Internet of Things label scale with AI identification module

Abstract

According to the Internet of Things label scale with the AI identification module, a user presses the classification set change button as an instruction, classification matching objects are limited to the weighed objects which are subjected to classification matching binding with the classification set change button, the matching range of subsequent classification features of the to-be-weighed objects is narrowed, and the classification speed is increased. According to the provided classification algorithm, the boundary shape image between the first contour of the to-be-weighed object and the second contour of the set visual field range of the AI identification module in the label scale for the to-be-weighed object is used as the classification feature, and the fine feature difference between the to-be-weighed object and the weighed object with the similar appearance feature but different types does not need to be captured any more. On the premise that the classification precision does not need to be improved from the perspective of improving the classification algorithm, the classification accuracy of the to-be-weighed objects with similar appearance characteristics but different types is improved. According to the label scale internet-of-things batch parameter updating method provided by the invention, synchronous updating of batch parameters of the weighed objects in the same classification set is realized by pressing the classification set change button.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent label scales, and in particular to an Internet of Things label scale with an AI recognition module. Background Art

[0002] Smart label scales are currently widely used in the market. For example, after placing fruit trees such as pineapples and tomatoes on the label scale, the label scale outputs the classification results and obtains the unit price of the goods after image recognition of the placed objects, and then weighs and prices the successfully classified goods. However, the existing smart label scales have the following three main problems:

[0003] 1. When the appearance of the weighed objects, such as size features and color features, have a high degree of similarity, such as oranges and tangerines with highly similar size and color features, conventional classification methods based on image recognition may make mistakes, and a classification algorithm that can capture more detailed features is needed for identification and classification. However, the higher the accuracy of the algorithm, the longer the calculation time is usually, which is usually not allowed in the scenario where there is a high requirement for reducing waiting time in the field of commodity weighing.

[0004] 2. For weighing objects of the same type but different sizes, the image recognition classification method is more prone to errors. For example, for the same type of netted melons, whose appearance features including shape features and color features are highly similar but whose sizes are different, the existing image classification algorithms are prone to pricing errors. For example, the same type of netted melons have basically the same appearance features except for size, and their sizes are also similar, but the pricing is different due to the different sizes. For two such weighing objects, how to identify the size difference features of the two frames, and then reduce pricing classification errors, is also a technical problem that needs to be solved urgently in this field.

[0005] 3. Multiple smart label scales are usually used in the same shopping mall. When setting label scales in batches, the weighing objects are generally set one by one. For example, a pineapple is placed on the label scale. The label scale recognizes the pineapple through image recognition and jumps to the parameter input page. The user enters the information of the pineapple, such as the unit price, in the page that jumps. After the setting is completed, other label scales in the setting mode connected to the label scale will modify the parameter information of the pineapple synchronously. This one-by-one setting method is inefficient and takes a long time when the number of weighing objects to be set is large. Summary of the invention

[0006] The present invention aims to improve the classification accuracy of weighing objects with similar appearance features but different types, and weighing objects of the same type and with difference features captured by conventional classification algorithms as size features, and to realize batch setting and updating of parameter information of weighing objects of various smart label scales of the Internet of Things. An Internet of Things label scale with an AI recognition module is provided.

[0007] To achieve this object, the present invention adopts the following technical solutions:

[0008] An Internet of Things label scale with an AI recognition module is provided. The object to be weighed has a classification set to which it belongs. When the object to be weighed fails to be classified by image recognition, the following steps are adopted to identify, classify and price the object to be weighed, and the steps include:

[0009] S1, pressing the category set change button corresponding to the object to be weighed, the label scale adjusts the set field of view and the set field of view depth of the AI ​​recognition module according to the button pressing as an instruction, and obtains the standard classification features of each historical weighed object in the category set to which the object to be weighed belongs;

[0010] S2, selecting the object to be weighed with a minimum circumscribed rectangle, and then calculating the center point of the minimum circumscribed rectangle;

[0011] S3, adjusting the position of the AI ​​recognition module under the set field of view depth so that the central point is the field of view center point of the AI ​​recognition module generating the set field of view range;

[0012] S4, extracting the boundary shape image between the first outline of the object to be weighed selected by the minimum circumscribed rectangle and the second outline of the set field of view as the classification feature of the object to be weighed, and then performing maximum similarity matching with each of the standard classification features associated with each of the historical weighed objects obtained in step S1, and the matching result is used as the classification result and priced.

[0013] Preferably, the weighing objects belonging to different classification sets have different size characteristics; the weighing objects in the same classification set have similar appearance characteristics but are of different types; the similar appearance characteristics of the weighing objects include any one or more of similar size characteristics, similar color characteristics, and similar shape characteristics; wherein, similarity in size characteristics means: the difference in coverage area of ​​the minimum circumscribed rectangle of the weighing objects is less than a preset first difference threshold; similarity in color characteristics means: the difference in color mean values ​​of the weighing objects is less than a preset second difference threshold; similarity in shape characteristics means: the difference in contour shape of the weighing objects is less than a preset third difference threshold.

