Remote warehouse learning method and device based on artificial intelligence, and medium

Abstract

The invention discloses a remote warehouse learning method and device based on artificial intelligence, and a medium, and relates to the technical field of data processing based on management purposes. The method comprises the following steps: acquiring a cargo image of a to-be-identified cargo in a warehouse, and detecting the cargo image to determine whether the to-be-identified cargo meets a storage constraint condition corresponding to a cargo type to which the to-be-identified cargo belongs; if not, target goods which do not meet the storage constraint condition are screened out, and a remote loading instruction for image acquisition equipment is generated according to the distribution relation between the key recognition area in the target goods and the image acquisition equipment arranged in the warehouse; and based on the remote loading instruction, the image acquisition device is controlled to identify the key identification area so as to generate cargo evaluation information for the target cargo according to an obtained identification result, and early warning prompt is performed according to the cargo evaluation information.

Description

Technical Field

[0001] This application relates to the technical field of data processing for management purposes, and particularly to a remote warehouse goods recognition method, device, and medium based on artificial intelligence. Background Art

[0002] In current warehouse management, the goods recognition process mainly relies on on-site personnel to conduct on-site inspections. However, when the personnel are far away from the warehouse where the goods are stored, this traditional goods recognition method is particularly inconvenient. It is not only inefficient but also may lead to management oversights due to information lag caused by the distance. With the development of video goods recognition technology, although the convenience of goods recognition has been improved to a certain extent, it is limited by the video resolution and the recognition ability of the human eye. Especially for some special goods, it is often difficult to guarantee the recognition accuracy, and there is a relatively high risk of error. Summary of the Invention

[0003] To solve the above problems, this application proposes a remote warehouse goods recognition method based on artificial intelligence, including:

[0004] Collect the goods images of the goods to be recognized in the warehouse, and detect the goods images to determine whether the goods to be recognized meet the storage constraint conditions corresponding to their respective goods types;

[0005] If not, screen out the target goods that do not meet the storage constraint conditions, and generate a remote loading instruction for the image acquisition device according to the distribution relationship between the key recognition regions in the target goods and the image acquisition devices set in the warehouse;

[0006] Based on the remote loading instruction, control the image acquisition device to recognize the key recognition regions, generate goods evaluation information for the target goods according to the obtained recognition results, and perform early warning prompts according to the goods evaluation information.

[0007] In an implementation manner of this application, the storage constraint conditions include quality constraint conditions and space constraint conditions. Generating a remote loading instruction for the image acquisition device according to the distribution relationship between the key recognition regions in the target goods and the image acquisition devices set in the warehouse specifically includes:

[0008] When the target goods do not meet the quality constraint conditions, screen out the designated image acquisition devices within the coverage range of the area where the target goods are located from the image acquisition devices in the warehouse according to the abnormal risk information corresponding to each key recognition region in the target goods;

[0009] Generate a first remote loading instruction for the image acquisition device according to the distribution relationship of the spatial positions between the specified image acquisition device and the target goods, and the concentration of the target goods around the specified image acquisition device.

[0010] In an implementation manner of the present application, after generating the remote loading instruction for the image acquisition device, the method further includes:

[0011] Obtain the appearance frequency of the target goods within a preset observation period, and determine the instruction type corresponding to the target goods according to the appearance frequency; wherein, the instruction type includes a single instruction and a combined instruction;

[0012] When the instruction type of the target goods is the combined instruction, identify the auxiliary loading instruction sent by the user, and correct the auxiliary loading instruction according to the detection requirements of the key identification area, so as to jointly load the corrected auxiliary loading instruction and the remote loading instruction when identifying the key identification area.

[0013] In an implementation manner of the present application, detecting the goods image to determine whether the goods to be identified meet the warehousing constraint conditions corresponding to its goods type specifically includes:

[0014] Detect the goods image to determine the goods type corresponding to the goods to be identified, and identify the position identification area and the quality detection dimension corresponding to the goods to be identified;

[0015] Obtain the warehousing information corresponding to the goods type, and compare the position identification information carried in the position identification area with the warehousing information to determine whether the position identification information matches the warehousing information;

[0016] And, based on the quality detection dimension, match the goods image with the standard image of the goods to be identified to identify whether there are quality defects in the goods image;

[0017] In the case that the position identification information does not match the warehousing information and / or there are quality defects in the goods image, determine that the goods to be identified do not meet the warehousing constraint conditions corresponding to its goods type.

