Article Identification Method, Device, Electronic Device and Computer Readable Medium

Abstract

The embodiments of the present disclosure disclose an object identification method, device, electronic device and computer-readable medium. A specific implementation of the method includes: obtaining a set of screenshots of the object information display page; inputting the display area image into a pre-trained optical character recognition model to obtain weight identification information; in response to the weight identification information including a weight value and the value being greater than a preset value, executing the following steps: obtaining a carrier plate image; performing object identification on the carrier plate image; in response to the weight identification information being empty, obtaining a carrier plate image; inputting the carrier plate image into a pre-trained empty plate detection model; in response to the empty plate identification result indicating that the object carrier plate carries an object, performing object identification on the carrier plate image to obtain object identification information. This implementation enables item identification to be triggered even if the weight of the object cannot be read when the weighing plate fails during the weighing process. This broadens the application scenarios of object identification.

Description

Technical Field

[0001] Embodiments of the present disclosure relate to the field of computer technology, and more particularly to an object identification method, apparatus, electronic device, and computer-readable medium. Background Art

[0002] Item recognition is a technology that identifies items. Currently, the method commonly used to identify items is to directly weigh the items and trigger item information recognition based on the change in weight. The identified item information is then printed on a label.

[0003] However, when using the above method to identify objects, the following technical problems often occur:

[0004] First, by directly weighing the items, item information recognition is triggered based on the change in weight. If the weighing plate fails during the weighing process, the weight of the item cannot be read, resulting in the inability to trigger item recognition.

[0005] Second, when directly capturing the image of the carrier plate to identify the item information, the background interference of the carrier plate image results in a low accuracy rate of the identified item information. For labels printed with incorrect item information, the item information needs to be re-identified, resulting in waste of label paper. Summary of the Invention

[0006] The content of this disclosure is used to briefly introduce concepts that will be described in detail in the detailed description section below. The content of this disclosure is not intended to identify key features or essential features of the claimed technical solution, nor is it intended to limit the scope of the claimed technical solution.

[0007] Some embodiments of the present disclosure provide object identification methods, devices, electronic devices, and computer-readable media to solve one or more of the technical problems mentioned in the above background technology section.

[0008] In a first aspect, some embodiments of the present disclosure provide an object identification method, the method comprising: obtaining a set of page screenshots of an object information display page within a preset time period. The page screenshots in the above-mentioned page screenshot set include a weight display area; generating a weight information display area image as a display area image based on the above-mentioned page screenshot set; inputting the above-mentioned display area image into a pre-trained optical character recognition model to obtain weight identification information in the above-mentioned display area image; in response to determining that the above-mentioned weight identification information includes a weight identification value, and the above-mentioned weight identification value is greater than a preset value, performing the following steps: determining that an object identification trigger condition is satisfied. The above-mentioned object carrier corresponds to the above-mentioned object information display page; obtaining an object carrier image of the above-mentioned object carrier as a first object carrier image; performing object identification on the above-mentioned first object carrier image to obtain object identification information; in response to determining that the above-mentioned weight identification information is empty, obtaining an object carrier image of the above-mentioned object carrier as a second object carrier image; inputting the above-mentioned second object carrier image into a pre-trained empty tray detection model to obtain an empty tray identification result. Among them, the above-mentioned empty plate recognition result indicates whether the above-mentioned item carrying plate carries an item; in response to determining that the above-mentioned empty plate recognition result indicates that the above-mentioned item carrying plate carries an item, it is determined that the item recognition trigger condition is met; in response to determining that the item recognition trigger condition is met, the above-mentioned second item carrying plate image is subjected to item recognition to obtain item identification information.

[0009] In a second aspect, some embodiments of the present disclosure provide an object identification device, the device comprising: a first acquisition unit, configured to acquire a set of page screenshots of an object information display page within a preset time period; a first input unit, configured to input the above-mentioned display area image into a pre-trained optical character recognition model to obtain weight identification information in the above-mentioned display area image; a first determination unit, configured to determine that an object identification trigger condition is satisfied, acquire an object carrying tray image of the above-mentioned object carrying tray as a first object carrying tray image, perform object identification on the above-mentioned first object carrying tray image, and obtain object identification information; a second acquisition unit, configured to acquire an object carrying tray image of the above-mentioned object carrying tray as a second object carrying tray image in response to determining that the above-mentioned weight identification information is empty; a second input unit, configured to input the above-mentioned second object carrying tray image into a pre-trained empty tray detection model to obtain an empty tray identification result; a second determination unit, configured to determine that an object identification trigger condition is satisfied in response to determining that the above-mentioned empty tray identification result indicates that an object is carried on the above-mentioned object carrying tray; and an identification unit, in response to determining that the object identification trigger condition is satisfied, perform object identification on the above-mentioned second object carrying tray image to obtain object identification information.

[0010] In a third aspect, some embodiments of the present disclosure provide an electronic device comprising: one or more processors; a storage device on which one or more programs are stored, and when the one or more programs are executed by one or more processors, the one or more processors implement the method described in any implementation of the first aspect above.

[0011] In a fourth aspect, some embodiments of the present disclosure provide a computer-readable medium having a computer program stored thereon, wherein when the program is executed by a processor, the method described in any implementation of the first aspect is implemented.