[0014] Preferably, step S2 specifically includes the steps of:

[0015] S21, using the standard classification feature as an instruction, activating and lighting a light projection module disposed under the transparent base of the label scale, and assigning an initial projection aperture range of the light projection module to the set field of view range;

[0016] S22, the light projection module detects whether the object to be weighed is within the initial projection aperture circle defined by the set visual field range projected upward from its initial position,

[0017] If yes, go to step S23;

[0018] If not, adjust the position until the object to be weighed is detected and then proceed to step S23;

[0019] S23, the light projection module reduces the projection aperture of the initial projection aperture range projected upward according to the set aperture reduction step, and the light receiving module arranged on the top of the label scale identifies and draws the vanishing point position when the projection light disappears;

[0020] S24, the light receiving module selects the shape of the object to be weighed represented by each of the vanishing points in a minimum circumscribed rectangle selection manner, and then calculates the center point of the minimum circumscribed rectangle.

[0021] Preferably, two diagonal lines are connected to each other for the minimum circumscribed rectangle, and the intersection of the two connecting lines is used as the center point of the minimum circumscribed rectangle.

[0022] Preferably, the set field of view is a perfect circle, and the classification features of the object to be weighed include boundary shape features in the boundary shape image, and / or gap width features between the first contour and the second contour of the boundary shape image.

[0023] Preferably, the first contour of the object to be weighed is the minimum circumscribed rectangle that frames the object to be weighed; the gap width feature is the average of the distances between each vertex of the minimum circumscribed rectangle and the boundary of the second contour, and the connecting line of the distances points to the center of the set field of view.

[0024] Preferably, the method for forming the standard classification characteristics of the historical weighed objects obtained in step S1 comprises the steps of:

[0025] A1, for each vertex in the minimum circumscribed rectangle corresponding to each historical weighing object of the same type and the same price and having similarity in boundary shape features, calculate the distance to the second contour of the set visual field bound to the historical weighing object,

[0026] A2, calculating the average value of the distances corresponding to the vertices having a point-to-point correspondence relationship in the minimum circumscribed rectangle corresponding to each of the historical weighed objects for which the distance calculation is completed in step A1;

[0027] A3, for any of the historical weighed objects obtained in step A1, the boundary shape is expanded or contracted by the average value corresponding to each of the vertices solved in step A2 to obtain a standard boundary shape, and the classification feature of the standard boundary shape is used as the standard classification feature.

[0028] Preferably, in step S4, the method for performing maximum similarity matching comprises the steps of:

[0029] S41, performing boundary shape feature similarity matching between the standard classification features corresponding to the object to be weighed and each of the historical weighed objects obtained in step S1;

[0030] S42, determining whether the number of similarity matching results in step S41 is greater than or equal to "2",

[0031] If yes, go to step S43;

[0032] If not, when the number of similarity matching results outputted in step S41 is "1", the matching result is used as the classification result of the object to be weighed, and when the number of similarity matching results outputted in step S41 is "0", it is determined that the classification fails;

[0033] S43, performing gap width feature similarity matching between the object to be weighed and each similarity matching result outputted in step S41;

[0034] S44, determining whether the similarity matching result of step S43 is unique.

[0035] If yes, the matching result is used as the classification result of the object to be weighed;

[0036] If not, the classification is considered to have failed.

[0037] The present application also provides a method for updating batch parameters of a label scale IoT, which updates the parameters of weighing objects in batches for the IoT label scale with an AI recognition module, including the steps of:

[0038] L1, pressing the same category set change button on each smart label scale of the Internet of Things, each smart label scale uses the button pressing as an instruction to obtain the category set bound to the button, and the appearance features of the weighing objects under the category set are similar but not of the same type;

[0039] L2, placing each weighing object to be updated with parameters under the classification set on a corresponding smart label scale for classification and identification;

[0040] L3, for the first weighing object successfully classified on the first smart label scale, in the classification set associated with the classification set change button currently pressed on each of the other second smart label scales connected to the first smart label scale, mark the first weighing object for parameter synchronization update;

[0041] L4, each of the second smart label scales takes the completion of parameter update of the first weighing object by the user on the first smart label scale as an instruction, and performs parameter synchronization update on the first weighing object marked for synchronization update in the classification set associated with the classification set change button currently pressed in step L3.

[0042] The present invention has the following beneficial effects:

[0043] 1. The label scale takes the user pressing the classification set change button as an instruction, and limits the classification matching objects to the weighing objects that have been classified and matched with the classification set change button, which narrows the matching range of the subsequent classification features of the objects to be weighed, which is conducive to greatly improving the classification speed.