[0018] In an implementation manner of the present application, after matching the goods image with the standard image of the goods to be identified based on the quality detection dimension, the method further includes:

[0019] According to the goods type, determine whether the goods to be identified need to be subjected to linkage detection. If so, determine the linkage device corresponding to the goods to be identified;

[0020] Obtain the sensing data collected by the linkage device, and identify whether there are quality defects in the quality inspection area based on the sensing data and the standard image.

[0021] In an implementation manner of the present application, the remote loading instruction further includes a second remote loading instruction, and a remote loading instruction for the image acquisition device is generated according to the distribution relationship between the key recognition area in the target goods and the image acquisition devices arranged in the warehouse, specifically including:

[0022] When the target goods do not meet the space constraint conditions, determine the target image acquisition device with the closest spatial distance to the target goods, and generate a second remote loading instruction for the target image acquisition device according to the distribution relationship between the target image acquisition device and the key recognition area.

[0023] An embodiment of the present application provides an artificial intelligence-based remote warehouse goods recognition device, and the device includes:

[0024] At least one processor;

[0025] And a memory communicatively connected to the at least one processor;

[0026] Wherein, the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can:

[0027] Collect the goods images of the goods to be recognized in the warehouse, and detect the goods images to determine whether the goods to be recognized meet the storage constraint conditions corresponding to their respective goods types;

[0028] If not, screen out the target goods that do not meet the storage constraint conditions, and generate a remote loading instruction for the image acquisition device according to the distribution relationship between the key recognition area in the target goods and the image acquisition devices arranged in the warehouse;

[0029] Based on the remote loading instruction, control the image acquisition device to recognize the key recognition area, generate goods evaluation information for the target goods according to the obtained recognition result, and perform early warning prompts according to the goods evaluation information.

[0030] In an implementation manner of the present application, the storage constraint conditions include quality constraint conditions and space constraint conditions, and the processor can also:

[0031] In the case that the target goods do not meet the quality constraint conditions, according to the abnormal risk information corresponding to each key identification area in the target goods, specify the image acquisition devices within the coverage area of the area where the target goods are located from the image acquisition devices in the warehouse;

[0032] Generate a first remote loading instruction for the image acquisition device according to the distribution relationship of the spatial positions between the image acquisition device and the target goods, and the concentration of the target goods around the specified image acquisition device.

[0033] In an implementation manner of the present application, the processor can also:

[0034] Obtain the occurrence frequency of the target goods within a preset observation period, and determine the instruction type corresponding to the target goods according to the occurrence frequency; wherein, the instruction type includes a single instruction and a combined instruction;

[0035] In the case that the instruction type of the target goods is the combined instruction, identify the auxiliary loading instruction sent by the user, and correct the auxiliary loading instruction according to the detection requirements required by the key identification area, so as to jointly load the corrected auxiliary loading instruction and the remote loading instruction when identifying the key identification area.

[0036] An embodiment of the present application provides a non-volatile computer storage medium storing computer-executable instructions, and the computer-executable instructions are set as:

[0037] A method for remotely identifying goods in a warehouse based on artificial intelligence as described in any one of the above.

[0038] A method for remotely identifying goods in a warehouse based on artificial intelligence proposed by the present application can bring the following

[0039] Beneficial effects:

[0040] By remotely identifying and evaluating goods, there is no need for personnel to go to the site for inspection, especially suitable for the situation where personnel are far from the warehouse, greatly improving the convenience and efficiency of identifying goods. At the same time, based on the preliminary goods image recognition results, control the image acquisition device to perform secondary detection on the key identification areas of the goods within a certain range. Compared with the traditional video goods identification technology limited by resolution and human eye recognition, it can effectively improve the identification accuracy of special goods and reduce the error risk. In addition, generate goods evaluation information in a timely manner based on the recognition results and give early warning prompts, enabling managers to quickly understand the goods situation, handle problems in a timely manner, avoid management omissions, and enhance the timeliness and accuracy of warehouse management. Description of the Drawings

[0041] The accompanying drawings described herein are used to provide a further understanding of the present application and form a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation to the present application. In the drawings:

[0042] Figure 1 It is a schematic flowchart of a method for identifying goods in a remote warehouse based on artificial intelligence provided by an embodiment of the present application;

[0043] Figure 2 It is a schematic diagram of a device for identifying goods in a remote warehouse based on artificial intelligence provided by an embodiment of the present application. Detailed implementation manners

[0044] To make the objectives, technical solutions and advantages of the present application clearer, the technical solutions of the present application will be clearly and completely described below in conjunction with the specific embodiments of the present application and the corresponding accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present application.

[0045] The following will, with reference to the accompanying drawings, detail the technical solutions provided by each embodiment of the present application.