[0012] The above-described embodiments of the present disclosure have the following advantageous effects: Through the object identification methods of some embodiments of the present disclosure, even if the weighing plate fails to read the item's weight during weighing, object identification can still be triggered. This broadens the application scenarios of object identification. Specifically, the inability to trigger object information recognition is due to the fact that, by directly weighing the item, object information recognition is triggered based on weight changes. If the weighing plate fails during weighing, the item's weight cannot be read, resulting in the inability to trigger object identification. Based on this, the object identification methods of some embodiments of the present disclosure first obtain a set of page screenshots of the item information display page within a preset time period. The page screenshots in the set of page screenshots include the weight display area. Thus, individual screenshots of the item information display page within the preset time period are obtained. Secondly, based on the set of page screenshots, an image of the weight information display area is generated as a display area image. Thus, an image displaying weight information is obtained. Thirdly, the display area image is input into a pre-trained optical character recognition model to obtain weight identification information in the display area image. Thus, the optical character recognition model can recognize weight-related information displayed in the display area image. Next, in response to determining that the weight identification information includes a weight identification value and that the weight identification value is greater than a preset value, the following steps are performed: First, determine that an item identification trigger condition is satisfied. The item tray corresponds to the item information display page. Second, obtain an item tray image of the item tray as a first item tray image. Third, perform item identification on the first item tray image to obtain item identification information. Thus, item identification can be triggered and item identification information can be obtained when the identified weight identification information includes a weight identification value greater than a preset value. Next, in response to determining that the weight identification information is empty, obtain an item tray image of the item tray as a second item tray image. Thus, an item tray image can be obtained when the identified weight identification information is empty. The second item tray image is then input into a pre-trained empty tray detection model to obtain an empty tray identification result. The empty tray identification result indicates whether the item tray carries an item. Thus, the empty tray detection model can be used to identify whether the item tray carries an item. Then, in response to determining that the empty tray recognition result indicates that the item carrier tray carries an item, it is determined that an item recognition trigger condition has been met. Thus, upon determining that the item carrier tray carries an item, item recognition can be triggered. Then, in response to determining that the item recognition trigger condition has been met, item recognition is performed on the second item carrier tray image to obtain item identification information. Thus, item recognition can be performed on the acquired item carrier tray image to obtain item identification information.Even when the weight identification information is empty, the image of the item tray can be used to identify whether the tray is loaded with an item. This allows item identification to continue when the tray is confirmed to be loaded with an item. This allows item identification to be triggered even if the weighing tray fails during weighing, even if the item weight cannot be read. This broadens the application scenarios for item identification. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] The above and other features, advantages, and aspects of the various embodiments of the present disclosure will become more apparent with reference to the following detailed description in conjunction with the accompanying drawings. Throughout the drawings, the same or similar reference numerals represent the same or similar elements. It should be understood that the drawings are schematic and that components and elements are not necessarily drawn to scale.

[0014] Figure 1 is a flow chart of some embodiments of the object identification method according to the present disclosure;

[0015] Figure 2 is a schematic structural diagram of some embodiments of the object identification device according to the present disclosure;

[0016] Figure 3 It is a structural diagram of an electronic device suitable for implementing some embodiments of the present disclosure. DETAILED DESCRIPTION

[0017] Embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although certain embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be construed as being limited to the embodiments described herein. On the contrary, these embodiments are provided to provide a more thorough and complete understanding of the present disclosure. It should be understood that the drawings and embodiments of the present disclosure are for illustrative purposes only and are not intended to limit the scope of protection of the present disclosure.

[0018] It should also be noted that, for ease of description, only the parts related to the invention are shown in the drawings. In the absence of conflict, the embodiments and features in the embodiments of the present disclosure may be combined with each other.

[0019] It should be noted that the concepts of "first" and "second" mentioned in this disclosure are only used to distinguish different devices, modules or units, and are not used to limit the order or interdependence of the functions performed by these devices, modules or units.

[0020] It should be noted that the modifications of "one" and "multiple" mentioned in the present disclosure are illustrative rather than restrictive, and those skilled in the art should understand that unless otherwise clearly indicated in the context, they should be understood as "one or more".

[0021] The names of the messages or information exchanged between multiple devices in the embodiments of the present disclosure are only used for illustrative purposes and are not used to limit the scope of these messages or information.

[0022] The present disclosure will be described in detail below with reference to the accompanying drawings and in conjunction with embodiments.

[0023] Figure 1 The process 100 of some embodiments of the object identification method according to the present disclosure is shown. The object identification method includes the following steps:

[0024] Step 101: Obtain a set of page screenshots of item information display pages within a preset time period.

[0025] In some embodiments, an entity executing the item identification method (e.g., a computing device) may obtain a collection of page screenshots. The page screenshots may include an image that includes a weight display area. The item information display page may be a page that displays item information. The page screenshots in the collection of page screenshots may be arranged in ascending order based on the time the page images were captured. In practice, the entity may capture individual page screenshots of the item information display page within a preset time period as the collection of page screenshots.

[0026] Step 102: Generate a weight display area image as a display area image based on the page screenshot set.

[0027] In some embodiments, the execution entity may generate the weight display area image as the display area image based on the page screenshot set.

[0028] In some optional implementations of some embodiments, the execution entity may generate a weight display area image as a display area image based on the page screenshot set through the following steps:

[0029] The first step is to grayscale each page screenshot in the page screenshot set to obtain a page grayscale screenshot set. The page grayscale screenshots in the page grayscale screenshot set can be images in which each pixel in the page grayscale image has only one grayscale value. The grayscale values can be represented by integers between 0 and 255, where 0 represents black, 255 represents white, and intermediate values represent different grayscale levels.

[0030] In the second step, the page grayscale screenshot set is divided in a sliding window manner to obtain a page grayscale screenshot group set. Among them, the page grayscale screenshot group in the page grayscale screenshot group set includes a first page grayscale screenshot and a second page grayscale screenshot. For example, the first page grayscale screenshot can be an image in the page grayscale screenshot group whose screenshot time is earlier than another page grayscale screenshot. The second page grayscale screenshot is an image in the grayscale screenshot group whose screenshot time is later than another page grayscale screenshot. As an example, the page grayscale screenshot set can be (screenshot a, screenshot b, screenshot c). Among them, the time interval of each screenshot is 1, the time window size is set to 2, the sliding duration is 1, and the initial window position is 0. After division, the page grayscale screenshot group set obtained is ((screenshot a, screenshot b), (screenshot b, screenshot c)).

[0031] In the third step, for each page grayscale screenshot group in the above page grayscale screenshot group set, perform the following steps:

[0032] In a first sub-step, the pixel values of the respective pixels in the first page grayscale screenshot included in the page grayscale screenshot group are determined as a first pixel value set.

[0033] In the second sub-step, the pixel value of each pixel point in the second page grayscale screenshot included in the page grayscale screenshot group is determined as a second pixel value set.

[0034] The third sub-step is to determine the average pixel difference value based on the first pixel value set and the second pixel value set. In practice, the absolute value of the difference between each first pixel value in the first pixel value set and each second pixel value in the corresponding second pixel value set is taken to obtain a pixel difference value set. The pixel difference values in the pixel difference value set are averaged to obtain the pixel difference average value. For example, the first pixel value set may be (150, 160, 170, 180), and the second pixel value set may be (50, 60, 70, 80), then the pixel difference value set may be (100, 100, 100, 100), and the pixel difference average value may be 100.