[0044] 2. The classification algorithm uses the boundary shape image between the first contour of the object to be weighed and the second contour of the set field of view of the AI ​​recognition module in the label scale for the object to be weighed as the classification feature. It is no longer necessary to capture the subtle feature differences between the object to be weighed and weighing objects with similar but different appearance features. That is, without having to improve the classification accuracy from the perspective of improving the classification algorithm, the classification accuracy of objects to be weighed with similar but different appearance features is improved.

[0045] 3. For objects to be weighed that are of different but similar sizes and different prices but of the same type, by setting the same depth of field of view for the AI ​​recognition module, the different size characteristics between the same type of objects to be weighed are reflected in the gap width between the first contour and the second contour mentioned above. The gap width is used as one of the classification features of the boundary shape image, and combined with the use of the classification set change button, accurate classification of objects to be weighed that are of different but similar sizes and different prices is achieved.

[0046] 4. When updating the parameters of weighing objects in batches on each smart label scale of the Internet of Things, firstly, the weighing objects to be updated under the category set bound to the same category set change button currently pressed by each smart label scale of the Internet of Things are classified and identified by the method recorded in steps S1-S4, and then the first weighing object successfully classified on the first smart label scale is marked for parameter synchronization update in the category set associated with the category set change button currently pressed on other second smart label scales of the Internet of Things of the first smart label scale. Finally, after the user completes the parameter update of the first weighing object on the first smart label scale, each second smart label scale uses this as an instruction to perform parameter synchronization update on the first weighing object marked for synchronization update, thereby realizing batch parameter synchronization update of each weighing object belonging to the same category set on multiple smart label scales of the Internet of Things.

[0047] 5. The shape features of the same type of weighing objects are smoothed using the minimum circumscribed rectangle to reduce the difference in boundary shape between the first outline (preferably the minimum circumscribed rectangle) of each weighing object of the same type with different shapes and the second outline of the set field of view. This is conducive to reducing the complexity of the subsequent classification algorithm using the boundary shape image as the classification feature, and improving the classification accuracy of weighing objects of different types with similar appearance features.

[0048] 6. Using the center point of the minimum circumscribed rectangle as a reference point for finding the center point of the field of view of the set field of view makes it possible for the AI ​​recognition module to adjust its position so that the field of view set for it covers the first outline of the weighing object, and further makes it possible to find subtle differences between weighing objects of similar appearance but different types, namely, boundary shape image features.

[0049] 7. The set field of view is set to a perfect circle. For weighing objects that are not perfect circles in appearance, no matter how the weighing object is rotated within a 360° range, the boundary shape formed by the first outline obtained by projecting light on the weighing object and the second outline of the set field of view that is a perfect circle are the same. This ensures the consistency of the classification features of the weighing object, which is beneficial to improving the classification accuracy of weighing objects of similar appearance features but different types. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the embodiments of the present invention. Obviously, the drawings described below are only some embodiments of the present invention, and for ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0051] Figure 1 This is a method step diagram of identifying and classifying objects to be weighed by an Internet of Things tag scale with an AI recognition module provided by an embodiment of the present invention;

[0052] Figure 2 It is a schematic diagram of a boundary shape image collected between a first contour of the weighing object and a second contour of the set visual field corresponding to the weighing object under a set visual field range and a set visual field depth bound to the weighing object;

[0053] Figure 3 It is a schematic diagram of the light-emitting module projecting light upward from the transparent bottom of the label scale and being received by the light-receiving module;

[0054] Figure 4 This is an example diagram of a boundary shape image between a first outline of an object to be weighed selected by a minimum circumscribed rectangle and a second outline of a set field of view. DETAILED DESCRIPTION

[0055] The technical solution of the present invention is further described below with reference to the accompanying drawings and through specific implementation methods.

[0056] Among them, the drawings are only used for illustrative explanations, and they only represent schematic diagrams rather than actual pictures, and should not be understood as limitations on this patent; in order to better illustrate the embodiments of the present invention, some parts of the drawings may be omitted, enlarged or reduced, and do not represent the size of the actual product; for those skilled in the art, it is understandable that some well-known structures and their descriptions in the drawings may be omitted.

[0057] The same or similar numbers in the drawings of the embodiments of the present invention correspond to the same or similar parts; in the description of the present invention, it should be understood that if the terms "upper", "lower", "left", "right", "inner", "outer", etc. appear, the orientation or position relationship indicated is based on the orientation or position relationship shown in the drawings, which is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation. Therefore, the terms describing the position relationship in the drawings are only used for illustrative purposes and cannot be understood as a limitation on this patent. For ordinary technicians in this field, the specific meanings of the above terms can be understood according to specific circumstances.

[0058] In the description of the present invention, unless otherwise clearly specified and limited, if the term "connection" or the like appears to indicate the connection relationship between components, the term should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection or an indirect connection through an intermediate medium, it can be the internal connection of two components or the interaction relationship between two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0059] First of all, it should be noted that the objects to be weighed in this embodiment are various items to be weighed and priced, and the objects to be weighed have classification sets to which they belong. In this embodiment, the classification sets are divided based on the differences in size characteristics between the objects to be weighed, and the appearance characteristics of the objects to be weighed in the same classification set are similar but of different types. For example, the sizes of oranges and textured melons are d1 and d2 respectively. Assuming that there is a size difference between d1 and d2 according to the size classification rules, oranges are classified into the first classification set, and textured melons are classified into the second classification set.