[0046] As Figure 1 shown, a method for identifying goods in a remote warehouse based on artificial intelligence provided by an embodiment of the present application includes:

[0047] S101: Collect the goods images of the goods to be identified in the warehouse, and detect the goods images to determine whether the goods to be identified meet the storage constraint conditions corresponding to their respective goods types.

[0048] During the process of warehouse management, various labels will be attached to the goods in the warehouse, such as two-dimensional codes, barcodes or radio frequency tags, etc. These labels usually contain the identification information of the goods (such as item numbers, names, specifications, etc.) and the storage location information (such as shelf numbers, bin numbers, etc.). At the same time, multiple image acquisition devices, such as high-definition cameras, will be installed in the warehouse for image acquisition and monitoring of the goods. These image acquisition devices are distributed at different positions in the warehouse to ensure that the entire warehouse area can be covered.

[0049] For each type of goods in the warehouse, there are corresponding warehousing constraint conditions. Warehousing constraint conditions refer to the restrictions and conditions set for the storage of goods. One type is quality constraint conditions, which are used to determine whether quality problems occur during the warehousing process for certain special goods (such as fresh food, medicines, textiles, etc.). Another type is space constraint conditions, which are used to determine whether the goods are stored in the correct warehousing location. Therefore, when conducting remote goods identification, the image acquisition device in the warehouse can be called to collect the goods image of the goods to be identified. By detecting the goods image, it can be determined whether the goods to be identified meet the warehousing constraint conditions corresponding to their goods types, so as to judge whether there are abnormalities in the currently stored goods to be identified.

[0050] Specifically, before identifying the goods image, it is first necessary to obtain a pre-trained neural network model. Here, the neural network model can adopt models such as YOLOv8. The neural network model is used to detect the goods image, so as to determine the goods type corresponding to the goods to be identified as belonging to categories such as food, electronic products, clothing, etc. through the probability distribution output by the Softmax layer. At the same time, the model can also identify the area containing the goods position identification information, such as QR codes, RFID tags or text coding areas, etc. These position identification areas can judge whether the goods are stored in the correct warehousing location. In addition, according to the identified goods type, the model can also determine the quality inspection dimensions used for detecting the goods. The quality inspection dimensions include the integrity, size specifications, color, etc. of the goods. Taking the goods type of steel as an example, its corresponding quality inspection dimensions may include the length, width, thickness of the steel, and whether there are defects on the surface, etc.

[0051] Therefore, after identifying the position identification area and quality inspection dimensions corresponding to the goods to be identified, the goods image is identified based on these two types of information, so as to determine whether the goods to be identified meet the warehousing constraint conditions corresponding to their goods types.

[0052] On the one hand, for the space constraint conditions, it is necessary to determine whether the currently identified goods are stored in the correct warehousing location based on the warehousing information corresponding to the goods type. By retrieving the database according to the goods type, the warehousing information corresponding to the goods type can be obtained. The warehousing information includes the goods number, name, specifications, storage location requirements, etc. of the goods. The storage location requirements limit the storage location of the goods. For example, for fragile ceramic products, they are required to be stored on shelves with a lower height and better buffer packaging during warehousing. Therefore, after obtaining the warehousing information corresponding to the goods type, the position identification information carried in the position flag area is compared with the storage location requirements in the warehousing information. By judging whether the position identification information matches the warehousing information, it can be determined whether the goods are stored in the correct location. If they do not match, it means that the goods to be identified do not meet their corresponding space constraint conditions. At this time, the warehousing location of the goods to be identified needs to be changed.

[0053] On the other hand, as for whether the quality constraint conditions are met, it can only be determined after the quality inspection dimension conducts quality inspection on the goods image. First, obtain the standard image corresponding to the goods to be identified. The standard image refers to the standard appearance image of this type of goods when the quality is qualified. Then, match the goods image with the standard image, detect the corresponding areas in the goods image based on the image features in each quality inspection dimension, and through feature comparison, it can be identified whether there are quality defects in the goods image. If there are quality defects, it means that the goods to be identified do not meet their corresponding quality constraint conditions. At this time, there are certain problems with the quality of the goods to be identified, and the goods need to be processed in time, and at the same time, such goods should be avoided from being sold externally.