[0035] In the fourth step, the determined average values of the pixel differences are determined as a pixel difference average value set.

[0036] In the fifth step, a pixel difference average value that satisfies a preset numerical condition is selected from the above pixel difference average value set as a target pixel difference average value, wherein the above preset numerical condition may be the maximum pixel difference average value in the pixel difference average value set.

[0037] In the sixth step, the page grayscale screenshot group corresponding to the target average pixel difference value is determined as the target page grayscale screenshot group.

[0038] In the seventh step, a difference map is generated based on the first and second grayscale screenshots of the target page grayscale screenshot group. In practice, the execution entity may perform difference processing on the first and second grayscale screenshots of the target page grayscale screenshot group to generate a difference map.

[0039] In the eighth step, the pixel value of each pixel point in the above differential image is determined as a pixel value set.

[0040] In step 9, in response to determining that each pixel value in the pixel value set is greater than a preset threshold, the position information of each pixel point corresponding to the pixel value set is determined as a set of pixel point position information in a changed area. For example, the preset threshold value may be 50.

[0041] In step 10, based on the above-mentioned set of pixel position information of the changed region, the weight display area in the second page grayscale screenshot included in the target page grayscale screenshot group is determined. The pixel position information of the changed region in the above-mentioned set of pixel position information of the changed region may be coordinate values in a two-dimensional coordinate system. In practice, a two-dimensional coordinate system can be established for the second page grayscale screenshot, and polygonal region extraction processing can be performed on the second page grayscale screenshot based on the coordinates of each pixel point to obtain the weight display area.

[0042] Step 11: Generate a weight display area image based on the determined weight display area. In practice, the execution subject can generate a weight display area image based on the determined weight display area by using template matching image processing technology.

[0043] In some optional implementations of some embodiments, the execution entity may perform grayscale processing on each page screenshot in the page screenshot set to obtain a page grayscale screenshot set through the following steps:

[0044] First, for each page screenshot in the above page screenshot collection, perform the following steps:

[0045] In the first sub-step, each pixel in the above page screenshot is determined as a pixel set.

[0046] The second sub-step is to perform the following pixel processing steps for each pixel in the above pixel set.

[0047] In the first sub-step, the values of the red, green and blue color channels of the above pixel points are determined as a color value set.

[0048] The second sub-step is to generate a color average as a grayscale value based on each color value in the color value set, wherein the grayscale value can be a color value corresponding to a channel.

[0049] The third sub-step is to update the pixel value of the above pixel point to the above grayscale value.

[0050] In the third sub-step, the page screenshot obtained after each pixel is processed by the above pixel processing step is determined as a page grayscale screenshot.

[0051] In the second step, the determined grayscale screenshots of each page are determined as a page grayscale screenshot set.

[0052] In some optional implementations of some embodiments, the execution entity may generate a difference image based on the first page grayscale screenshot and the second page grayscale screenshot included in the target page grayscale screenshot group through the following steps:

[0053] In the first step, the pixel value of each pixel point in the second page grayscale screenshot is determined as a second page grayscale screenshot pixel value set.

[0054] In the second step, for each pixel in the grayscale screenshot of the first page, perform the following steps:

[0055] The first sub-step is to determine the pixel value of the pixel point as the pixel value of the first page grayscale screenshot.

[0056] The second sub-step is to generate a target pixel value based on the pixel value of the grayscale screenshot of the first page and the pixel value corresponding to the grayscale screenshot pixel value of the first page in the grayscale screenshot pixel value set of the second page. In practice, first, the execution entity may determine the difference between the grayscale screenshot pixel value of the first page and the grayscale screenshot pixel value of the second page in the grayscale screenshot pixel value set corresponding to the first page as the initial target pixel value. The execution entity may take the absolute value of the initial target pixel value to obtain the target pixel value. As an example, the grayscale screenshot pixel value of the first page may be 60, and the grayscale screenshot pixel value of the second page in the grayscale screenshot pixel value set corresponding to the second page may be 40, then the initial target pixel value may be 20, and the target pixel value may be 20.

[0057] The third sub-step is to update the pixel value of the pixel point to the target pixel value to obtain an updated pixel point. As an example, the pixel value of the updated pixel point can be 20.

[0058] The third step is to generate a difference map based on each updated pixel point. In practice, the difference map can be generated by performing digital image processing on each updated pixel point.

[0059] In step 103, the display area image is input into a pre-trained optical character recognition model to obtain weight recognition information in the display area image.

[0060] In some embodiments, the execution entity may input the display area image into a pre-trained optical character recognition model to obtain weight identification information from the display area image. The optical character recognition model may be a neural network model that uses the display area image as input data and outputs the weight identification information. For example, the neural network model may be a SVM model or a CNN model.

[0061] Optionally, the optical character recognition model can be trained by the following steps:

[0062] The first step is to obtain a sample set, wherein the samples in the sample set include a screenshot of the sample weight display area and numerical information of the sample target weight corresponding to the screenshot of the sample weight display area.

[0063] In the second step, the following training steps are performed based on the sample set:

[0064] In the first sub-step, a screenshot of the sample weight display area of at least one sample in the sample set is input into the initial neural network to obtain sample identification weight numerical information corresponding to each sample in the at least one sample.

[0065] In a second sub-step, the sample identification weight numerical information corresponding to each of the at least one sample is compared with the corresponding sample target weight numerical information to obtain an average loss. In practice, the execution entity may use a cross-entropy loss function to compare the sample identification weight numerical information corresponding to each of the at least one sample and the corresponding sample target weight numerical information to determine the difference between them.

[0066] The third sub-step is to optimize the parameters of the initial neural network based on the loss average and a preset optimizer, wherein the preset optimizer can be an Adam optimizer.

[0067] In a fourth sub-step, in response to the current number of training times being the preset number of training times, the initial neural network after parameter optimization is used as a trained optical character recognition model.

[0068] In a fifth sub-step, in response to the current number of training times being less than the predetermined number of training times, the initial neural network after parameter optimization is used as the initial neural network, and the sample set is composed of unused samples, and the above training steps are performed again. As an example, the network parameters of the initial neural network can be adjusted using a back propagation algorithm (BP algorithm) and a gradient descent method (e.g., a mini-batch gradient descent algorithm).