[0060] The similarity of the appearance features of the weighing objects in the same classification set includes any one or more of the similarity of size features, similarity of color features, and similarity of shape features. Among them, the similarity of size features means that the difference in the coverage area of ​​the minimum circumscribed rectangle of the weighing objects is less than the preset first difference threshold. For example, assuming that oranges and oranges belong to the same first classification set, the coverage area of ​​the minimum circumscribed rectangle of oranges is assumed to be a1, and the coverage area of ​​the minimum circumscribed rectangle of oranges is assumed to be a2, and a1-a2 is less than the preset area difference threshold (first difference threshold), then the size features of oranges and oranges are determined to be similar. The method of selecting weighing objects with the minimum circumscribed rectangle will be described in detail in the subsequent content. The similarity of color features means that the difference in the color mean of the weighing objects is less than the preset second difference threshold. For example, for oranges and oranges, the color features of each pixel of the weighing object image selected by the minimum circumscribed rectangle are identified, and then the color value of each pixel is averaged (i.e., the color mean). The difference in color mean is also the absolute value of the difference in the color mean of the two weighing objects. The similarity of shape features means that the difference in the contour shape of the weighing objects is less than the preset third difference threshold. The outline shape of the weighing object referred to here is Figure 2 The boundary shape image between the first contour of the weighing object and the second contour of the set field of view corresponding to the weighing object. The contour shape difference of the weighing object refers to: the dissimilarity value of the boundary shape images of the two weighing objects, and the dissimilarity value is 1-similarity value. How to calculate the similarity value of the boundary shape images of the two weighing objects will be explained in detail in the subsequent content.

[0061] In an embodiment of the present invention, an Internet of Things label scale with an AI recognition module is provided. The object to be weighed has a classification set to which it belongs. When the object to be weighed fails to be classified by a built-in conventional image recognition method, Figure 1 As shown, the following steps are used to identify, classify and price the objects to be weighed:

[0062] S1, press the category set change button corresponding to the object to be weighed, and the label scale adjusts the set field of view range and set field of view depth of the AI ​​recognition module based on the button press as a command;

[0063] For example, the object to be weighed is an orange, and the classification set to which it belongs is the first classification set. The first classification set corresponds to the classification set change button b1, that is, the classification set change button corresponding to the orange is b1. When the label scale fails to classify the orange, the label scale alarms and prompts the user to press the classification set change button b1 set on the label scale. When the user presses button b1, the label scale uses this as an instruction to adjust the field of view of the AI ​​recognition module to the set field of view corresponding to button b1, and adjusts the field of view depth to the set field of view depth corresponding to button b1. Figure 2The circle 100 in the figure is the set field of view, and the height H is the set field of view depth. The size of the set field of view range and the set field of view depth is related to the size of the weighing object. Since how to set the set field of view range and the set field of view depth for each weighing object with size difference characteristics is not within the scope of the rights protection required by this application, and this technical point does not affect the solution of the technical problem, it will not be elaborated in detail.

[0064] The label scale automatically adjusts the set field of view and the set field of view depth of the AI ​​recognition module according to the user's button pressing instructions, and also obtains the standard classification features of each historical weighed object in the classification set to which the object to be weighed belongs. The standard classification features are related to the distance between each vertex of the minimum circumscribed rectangle of the historical weighed object and the second contour of the set field of view bound to the historical weighed object. The method of forming the standard classification features will be specifically explained after the technical points such as the generation method of the minimum circumscribed rectangle are explained later.

[0065] Through step S1, the visual field range and visual field depth of the AI ​​recognition module are adjusted to the set visual field range and set visual field depth corresponding to the object to be weighed. Figure 1 As shown, the method for identifying, classifying and pricing objects to be weighed by the label scale provided in this embodiment enters the steps:

[0066] S2, select the object to be weighed with the minimum circumscribed rectangle, and then calculate the center point of the minimum circumscribed rectangle, which specifically includes the following steps:

[0067] S21, using the standard classification features of each historical weighed object in the classification set to which the object to be weighed belongs as an instruction, activating and lighting the light-projecting module set under the transparent base of the label scale, and assigning the initial projection aperture range of the light-projecting module to be the set field of view range adjusted for the AI ​​recognition module in step S1;

[0068] It should be noted here that the light projection module is set under the transparent base of the label scale, and the light projection direction is vertically upward, and the light is projected upward in the form of an aperture. After the light projection module is activated, the initial projection aperture range of the light projected upward is the set field of view range.