[0054] It should be noted that the quality defects of the goods cannot be fully identified only based on the goods image. Before conducting quality inspection on the goods to be identified through the standard image, it is also necessary to judge whether linkage inspection is required according to the type of the goods to be identified. Different types of goods may have different characteristics and quality inspection requirements. Some goods can more accurately judge the quality only through image matching, while some goods need to combine other inspection means to more comprehensively evaluate their quality status. If it is determined that linkage inspection is required, then the linkage device corresponding to the goods to be identified needs to be determined, and the sensing data collected by the linkage device is combined with the standard image to comprehensively analyze whether there are quality defects in the quality inspection area. For example, for fresh goods, in addition to image monitoring, the environmental data collected by sensors can also be used to assist in detecting the quality of the goods. If the environmental data does not meet the best storage requirements of the goods to be identified, it can be judged that the goods to be identified may have certain quality defects due to unsuitable environment. The standard image provides the quality standard of the goods in terms of appearance, while the sensing data provides more in-depth quality information. By combining image matching and linkage inspection, the quality status of the goods can be comprehensively evaluated from multiple angles and dimensions, avoiding missed inspections or misjudgments that may occur by relying solely on single image matching, and improving the accuracy and reliability of quality inspection.

[0055] According to the above detection process, when the position identification information does not match the warehousing information, and / or there are quality defects in the goods image, it means that the goods to be identified do not meet the warehousing constraint conditions.

[0056] S102: If not, screen out the target goods that do not meet the warehousing constraint conditions, and generate a remote loading instruction for the image acquisition device according to the distribution relationship between the key identification area in the target goods and the image acquisition devices set in the warehouse.

[0057] Through the detection method described above, target goods that do not meet the warehousing constraint conditions can be screened out from the goods to be recognized. These target goods may have abnormalities in warehousing location or quality inspection. However, this recognition method is a global image recognition based on the overall image of the goods to be recognized. Although it can quickly screen out goods with abnormalities, limited by the video resolution, for some subtle abnormalities in key parts, such as small characters on the goods label, tiny defects on the product surface, etc., it may not be able to accurately recognize them.

[0058] To solve the problem of limited accuracy of goods recognition, a remote loading instruction for the image acquisition device can be generated according to the distribution relationship between the key recognition areas in the target goods and the image acquisition devices set in the warehouse. The key recognition area refers to the area containing important information, such as the location identification area, and the key quality inspection areas on the surface of the goods, such as the appearance defect area, the size specification inspection area, etc. Capturing the key recognition areas of the goods to be recognized and performing secondary recognition on them can obtain a clearer image, so as to more accurately judge the storage location and quality of the goods. The remote loading instruction is an instruction used to remotely control the shooting angle, shooting frequency, resolution, etc. of the image acquisition device. According to the distribution relationship between the key recognition areas of the target goods and the image acquisition devices set in the warehouse, adjusting the parameters of the image acquisition device can obtain clearer image information, improving the accuracy and reliability of remote goods recognition. At the same time, the remote loading instruction can automatically adjust the working state of the image acquisition device, and a clear image of the key recognition area can be obtained without manual intervention. This greatly reduces the time and workload of manual inspection and adjustment, improving the efficiency of remote goods recognition.

[0059] In one embodiment, when the target goods do not meet the quality constraint conditions, it is necessary to screen out the designated image acquisition devices within the coverage area of the target goods from the image acquisition devices in the warehouse based on the abnormal risk information corresponding to each key recognition area in the target goods, ensuring that the selected designated image acquisition devices can clearly capture the images of the key recognition areas.

[0060] The abnormal risk information refers to the degree of risk of abnormality occurring in the key recognition area, which determines the size of the spatial range to be considered when screening the designated image acquisition device. The greater the risk, the higher the probability of abnormality occurring in the key recognition area. In order to identify whether there is an abnormality in this area, a wider and more comprehensive area range needs to be considered to capture the key recognition areas that may have abnormalities from multiple perspectives. When determining the abnormal risk information corresponding to the key recognition area, it can be achieved through the following steps:

[0061] First, extract the historical identification record information of goods of the same type as the target goods in the past period from the warehouse management system or database. The historical identification record information contains various information when the target goods of this type were identified before, such as the performance of the key identification area of the goods in the image, whether there are abnormal situations, and abnormal information, etc.

[0062] Then, by analyzing the historical identification record information, calculate the abnormal change rate corresponding to each key identification area of the target goods at different time points. The abnormal change rate represents the abnormal development situation in the key identification area. The larger the abnormal change rate, the faster the abnormal change of the target goods, and the more attention should be paid to the abnormal situation in this area.

[0063] Secondly, determine the abnormal risk coefficient corresponding to the key identification area according to the abnormal change rate. The abnormal risk coefficient is a quantitative indicator used to measure the likelihood of an abnormality occurring in the key identification area during future warehousing. For example, for the key identification area on the surface of the electronic product shell, the abnormal change rate is relatively high, and the standard value of the abnormal risk coefficient is 1. After evaluation, its abnormal risk coefficient may be determined to be 0.8, indicating a very high risk.

