jewelry monitoring system of claim 4, wherein: In the hierarchical supervision module, the risk feature QA in the risk feature list AF is i , according to the jth monitoring image PA j Image score U j The portrait score U0 compared with the initial portrait; When U j When ≥U0, weak risk supervision is implemented; When U j When <U0, QA according to risk characteristics i The characteristic difference △W in the jth image period ij , get the image change trend in the jth image period When E j When ≥E0, strong risk supervision is implemented; When E j When <E0, weak risk supervision is implemented.

6. The feature-based jewelry monitoring system of claim 5, wherein: The regulatory model of weak risk supervision is: At the j+1th monitoring time T j+1 Get the j+1th image PA of jewelry A j+1 , T j+1 =T j +△T0; Among them, we get the i-th feature QA i At the j+1th monitoring time T j+1 The eigenvalue W i(j+1) , and the j+1th image PA of jewelry A j+1 Image score Get the characteristic difference △W in the j+1th image period i(j+1) =W i(j+1) -W ij ; When △W i(j+1) <△W ij When the When △W i(j+1) ≥△W ij When the time comes, it will switch to strong risk supervision.

7. The feature-based jewelry monitoring system of claim 5, wherein: The supervisory model of strong risk supervision is: Get risk feature list AF QA i The number of occurrences num, get QA i The risk distribution θ i =num / j; Adjust the image period to △T1, where: At the j+1th monitoring time T j+1 Get the j+1th image PA j+1 , T j+1 =T j +△T1; Among them, QA i At the j+1th monitoring time T j+1 The eigenvalue W i(j+1) , and the j+1th image PA j+1 Image score Get the characteristic difference △W in the j+1th image period i(j+1) =W i(j+1) -W ij ; When △W i(j+1) <△W ij When the risk is too high, it will be switched to weak risk supervision; When △W i(j+1) ≥△W ij While maintaining strong risk supervision.

8. The feature-based jewelry monitoring system of claim 7, wherein: Set the risk distribution threshold θ max , when QA i The risk distribution θ i ≥θ max When QA i Issue early warnings and push risk-related information to servers and terminals.

Images