[0069] S22, the light projection module detects whether there is an object to be weighed within the initial projection aperture circle defined by the set field of view projected upward from its initial position,

[0070] If yes, go to step S23;

[0071] If not, the position is adjusted according to the preset trajectory until the object to be weighed is detected and then the process goes to step S23;

[0072] There are many existing methods for the light projection module to detect the object to be weighed. For example, the object to be weighed can be detected by a distance sensor. The light projection module sends a distance detection signal upward within the initial projection aperture range. When the object to be weighed is encountered, the detection signal is blocked, and the distance between the light projection module and the object to be weighed can be calculated based on the sending speed and sending duration of the detection signal. This technology can also be used to determine whether the object to be weighed is within the initial projection aperture range. For example, if the detection signals at each boundary point of the aperture are blocked and the blocking distance is close, it is determined that there is an object to be weighed within the initial projection aperture range.

[0073] It should be noted here that the number of objects to be weighed detected within the initial projection aperture may not be one. For example, if a consumer buys 5 oranges and weighs and prices them together, multiple oranges may be detected within the initial projection aperture. The method of adjusting the position of the projection module according to the preset trajectory is, for example, to adjust the position from the initial position to the right and from the top to the bottom according to the set detection step length for detection.

[0074] S23, the light projection module projects the aperture according to the set aperture reduction step size for the initial projection aperture range projected upward, and the light receiving module arranged on the top of the label scale identifies and draws the vanishing point position when the projection light disappears;

[0075] For example, Figure 3 In the process, the light-projecting module 200 projects light upward from the transparent bottom of the label scale and is received by the light-receiving module 300. When the light-projecting module projects light of the initial projection aperture range upward, all the light projected upward from the boundary of the initial projection aperture range is received by the light-receiving module 300. As the light-projecting module gradually reduces the projection aperture of the initial projection aperture range according to the set aperture reduction step, once the aperture boundary encounters an obstacle of the object to be weighed, the light-receiving module 300 will not be able to receive the light projected upward from the boundary point. At this time, the light-receiving module 300 will record the disappearance moment of the projected light from the boundary point and the boundary point until all the upward projected light from all the boundary points disappears. Then, draw out each boundary point at each disappearance moment of the light generated by the aperture boundary to characterize the shape characteristics of the object to be weighed.

[0076] It should be noted here that in conventional schemes, the shape features of the object to be weighed are identified through image recognition. However, in weighing application scenarios, due to different lighting conditions, the shape features of the object to be weighed are identified from top to bottom through image recognition. Due to factors such as light and shadow, the identified shape features of the object to be weighed may be inaccurate. This is unacceptable in the scenario where the present application is to improve the classification accuracy of objects to be weighed with similar appearance features but different types, especially when conventional image classification algorithms have failed to classify objects to be weighed with similar appearance features but different types.

[0077] S24, the light receiving module selects the shape of the object to be weighed represented by each vanishing point in a minimum circumscribed rectangle selection manner, and then calculates the center point of the minimum circumscribed rectangle;

[0078] The shape of the object to be weighed is framed with a minimum enclosing rectangle in such a way that the minimum enclosing rectangle is obtained, and the existing method is adopted and no specific explanation is given.

[0079] In this embodiment, the method for calculating the center point of the minimum circumscribed rectangle is as follows:

[0080] like Figure 2 As shown in , two diagonal lines are connected for the minimum circumscribed rectangle 500, and the intersection point 400 of the two connecting lines is used as the center point of the minimum circumscribed rectangle 500. It should be noted here that the minimum circumscribed rectangle is formed on the light receiving module, that is, the top of the label scale.

[0081] It should be noted here that the center point of the minimum circumscribed rectangle is used as the center point of the field of view for finding the set field of view, so that it is possible to adjust the position of the AI ​​recognition module to cover the first contour of the object to be weighed for its field of view, and further make it possible to find the subtle difference features between the weighing objects of similar appearance but different types, that is, the boundary shape image features. In addition, the shape features of the same type of weighing objects are smoothed using the minimum circumscribed rectangle to reduce the difference in boundary shape between the first contour of each weighing object with different shapes under the same type and the second contour of the set field of view, which is conducive to reducing the algorithm complexity of the subsequent classification using the boundary shape image as the classification feature (i.e., the maximum similarity matching of step S4), because only the difference in boundary shape between the morphological contour (first contour) of the minimum circumscribed rectangle and the second contour formed by the set field of view needs to be considered, without excessive consideration of the details of the morphological contour of the weighing object framed in the minimum circumscribed rectangle, thereby improving the classification accuracy of weighing objects of similar appearance features but different types.