[0069] Step 104: In response to determining that the weight identification information includes a weight identification value, and the weight identification value is greater than a preset value, perform the following steps:

[0070] Step 1041: Determine whether the item identification trigger condition is met, wherein the item carrier corresponds to the item information display page.

[0071] In some embodiments, the execution entity may determine that an item recognition trigger condition is satisfied, wherein the item recognition trigger condition may be that the item carrier is not empty.

[0072] Step 1042 : Acquire an item carrying tray image of the item carrying tray as a first item carrying tray image.

[0073] In some embodiments, the execution entity may obtain an image of the item carrying tray as the first item carrying tray image. In practice, the execution entity may obtain an image of the item carrying tray captured by a camera as the first item carrying tray image.

[0074] Step 1043: perform object recognition on the image of the first object carrier tray to obtain object recognition information.

[0075] In some embodiments, the execution entity may perform object recognition on the image of the first object carrier to obtain object recognition information, wherein the object recognition information may include object coding information, object name, and object value.

[0076] Step 105 : In response to determining that the weight identification information is empty, obtaining an item carrying tray image of the item carrying tray as a second item carrying tray image.

[0077] In some embodiments, the execution entity may obtain an image of the item carrying tray as the second item carrying tray image in response to determining that the weight identification information is empty. In practice, the execution entity may obtain an image of the item carrying tray captured by a camera as the second item carrying tray image.

[0078] Step 106: Input the second article carrier tray image into a pre-trained empty tray detection model to obtain an empty tray recognition result.

[0079] In some embodiments, the execution entity may input the second item carrier tray image into a pre-trained empty tray detection model to obtain an empty tray recognition result, wherein the empty tray recognition result may indicate whether the item carrier tray carries an item.

[0080] Optionally, the empty tray detection model can be trained by the following steps:

[0081] The first step is to obtain a first sample set. The first sample in the first sample set includes a sample item tray image, a true background-removed item tray image corresponding to the sample item tray image, and sample target label information indicating whether the tray corresponding to the sample item tray image is empty.

[0082] In the second step, the following empty disk detection model training steps are performed based on the first sample set:

[0083] In a first sub-step, the sample item carrier plate image of at least one first sample in the first sample set is input into a first initial neural network to obtain a sample predicted background-removed item carrier plate image corresponding to each second sample in the at least one first sample.

[0084] In a third sub-step, the predicted sample background-removed object carrier tray image corresponding to each of the at least one first sample is compared with the corresponding true background-removed object carrier tray image to obtain a first comparison result. In practice, the execution entity may use a cross-entropy loss function to determine the difference between the predicted sample background-removed object carrier tray image and the true background-removed object carrier tray image.

[0085] In a fourth sub-step, for each sample item carrier tray image in the at least one first sample, the sample item carrier tray image in the first sample set is updated to a sample predicted background-removed item carrier tray image corresponding to the sample item carrier tray image.

[0086] The fifth sub-step is to determine the updated first sample set as the second sample set.

[0087] The sixth sub-step is to input the background-removed object carrier tray image of at least one second sample in the second sample set into the second initial neural network to obtain sample prediction label information corresponding to each second sample in the above at least one second sample, which characterizes whether the carrier tray corresponding to the background-removed object carrier tray image of the second sample is empty.

[0088] In a seventh sub-step, the sample predicted label information corresponding to each second sample in the at least one second sample is compared with the corresponding sample target label information to obtain a second comparison result. In practice, the execution entity may use a cross-entropy loss function to compare the sample predicted label information and the corresponding sample target label information to determine the difference between the sample predicted label information and the corresponding sample target label information.

[0089] In an eighth sub-step, based on the first comparison result, the second comparison result, and the preset weights, it is determined whether the first initial neural network and the second initial neural network achieve a preset optimization goal. The optimization goal may be minimizing a loss function or maximizing a likelihood function.

[0090] A ninth sub-step, in response to determining that the first initial neural network and the second initial neural network achieve the above-mentioned optimization goal, uses the first initial neural network and the second initial neural network as trained optical character recognition models.

[0091] In a tenth sub-step, in response to determining that the first initial neural network and the second initial neural network have not achieved the optimization goal, network parameters of the first initial neural network and the second initial neural network are adjusted, and the first sample set is formed using unused first samples. The empty disk detection model training step is performed again using the adjusted first initial neural network and the second initial neural network. As an example, the network parameters of the first initial neural network and the second initial neural network can be adjusted using a back propagation algorithm (BP algorithm).

[0092] Step 107 : In response to determining that the empty tray identification result indicates that the item carrying tray carries an item, it is determined that an item identification trigger condition is satisfied.

[0093] In some embodiments, the execution entity may determine that an item recognition trigger condition is satisfied in response to determining that the empty tray recognition result indicates that the item carrying tray carries an item.

[0094] Step 108 : In response to the object recognition trigger condition being met, object recognition is performed on the second object carrier plate image to obtain object recognition information.

[0095] In some embodiments, the execution entity may perform object recognition on the second object carrying tray image in response to satisfying an object recognition trigger condition to obtain object recognition information.

[0096] In some optional implementations of some embodiments, the execution entity may perform object recognition on the second object carrying tray image to obtain object recognition information in response to determining that an object recognition trigger condition is satisfied through the following steps.

[0097] The first step is to obtain a set of item images of a preset item circulation source as a target image set, wherein the target image set includes target images of items corresponding to the second item carrier image.

[0098] The second step is to obtain the item code of each item corresponding to the target image set to obtain an item code set. The item code can be an item number, a self-encoding code, or a mnemonic code.

[0099] The third step is to obtain the item information group of each item corresponding to the target image set to obtain an item information group set. The item information in the item information group may include the item name and item value.

[0100] The fourth step is to generate an item data set based on the item code set and the item information group set. The item data in the item data set includes the item code and the item information group.

[0101] In the fifth step, the second item carrier tray image is input into a pre-trained image extraction model to obtain an image of the item in the second item carrier tray image. The image extraction model may be a neural network model that uses the second item carrier tray image as input data and outputs the item image. For example, the neural network model may be a LeNet model or a CNN model.

[0102] In the sixth step, the above-mentioned object image is input into the pre-trained code recognition model to obtain the object code as the target object code.

[0103] In the seventh step, the item data corresponding to the target item code in the item data set is determined as item identification information.