[0082] Through step S2, after the object to be weighed is selected with the minimum circumscribed rectangle and the center point of the minimum circumscribed rectangle is calculated, the method for identifying, classifying and pricing the object to be weighed by the Internet of Things label scale with AI recognition module provided in this embodiment is transferred to step:

[0083] S3, under the set field of view depth, adjusting the position of the AI ​​recognition module so that the center point is the field of view center point of the AI ​​recognition module generating the set field of view range;

[0084] For example, Figure 2As shown in , under the set field of view depth H, the position of the AI ​​recognition module is adjusted so that the center point 400 is the field of view center point of the AI ​​recognition module generating the set field of view range, that is, the center point 400 overlaps with the field of view center point. In this embodiment, the set field of view range is preferably a perfect circle. For a weighing object with a non-perfect circle appearance, no matter how the weighing object is rotated and placed within a 360° range, the boundary shape formed between the first contour obtained by projecting light on the weighing object and the second contour of the set field of view range that is a perfect circle is the same, which ensures the consistency of the classification features of the weighing object, and is conducive to improving the classification accuracy of weighing objects of similar appearance features but different types.

[0085] S4, extract the boundary shape image between the first outline of the object to be weighed selected by the minimum circumscribed rectangle and the second outline of the set field of view as the classification feature of the object to be weighed, and then perform maximum similarity matching with each standard classification feature of each historical weighed object obtained in step S1, and use the matching result as the classification result and calculate the price.

[0086] Figure 4 In the figure, the gray part is an example of a boundary shape image. In this embodiment, the classification feature includes the boundary shape feature in the boundary shape image (i.e. Figure 4 The shape of the gray part shown in FIG. 1 ), and / or the gap width feature between the first contour (preferably the minimum circumscribed rectangle of the object to be weighed) and the second contour (preferably a perfect circle) of the boundary shape image. The gap width feature is the average of the distances between each vertex of the minimum circumscribed rectangle and the boundary of the second contour, and the connecting line forming the distance points to the center of the circle of the set field of view.

[0087] For example, Figure 4 In the figure, the reference numeral "600" is the center of the circle of the set field of view. The distances between the vertices p1, p2, p3, p4 of the minimum circumscribed rectangle and the boundary of the second contour are l1, l2, l3, l4 respectively, and the gap width characteristic of the object to be weighed is (l1+l2+l3+l4) / 4.

[0088] The following is an explanation of the method for forming the standard classification characteristics of historical weighing objects:

[0089] In this embodiment, the method for forming the standard classification feature includes the following steps:

[0090] A1, for each vertex of the minimum circumscribed rectangle corresponding to each historical weighing object of the same type and the same price and having similarity in boundary shape features, calculate the distance to the second contour of the set visual range bound to the historical weighing object;

[0091] The distance calculation method is as follows: Figure 4I have explained it as an example and will not repeat it here. What needs to be explained here is that the same type and the same pricing means, for example, it is necessary to form a standard classification feature corresponding to "oranges", and all the historical weighed objects obtained in step A1 are "oranges". The same pricing means that the selling prices of the "oranges" classified by steps S1-S4 are the same. The similarity of boundary shape features between the weighed objects means, for example, the boundary shape feature of a certain orange is Figure 4 As shown in the gray part of , if the overlap between the boundary shape feature of another orange and the gray part is greater than the preset overlap threshold, it is determined that the two have boundary shape feature similarity. As for the calculation of the overlap between the gray parts, the existing method is used and no specific description is given.

[0092] A2, calculating the average value of the distances corresponding to the vertices having point-to-point correspondence in the minimum circumscribed rectangle corresponding to each historical weighed object for which the distance calculation is completed in step A1;

[0093] For example, if the vertex px1 in the historical weighing object x1 obtained in step A1 has an overlapping relationship with the vertex px2 in the historical weighing object x2, it is determined that px1 and px2 have a point correspondence relationship, and then the average value of the distances calculated in step A1 corresponding to px1 and px2 is calculated;

[0094] A3, the boundary shape of any historical weighed object obtained in step A1 is expanded or contracted by the average value corresponding to each vertex solved in step A2 to obtain a standard boundary shape, and the classification feature of the standard boundary shape is used as the standard classification feature.

[0095] For example, in step A3, the historical weighing object to be expanded or reduced is Figure 4 The minimum circumscribed rectangle 700 is shown. When it is necessary to expand or shrink the vertex p1 in the minimum circumscribed rectangle 700, the average value of the distances between each vertex and the second contour in other historical weighing objects obtained in step A1 having a point-to-point correspondence relationship with the vertex p1 is extracted, and then the distance between the vertex p1 and the second contour is expanded or reduced to the average value.

[0096] In step S4, the method for performing maximum similarity matching includes the following steps:

[0097] S41, performing boundary shape feature similarity matching (such as the overlap matching in the above example) on the standard classification features corresponding to the object to be weighed and the historical weighed objects obtained in step S1;

[0098] S42, determining whether the number of similarity matching results in step S41 is greater than or equal to "2",

[0099] If yes, go to step S43;

[0100] If not, when the number of similarity matching results outputted in step S41 is "1", the matching result is used as the classification result of the object to be weighed, and when the number of similarity matching results outputted in step S41 is "0", the classification is judged to have failed and an alarm is prompted;

[0101] S43, further performing gap width feature similarity matching between the object to be weighed and each similarity matching result outputted in step S41;

[0102] S44, determining whether the similarity matching result of step S43 is unique.

[0103] If so, the matching result is used as the classification result of the object to be weighed;

[0104] If not, the classification is judged to have failed and an alarm is prompted.

[0105] It should be noted here that, for objects to be weighed that are of different but similar sizes and have different prices but are of the same type, such as two types of textured melons that are priced differently due to their different sizes, by setting the same depth of field for the AI ​​recognition module, the different size characteristics between the same type of weighing objects are reflected in the gap width between the first contour and the second contour mentioned above. The gap width is used as one of the classification features of the boundary shape image. Through steps S42-S44 and combined with the use of the classification set change button, accurate classification of objects to be weighed that are of different but similar sizes and have different prices is achieved.

[0106] This embodiment also provides a method for updating batch parameters of a label scale IoT, which updates the parameters of weighing objects in batches for an IoT label scale with an AI recognition module, and specifically includes the following steps:

[0107] L1, press the same category set change button for each smart label scale of the Internet of Things. Each smart label scale uses the button pressing as an instruction to obtain the category set bound to the button. The appearance features of each weighing object under the category set are similar but not of the same type; the similar appearance features are specifically explained in the above content and will not be repeated here; the method of networking between the label scales through the Internet of Things is an existing method and will not be explained here either.

[0108] L2, each weighing object to be updated under the classification set is placed on a corresponding smart label scale for classification and identification. The classification and identification method is the above steps S1-S4, which will not be repeated;

[0109] L3, for the first weighing object successfully classified on the first smart label scale, in the classification set associated with the currently pressed classification set change button on each of the other second smart label scales connected to the first smart label scale, the first weighing object is marked for parameter synchronization update;

[0110] For example, in step L2, the first smart label scale successfully classifies the first weighing object "orange". In the classification set associated with the currently pressed classification set change button on each other second smart label scale connected to the first smart label scale, the first weighing object "orange" is marked for parameter synchronization update;

[0111] L4, each second smart label scale takes the completion of parameter update of the first weighing object by the user on the first smart label scale as an instruction, and performs parameter synchronization update on the first weighing object marked with synchronization update in the category set associated with the category set change button currently pressed in step L3.

[0112] When the category set change button is pressed, it remains pressed and will pop up only when pressed again.

[0113] In summary, the label scale provided by the present application uses the user pressing the classification set change button as an instruction to limit the classification matching objects to the weighing objects that have been classified and matched with the classification set change button, thereby narrowing the matching range of the subsequent classification features of the objects to be weighed, which is conducive to greatly improving the classification speed. The provided classification algorithm uses the boundary shape image between the first contour of the object to be weighed and the second contour of the set field of view of the AI ​​recognition module in the label scale for the object to be weighed as the classification feature. It is no longer necessary to capture the subtle feature differences between the object to be weighed and the weighing objects with similar but different appearance features. That is, without improving the classification accuracy from the perspective of improving the classification algorithm, the classification accuracy of the objects to be weighed with similar but different appearance features is improved. The provided label scale IoT batch parameter update method realizes the synchronous update of batch parameters of each weighing object under the same classification set by pressing the classification set change button.

[0114] It should be noted that the above specific implementations are only preferred embodiments of the present invention and the technical principles used. Those skilled in the art should understand that various modifications, equivalent substitutions, changes, etc. can be made to the present invention. However, as long as these changes do not deviate from the spirit of the present invention, they should be within the scope of protection of the present invention. In addition, some terms used in the specification and claims of this application are not restrictive, but are only for the convenience of description.

Claims

1. An Internet of Things label scale with an AI recognition module, characterized in that: The object to be weighed has a classification set to which it belongs. When the classification of the object to be weighed by using image recognition fails, the following steps are used to identify, classify and price the object to be weighed, and the steps include: S1, pressing the category set change button corresponding to the object to be weighed, the label scale adjusts the set field of view and the set field of view depth of the AI ​​recognition module according to the button pressing as an instruction, and obtains the standard classification features of each historical weighed object in the category set to which the object to be weighed belongs; S2, selecting the object to be weighed with a minimum circumscribed rectangle, and then calculating the center point of the minimum circumscribed rectangle; S3, adjusting the position of the AI ​​recognition module under the set field of view depth so that the central point is the field of view center point of the AI ​​recognition module generating the set field of view range; S4, extracting the boundary shape image between the first outline of the object to be weighed selected by the minimum circumscribed rectangle and the second outline of the set field of view as the classification feature of the object to be weighed, and then performing maximum similarity matching with each of the standard classification features associated with each of the historical weighed objects obtained in step S1, and the matching result is used as the classification result and priced.