[0104] Optionally, the execution entity may further control an associated printing device to print the obtained item information onto a label paper, and then control an associated robotic arm to affix the label paper to the item carried on the item carrying tray.

[0105] Optionally, the above-mentioned encoding recognition model can be trained through the following steps:

[0106] The first step is to obtain a third sample set, wherein the third sample in the third sample set includes a sample object image and a sample target object code corresponding to the sample object image.

[0107] In the second step, for the third sample in the third sample set, the following encoding recognition model training steps are performed:

[0108] In a first sub-step, the sample item image from the third sample is input into the input layer of the code recognition network to obtain initial recognition information. The code recognition network includes the input layer, a first feature extraction network layer, a second feature extraction network layer, a third feature extraction network layer, a feature concatenation prediction layer, and an output layer. The input layer can be used for data input. The first feature extraction network layer can be used for feature extraction. The second feature extraction network layer can be used for converting linear features into nonlinear features. The third feature extraction network layer can be used for feature dimensionality reduction. The feature concatenation prediction layer can be used to combine features to determine a final predicted feature vector. The output layer can be used to convert the predicted feature vector into a predicted result and output the predicted result.

[0109] In a second sub-step, the initial identification information is input into the first feature extraction network layer to obtain initial first feature identification information. The first feature extraction network layer may be a convolutional layer. The first feature identification information may be an initial feature map of the object.

[0110] In a third sub-step, the initial first feature recognition information is input into the second feature extraction network layer to obtain second feature recognition information. The second feature extraction network layer may be an activation layer. The second feature recognition information may be a nonlinear feature map.

[0111] In a fourth sub-step, the second feature recognition information is input into the third feature extraction network layer to obtain third feature recognition information. The third feature extraction network layer may be a pooling layer. The third feature recognition information may be a target feature map that reduces the dimensionality of the nonlinear feature map.

[0112] The fifth sub-step is to perform feature vector conversion processing on the third feature identification information to obtain a feature vector. In practice, the execution subject can perform feature vector conversion processing by flattening the third feature identification information to convert the third feature identification information into a feature vector.

[0113] In a sixth sub-step, the feature vector is input to the feature concatenation prediction layer to obtain a predicted target feature vector. The feature concatenation prediction layer may be a fully connected layer.

[0114] The seventh sub-step is to input the above-mentioned predicted target feature vector into the above-mentioned output layer to obtain sample prediction coding information.

[0115] In the eighth sub-step, the sample prediction code information is compared with the sample target item code included in the third sample. In practice, the execution subject may use a cross-entropy loss function to compare and determine the difference between the sample prediction code information and the corresponding sample target item code.

[0116] The ninth sub-step is to determine whether the coding recognition network has achieved a preset optimization goal based on the comparison result.

[0117] In the tenth sub-step, in response to determining that the code recognition network has achieved the above-mentioned optimization goal, the initial neural network is used as the trained code recognition model.

[0118] In the eleventh sub-step, in response to determining that the code recognition network has not achieved the optimization goal, network parameters of the initial neural network are adjusted, and a third sample set is formed using unused third samples. The adjusted code recognition network is used as the code recognition network, and the code recognition model training step is performed again. As an example, the network parameters of the code recognition network can be adjusted using a back propagation algorithm (BP algorithm) and a gradient descent method.

[0119] The above technical solution and its related contents, as an inventive feature of an embodiment of the present disclosure, address the second technical problem mentioned in the background technology: "When performing item information recognition based on directly acquired carrier tray images, interference from the carrier tray image's background results in low accuracy in item information recognition. For labels printed with incorrect item information, item information must be re-recognized, resulting in wasted label paper." Factors that often lead to wasted label paper are as follows: When performing item information recognition based on directly acquired carrier tray images, interference from the carrier tray image's background results in low accuracy in item information recognition. For labels printed with incorrect item information, item information must be re-recognized, resulting in wasted label paper. If these factors are resolved, efficient item recognition can be achieved. To achieve this, the present disclosure employs the following steps: First, a set of item images from a preset item flow source is obtained as a target image set. This target image set includes target images of items corresponding to the second item carrier tray image. Thus, a set of item images from the item flow source is obtained. Second, item codes for each item corresponding to the target image set are obtained to obtain an item code set. Thus, an item code set is obtained. Next, the item information groups of each item corresponding to the target image set are obtained to obtain an item information group set. Thus, an item information group set can be obtained. Next, an item data set is generated based on the item code set and the item information group set. The item data in the item data set includes item codes and item information groups. Thus, an item data set can be obtained based on the item code set and the item information group set. Then, the second item carrier tray image is input into a pre-trained image extraction model to obtain an image of the item in the second item carrier tray image. Thus, the pre-trained image extraction model can be used to obtain an image of the item with the carrier tray background removed. Next, the item image is input into a pre-trained code recognition model to obtain an item code as the target item code. Thus, the pre-trained code recognition model is used to predict the item code corresponding to the item image. Finally, the item data in the item data set corresponding to the target item code is determined as item identification information. Thus, the target item code can be searched for in the data set to obtain item information including the item code. The use of an image extraction model also removes interference from the background of the carrier plate image, improves the accuracy of object information recognition, reduces the number of labels printed with incorrect object information, and thus reduces label waste.