2. The Internet of Things label scale with an AI recognition module according to claim 1, characterized in that: The weighing objects belonging to different classification sets have different size characteristics; the weighing objects in the same classification set have similar appearance characteristics but are of different types; the similar appearance characteristics of the weighing objects include any one or more of similar size characteristics, similar color characteristics, and similar shape characteristics; wherein, similar size characteristics refer to: the difference in the coverage area of ​​the minimum circumscribed rectangle of the weighing objects is less than a preset first difference threshold; similar color characteristics refer to: the difference in the color mean of the weighing objects is less than a preset second difference threshold; similar shape characteristics refer to: the difference in the contour shape of the weighing objects is less than a preset third difference threshold.

3. The Internet of Things label scale with an AI recognition module according to claim 1, characterized in that: Step S2 specifically includes the following steps: S21, using the standard classification feature as an instruction, activating and lighting a light projection module disposed under the transparent base of the label scale, and assigning an initial projection aperture range of the light projection module to the set field of view range; S22, the light projection module detects whether the object to be weighed is within the initial projection aperture circle defined by the set visual field range projected upward from its initial position, If yes, go to step S23; If not, adjust the position until the object to be weighed is detected and then proceed to step S23; S23, the light projection module reduces the projection aperture of the initial projection aperture range projected upward according to the set aperture reduction step, and the light receiving module arranged on the top of the label scale identifies and draws the vanishing point position when the projection light disappears; S24, the light receiving module selects the shape of the object to be weighed represented by each of the vanishing points in a minimum circumscribed rectangle selection manner, and then calculates the center point of the minimum circumscribed rectangle.

4. An Internet of Things label scale with an AI recognition module according to claim 1 or 3, characterized in that: Lines are connected for each of the two diagonals of the minimum circumscribed rectangle, and the intersection of the two connecting lines is used as the center point of the minimum circumscribed rectangle.

5. An Internet of Things label scale with an AI recognition module according to claim 1 or 3, characterized in that: The set visual field range is a perfect circle, and the classification feature of the object to be weighed includes a boundary shape feature in the boundary shape image and / or a gap width feature between the first contour and the second contour of the boundary shape image.

6. The Internet of Things label scale with an AI recognition module according to claim 5, characterized in that: The first outline of the object to be weighed is the minimum circumscribed rectangle that frames the object to be weighed; the gap width feature is the average of the distances between each vertex of the minimum circumscribed rectangle and the boundary of the second outline, and the connecting line of the distances points to the center of the set field of view.

7. The Internet of Things label scale with an AI recognition module according to claim 1, characterized in that: The method for forming the standard classification characteristics of the historical weighed object obtained in step S1 comprises the steps of: A1, for each vertex in the minimum circumscribed rectangle corresponding to each historical weighing object of the same type and the same price and having similarity in boundary shape features, calculate the distance to the second contour of the set visual field bound to the historical weighing object, A2, calculating the average value of the distances corresponding to the vertices having a point-to-point correspondence relationship in the minimum circumscribed rectangle corresponding to each of the historical weighed objects for which the distance calculation is completed in step A1; A3, for any of the historical weighed objects obtained in step A1, the boundary shape is expanded or contracted by the average value corresponding to each of the vertices solved in step A2 to obtain a standard boundary shape, and the classification feature of the standard boundary shape is used as the standard classification feature.

8. The Internet of Things label scale with an AI recognition module according to claim 5, characterized in that: In step S4, the method for performing maximum similarity matching includes the steps of: S41, performing boundary shape feature similarity matching between the standard classification features corresponding to the object to be weighed and each of the historical weighed objects obtained in step S1; S42, determining whether the number of similarity matching results in step S41 is greater than or equal to "2", If yes, go to step S43; If not, when the number of similarity matching results outputted in step S41 is "1", the matching result is used as the classification result of the object to be weighed, and when the number of similarity matching results outputted in step S41 is "0", the classification is judged to have failed; S43, performing gap width feature similarity matching between the object to be weighed and each similarity matching result outputted in step S41; S44, determining whether the similarity matching result of step S43 is unique. If yes, the matching result is used as the classification result of the object to be weighed; If not, the classification is considered to have failed.

9. A method for updating batch parameters of a label scale Internet of Things, for updating parameters of weighing objects in batches for an Internet of Things label scale with an AI recognition module as described in any one of claims 1 to 8, characterized in that: Includes steps: L1, pressing the same category set change button on each smart label scale of the Internet of Things, each smart label scale uses the button pressing as an instruction to obtain the category set bound to the button, and the appearance features of the weighing objects under the category set are similar but not of the same type; L2, placing each weighing object to be updated with parameters under the classification set on a corresponding smart label scale for classification and identification; L3, for the first weighing object successfully classified on the first smart label scale, in the classification set associated with the classification set change button currently pressed on each of the other second smart label scales connected to the first smart label scale, mark the first weighing object for parameter synchronization update; L4, each of the second smart label scales takes the completion of parameter update of the first weighing object by the user on the first smart label scale as an instruction, and performs parameter synchronization update on the first weighing object marked for synchronization update in the classification set associated with the classification set change button currently pressed in step L3.

Images