[0120] The above-described embodiments of the present disclosure have the following advantageous effects: Through the object identification methods of some embodiments of the present disclosure, even if the weighing plate fails to read the object's weight during weighing, object identification can be triggered. This broadens the application scenarios of object identification. Specifically, the inability to trigger object information recognition and identify an object is caused by the fact that the weighing plate may fail to read the object's weight, triggering object information recognition based on weight changes. Furthermore, the object information recognition may fail to read the object's weight, resulting in an inability to trigger object information recognition and identify the object. Based on this, the object identification methods of some embodiments of the present disclosure first obtain a set of page screenshots of an item information display page within a preset time period. The page screenshots in the set of page screenshots include the weight display area. Thus, individual screenshots of the item information display page within the preset time period are obtained. Next, based on the set of page screenshots, an image of the weight information display area is generated as a display area image. Thus, an image displaying weight information is obtained. The display area image is then input into a pre-trained optical character recognition model to obtain weight identification information in the display area image. The optical character recognition model can then recognize weight-related information displayed in the display area image. Next, in response to determining that the weight identification information includes a weight identification value and that the weight identification value is greater than a preset value, the following steps are performed: First, determine that an item identification trigger condition is satisfied. The item tray corresponds to the item information display page. Second, obtain an item tray image of the item tray as a first item tray image. Third, perform item identification on the first item tray image to obtain item identification information. Thus, item identification can be triggered and item identification information can be obtained when the identified weight identification information includes a weight identification value greater than a preset value. Next, in response to determining that the weight identification information is empty, obtain an item tray image of the item tray as a second item tray image. Thus, an item tray image can be obtained when the identified weight identification information is empty. The second item tray image is then input into a pre-trained empty tray detection model to obtain an empty tray identification result. The empty tray identification result indicates whether the item tray carries an item. Thus, the empty tray detection model can be used to identify whether the item tray carries an item. Then, in response to determining that the empty tray recognition result indicates that the item carrier tray carries an item, it is determined that an item recognition trigger condition is satisfied. Thus, upon determining that the item carrier tray carries an item, item recognition can be triggered. Then, in response to determining that the item recognition trigger condition is satisfied, item recognition is performed on the second item carrier tray image to obtain item identification information. Thus, item recognition can be performed on the acquired item carrier tray image to obtain item identification information.Furthermore, even when the weight identification information is empty, the image of the item tray can be used to determine whether the item tray is loaded. This allows item identification to continue when the item tray is confirmed to be loaded. Therefore, even if the weighing tray fails during weighing, item identification can still be triggered, even if the item weight cannot be read. This broadens the application scenarios for item identification.

[0121] Further references Figure 2 As an implementation of the methods shown in the figures, the present disclosure provides some embodiments of a web page generation device. These device embodiments are similar to Figure 1 Corresponding to the method embodiments shown, the device can be specifically applied to various electronic devices.

[0122] like Figure 2 As shown, in some embodiments, a webpage generation device 200 includes: a first acquisition unit 201, a generation unit 202, a first input unit 203, an execution unit 204, a second acquisition unit 205, a second input unit 206, a determination unit 207, and an identification unit 208. The first acquisition unit 201 is configured to acquire a set of page screenshots of an item information display page within a preset time period. The page screenshots in the set of page screenshots include a weight display area. The generation unit 202 is configured to generate a weight display area image as a display area image based on the set of page screenshots. The first input unit 203 is configured to input the display area image into a pre-trained optical character recognition model to obtain weight identification information in the display area image. The execution unit 204 is configured to, in response to determining that the weight identification information includes a weight identification value and that the weight identification value is greater than a preset value, perform the following steps: determining that an item identification trigger condition is satisfied. The item carrier tray corresponds to the item information display page; an item carrier tray image of the item carrier tray is obtained as a first item carrier tray image; item recognition is performed on the first item carrier tray image to obtain item recognition information; a second acquisition unit 205 is configured to, in response to determining that the weight recognition information is empty, obtain an item carrier tray image of the item carrier tray as a second item carrier tray image; a second input unit 206 is configured to input the second item carrier tray image into a pre-trained empty tray detection model to obtain an empty tray recognition result. The empty tray recognition result indicates whether the item carrier tray carries an item; the determination unit 207 is configured to, in response to determining that the empty tray recognition result indicates that the item carrier tray carries an item, determine that an item recognition trigger condition is satisfied; and the recognition unit 208 is configured to, in response to determining that the item recognition trigger condition is satisfied, perform item recognition on the second item carrier tray image to obtain item recognition information.

[0123] It is understood that the units described in the device 200 are similar to those described in the reference Figure 1 Therefore, the operations, features and beneficial effects described above for the method are also applicable to the device 200 and the units included therein, and will not be repeated here.

[0124] Reference below Figure 3 , which shows a structural diagram of an electronic device 300 suitable for implementing some embodiments of the present disclosure. Figure 3 The electronic device shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present disclosure.

[0125] like Figure 3 As shown, the electronic device 300 may include a processing device (e.g., a central processing unit, a graphics processing unit, etc.) 301, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 302 or a program loaded from a storage device 308 into a random access memory (RAM) 303. Various programs and data required for the operation of the electronic device 300 are also stored in the RAM 303. The processing device 301, the ROM 302, and the RAM 303 are connected to each other via a bus 304. An input / output (I / O) interface 305 is also connected to the bus 304.

[0126] Typically, the following devices may be connected to the I / O interface 305: an input device 306 including, for example, a touch screen, a touchpad, a keyboard, a mouse, a camera, a microphone, an accelerometer, a gyroscope, etc.; an output device 307 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, etc.; a storage device 308 including, for example, a magnetic tape, a hard disk, etc.; and a communication device 309. The communication device 309 may allow the electronic device 300 to communicate with other devices wirelessly or by wire to exchange data. Figure 3 The electronic device 300 is shown with various devices, but it should be understood that it is not required to implement or possess all of the devices shown. More or fewer devices may be implemented or possessed instead. Figure 3 Each block shown in the figure may represent one device, or may represent multiple devices as needed.

[0127] In particular, according to some embodiments of the present disclosure, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, some embodiments of the present disclosure include a computer program product comprising a computer program carried on a computer-readable medium, the computer program comprising program code for executing the method shown in the flowchart. In some such embodiments, the computer program can be downloaded and installed from a network via the communication device 309, or installed from the storage device 308, or installed from the ROM 302. When the computer program is executed by the processing device 301, the functions defined in the methods of some embodiments of the present disclosure are performed.

[0128] It should be noted that the computer-readable medium described in some embodiments of the present disclosure may be a computer-readable signal medium or a computer-readable storage medium, or any combination of the two. The computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or device, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In some embodiments of the present disclosure, the computer-readable storage medium may be any tangible medium that contains or stores a program that can be used by or in conjunction with an instruction execution system, device, or device. In some embodiments of the present disclosure, the computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, which carries computer-readable program code. Such a propagated data signal may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. A computer-readable signal medium may also be any computer-readable medium other than a computer-readable storage medium that can transmit, propagate, or transport a program for use by or in conjunction with an instruction execution system, apparatus, or device. The program code contained on the computer-readable medium may be transmitted using any suitable medium, including but not limited to wires, optical cables, RF (radio frequency), etc., or any suitable combination thereof.

[0129] In some embodiments, the client and server can communicate using any currently known or future developed network protocol, such as HTTP (HyperText Transfer Protocol), and can be interconnected with any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include a local area network ("LAN"), a wide area network ("WAN"), an internet (e.g., the Internet), and a peer-to-peer network (e.g., an ad hoc peer-to-peer network), as well as any currently known or future developed network.

[0130] The computer-readable medium may be included in the electronic device mentioned above; or it may exist independently without being assembled into the electronic device. The computer-readable medium carries one or more programs. When the one or more programs are executed by the electronic device, the electronic device: obtains a set of page screenshots of the item information display page within a preset time period, wherein the page screenshots in the above page screenshot set include a weight display area; generates a weight display area image as a display area image based on the above page screenshot set; inputs the above display area image into a pre-trained optical character recognition model to obtain weight identification information in the above display area image; in response to determining that the above weight identification information includes a weight identification value, and the above weight identification value is greater than a preset value, performs the following steps: determines that the item identification trigger condition is met, wherein the above item carrier plate corresponds to the above item information display page; Acquire an item carrying tray image of the item carrying tray as a first item carrying tray image; perform item recognition on the first item carrying tray image to obtain first item recognition information; in response to determining that the weight recognition information is empty, acquire an item carrying tray image of the item carrying tray as a second item carrying tray image; input the second item carrying tray image into a pre-trained empty tray detection model to obtain an empty tray recognition result, wherein the empty tray recognition result indicates whether the item carrying tray carries an item; in response to determining that the empty tray recognition result indicates that the item carrying tray carries an item, determine that an item recognition trigger condition is satisfied; in response to determining that the item recognition trigger condition is satisfied, perform item recognition on the second item carrying tray image to obtain second item recognition information.

[0131] Computer program code for performing the operations of some embodiments of the present disclosure may be written in one or more programming languages, or a combination thereof, including object-oriented programming languages such as Java, Smalltalk, C++, and conventional procedural programming languages such as "C" or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, as a stand-alone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving a remote computer, the remote computer may be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., through the Internet using an Internet service provider).

[0132] The flowcharts and block diagrams in the accompanying drawings illustrate the possible implementation architecture, functions and operations of the systems, methods and computer program products according to various embodiments of the present disclosure. In this regard, each box in the flowchart or block diagram can represent a module, program segment, or a part of code, and the module, program segment, or a part of code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order than that marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram and / or flowchart, and the combination of the boxes in the block diagram and / or flowchart, can be implemented with a dedicated hardware-based system that performs the specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0133] The units described in some embodiments of the present disclosure may be implemented in software or hardware. The units described may also be provided in a processor. For example, they may be described as follows: a processor comprising a first acquisition unit, a generation unit, a first input unit, an execution unit, a second acquisition unit, a second input unit, a determination unit, and an identification unit. The names of these units do not, in some cases, constitute limitations on the units themselves. For example, the first acquisition unit may also be described as a "unit for acquiring a collection of page screenshots of the item information display page within a preset time period."

[0134] The functions described above herein may be performed, at least in part, by one or more hardware logic components. For example, and without limitation, exemplary types of hardware logic components that may be used include: field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chip (SOCs), complex programmable logic devices (CPLDs), and the like.

[0135] The above description is only an illustration of some preferred embodiments of the present disclosure and the technical principles used. Those skilled in the art should understand that the scope of the invention involved in the embodiments of the present disclosure is not limited to the technical solutions formed by the specific combination of the above-mentioned technical features, but should also cover other technical solutions formed by any combination of the above-mentioned technical features or their equivalent features without departing from the above-mentioned inventive concept. For example, the above-mentioned features are replaced with (but not limited to) technical features with similar functions disclosed in the embodiments of the present disclosure.

Claims

1. A method for identifying an object, comprising: Obtaining a set of page screenshots of an item information display page within a preset time period, wherein the page screenshots in the set of page screenshots include a weight display area; generating a weight display area image as a display area image according to the page screenshot set; Inputting the display area image into a pre-trained optical character recognition model to obtain weight recognition information in the display area image; In response to determining that the weight identification information includes a weight identification value, and the weight identification value is greater than a preset value, performing the following steps: Determining that an item identification trigger condition is satisfied, wherein the item carrier plate corresponds to the item information display page; acquiring an item carrying tray image of the item carrying tray as a first item carrying tray image; performing object recognition on the first object carrying tray image to obtain object recognition information; In response to determining that the weight identification information is empty, acquiring an item carrying tray image of the item carrying tray as a second item carrying tray image; Inputting the second article carrying tray image into a pre-trained empty tray detection model to obtain an empty tray recognition result, wherein the empty tray recognition result indicates whether the article carrying tray carries an article; In response to determining that the empty tray identification result indicates that the item carrying tray carries an item, determining that an item identification trigger condition is satisfied; In response to determining that the item recognition trigger condition is met, item recognition is performed on the second item carrying tray image to obtain item recognition information.

2. The method according to claim 1, wherein Generating a weight display area image as a display area image according to the page screenshot set includes: Performing grayscale processing on each page screenshot in the page screenshot set to obtain a page grayscale screenshot set; Dividing the page grayscale screenshot set in a sliding window manner to obtain a page grayscale screenshot group set, wherein the page grayscale screenshot group in the page grayscale screenshot group set includes a first page grayscale screenshot and a second page grayscale screenshot; For each page grayscale screenshot group in the page grayscale screenshot group set, perform the following steps: Determining pixel values of respective pixels in the first page grayscale screenshot included in the page grayscale screenshot group as a first pixel value set; Determining the pixel value of each pixel point in the second page grayscale screenshot included in the page grayscale screenshot group as a second pixel value set; determining an average pixel difference value based on the first pixel value set and the second pixel value set; Determining the determined average values of the pixel differences as a pixel difference average value set; Selecting a pixel difference average value that meets a preset numerical condition from the pixel difference average value set as a target pixel difference average value; Determining a page grayscale screenshot group corresponding to the target pixel difference average value as a target page grayscale screenshot group; generating a difference image according to the first page grayscale screenshot and the second page grayscale screenshot included in the target page grayscale screenshot group; Determine the pixel value of each pixel point in the difference image as a pixel value set; In response to determining that each pixel value in the pixel value set is greater than a preset threshold, determining the position information of each pixel point corresponding to the pixel value set as a change area pixel point position information set; determining a weight display area in the second page grayscale screenshot included in the target page grayscale screenshot group according to the pixel position information set of the changed area; A weight display area image is generated based on the determined weight display area.

3. The method according to claim 2, wherein: The grayscale processing is performed on each page screenshot in the page screenshot set to obtain the page grayscale screenshot set, including: For each page screenshot in the page screenshot collection, perform the following steps: Determine each pixel in the page screenshot as a pixel set; For each pixel in the pixel set, perform the following pixel processing steps: Determine the values of the red, green and blue color channels of the pixel as a color value set; Generate a color average as a grayscale value according to each color value in the color value set; Updating the pixel value of the pixel point to the grayscale value; Determine the page screenshot after each pixel is processed by the pixel processing step as a page grayscale screenshot; The determined page grayscale screenshots are determined as a page grayscale screenshot set.

4. The method according to claim 2, wherein: Generating a difference image based on the first page grayscale screenshot and the second page grayscale screenshot included in the target page grayscale screenshot group includes: Determining the pixel value of each pixel point in the second page grayscale screenshot as a second page grayscale screenshot pixel value set; For each pixel in the grayscale screenshot of the first page, perform the following steps: Determine the pixel value of the pixel point as the pixel value of the grayscale screenshot of the first page; generating a target pixel value according to the first page grayscale screenshot pixel value and a pixel value corresponding to the first page grayscale screenshot pixel value in the second page grayscale screenshot pixel value set; Updating the pixel value of the pixel point to the target pixel value to obtain an updated pixel point; Generate a difference map based on each updated pixel point.

5. The method according to claim 1, wherein The optical character recognition model is trained by the following steps: Acquire a sample set, wherein the samples in the sample set include a screenshot of a sample weight display area and sample target weight numerical value information corresponding to the screenshot of the sample weight display area; The following training steps are performed based on the sample set: Inputting a screenshot of the sample weight display area of at least one sample in the sample set into the initial neural network to obtain sample identification weight numerical information corresponding to each sample in the at least one sample; Comparing the sample identification weight numerical information corresponding to each sample in the at least one sample with the corresponding sample target weight numerical information to obtain an average loss value; Optimizing parameters of the initial neural network based on the loss average and a preset optimizer; In response to the current number of training times being the preset number of training times, using the initial neural network after parameter optimization as a trained optical character recognition model; In response to the current number of training times being less than the preset number of training times, the initial neural network after parameter optimization is used as the initial neural network, and unused samples are used to form a sample set, and the above training steps are performed again.

6. The method according to claim 1, wherein The empty disk detection model is trained by the following steps: Acquire a first sample set, wherein a first sample in the first sample set includes a sample item carrier tray image, a true background-removed item carrier tray image corresponding to the sample item carrier tray image, and sample target label information indicating whether the carrier tray corresponding to the sample item carrier tray image is empty; Perform the following empty disk detection model training steps based on the first sample set: Inputting a sample item carrier tray image of at least one first sample in the first sample set into a first initial neural network to obtain a sample predicted background-removed item carrier tray image corresponding to each first sample in the at least one first sample; Comparing the sample predicted background-removed article-carrying tray image corresponding to each first sample of the at least one first sample with the corresponding true background-removed article-carrying tray image to obtain a first comparison result; For each sample item carrier tray image in the at least one first sample, updating the sample item carrier tray image in the first sample set to a sample predicted background-removed item carrier tray image corresponding to the sample item carrier tray image; determining the updated first sample set as the second sample set; Inputting a background-removed item carrier tray image of at least one second sample in the second sample set into a second initial neural network, obtaining sample prediction label information corresponding to each second sample in the at least one second sample, indicating whether a carrier tray corresponding to the background-removed item carrier tray image of the second sample is empty; Comparing the sample prediction label information corresponding to each second sample in the at least one second sample with the corresponding sample target label information to obtain a second comparison result; Determining whether the first initial neural network and the second initial neural network achieve a preset optimization goal based on the first comparison result, the second comparison result and the preset weight; In response to determining that the first initial neural network and the second initial neural network achieve the optimization goal, using the first initial neural network and the second initial neural network as trained optical character recognition models; In response to determining that the first initial neural network and the second initial neural network do not achieve the optimization goal, the network parameters of the first initial neural network and the second initial neural network are adjusted, and the first sample set is composed of unused first samples. The empty disk detection model training step is performed again using the adjusted first initial neural network and the second initial neural network.

7. The method according to claim 1, wherein The method further comprises: Controlling the associated printing device to print the obtained item information onto label paper; The associated mechanical arm is controlled to stick the label paper onto the items carried in the item carrying tray.

8. An object identification device comprising: a first acquiring unit configured to acquire a set of page screenshots of an item information display page within a preset time period, wherein the page screenshots in the set of page screenshots include a weight display area; a generating unit configured to generate a weight display area image as a display area image based on the page screenshot set; a first input unit configured to input the display area image into a pre-trained optical character recognition model to obtain weight recognition information in the display area image; The execution unit is configured to, in response to determining that the weight identification information includes a weight identification value and the weight identification value is greater than a preset value, perform the following steps: determining that an item identification trigger condition is satisfied, wherein the item carrier tray corresponds to the item information display page; obtaining an item carrier tray image of the item carrier tray as a first item carrier tray image; performing item identification on the first item carrier tray image to obtain item identification information; a second acquiring unit configured to acquire, in response to determining that the weight identification information is empty, an item carrying tray image of the item carrying tray as a second item carrying tray image; a second input unit configured to input the second article carrying tray image into a pre-trained empty tray detection model to obtain an empty tray recognition result, wherein the empty tray recognition result indicates whether the article carrying tray carries an article; a determining unit configured to, in response to determining that the empty tray identification result indicates that the item carrying tray carries an item, determine that an item identification trigger condition is satisfied; The recognition unit is configured to perform object recognition on the second object carrying tray image in response to determining that an object recognition trigger condition is met, to obtain object recognition information.

9. An electronic device comprising: one or more processors; a storage device having one or more programs stored thereon; When the one or more programs are executed by the one or more processors, the one or more processors implement the method according to any one of claims 1 to 7.

10. A computer-readable medium having a computer program stored thereon, wherein: When the program is executed by a processor, the method according to any one of claims 1 to 7 is implemented.

Images