[0064] Finally, based on the abnormal risk coefficient, determine the corresponding area range value of each key identification area in the warehouse space. The area range value is used to define the spatial range related to the key identification area that needs to be considered when screening image acquisition devices. Among them, there is a positive correlation between the abnormal risk coefficient and the area range value. The higher the abnormal risk coefficient, the larger the corresponding area range value, which means that when screening image acquisition devices, a wider area range needs to be considered because the key identification area with a higher risk requires a wider monitoring range to increase the possibility of detecting potential abnormalities, so as to ensure that measures can be taken in time to avoid losses of goods or warehousing management problems caused by abnormal situations.

[0065] After screening out the specified image acquisition device according to the risk information of the key identification area, it is necessary to generate a first remote loading instruction for the image acquisition device according to the distribution relationship of the spatial positions between the specified image acquisition device and the target goods, and the concentration of the target goods around the specified image acquisition device. The first remote loading instruction is an instruction to accurately photograph and locate the key identification area by controlling the moving route and shooting frequency of the image acquisition device when the goods to be identified do not meet the quality constraint conditions, and is used to verify whether there are quality abnormalities in the key identification area. Here, the key identification area refers to the key quality detection area.

[0066] Specifically, according to the spatial position distribution relationship between the specified image acquisition device in the warehouse and the target goods, with the central acquisition point in the specified image acquisition device as a reference, a movement route for the central acquisition point to move from the current position to the target goods is generated based on the spatial position of the target goods. This can ensure that the image acquisition device can accurately align with the target goods, improving the accuracy and effectiveness of image acquisition. It should be noted that the above process realizes the adjustment of the shooting angle of the specified image acquisition device. Also, during the process of controlling the movement of the specified image acquisition device, the focal length of the specified image acquisition device can be adjusted accordingly, so that the specified image acquisition device after adjusting the focal length can obtain an enlarged key recognition area, while ensuring that the shooting range can cover all target goods to be acquired.

[0067] Through the spatial position distribution relationship, the adjustment of the movement route of the specified image acquisition device is realized. During this process, the workload situation of each specified image acquisition device needs to be considered, that is, the distribution of the target goods that the specified image acquisition device needs to cover. Therefore, it is also necessary to calculate the concentration degree of the target goods around the specified image acquisition device. The concentration degree refers to the density of the target goods around the specified image acquisition device. The higher the concentration degree, the more target goods there are around the specified image acquisition device, and the more target goods need to be monitored simultaneously. Therefore, according to the concentration degree, the acquisition frequency of the specified image acquisition device can be adjusted. The higher the concentration of the surrounding goods, the more likely it is necessary to increase the acquisition frequency of the specified image acquisition device in order to obtain more image information and timely capture the quality anomalies in the key recognition area, realizing more detailed monitoring and analysis of the target goods.

[0068] In one embodiment, when the target goods do not meet the spatial constraint conditions corresponding to their respective goods types, that is, the storage location of the goods does not conform to the specified storage location in the warehouse, it is necessary to further determine its specific storage location and generate a corresponding remote loading instruction.

[0069] Specifically, among all the image acquisition devices in the warehouse, by calculating the spatial distance between each device and the target goods, the target image acquisition device with the closest spatial distance to the target goods is selected. This can calculate the straight-line distance from each camera to the target goods according to the installation position coordinates of each camera and the current position coordinates of the target goods, using a spatial distance formula such as the Euclidean distance formula. The camera with the smallest distance is the closest target image acquisition device.

[0070] According to the distribution relationship between the nearest target image acquisition device and the key identification area of the target goods, such as factors like the field of view range of the device, the relative angle with the goods, and the distance, a second remote loading instruction for controlling the target image acquisition device is generated. The second remote loading instruction is used to verify whether there is an abnormal storage position problem in the key identification area when the goods to be identified do not meet the spatial constraint conditions. Here, the key identification area refers to the position identification area. For example, the second remote loading instruction requires the target image acquisition device to adjust the shooting angle and focal length so that the key identification area enters the center of its shooting field of view and achieves the best focus, clearly capturing the image of this area for re - detection to determine whether the actual storage position of the goods is compliant. If it is found that the storage position of the goods is incorrect, combined with its quality status, a further decision can be made on whether to move the goods to the correct position or perform other processing.

[0071] In the embodiment of the present application, when the target goods do not meet the spatial constraint conditions, that is, the storage position is incorrect, the main task is to quickly determine its actual storage position. The nearest image acquisition device can most directly and quickly capture the image of the target goods, thereby quickly judging the specific position where the goods are located. The quality constraint conditions involve a comprehensive inspection of the quality of the goods, and it may be necessary to observe the goods from multiple angles and orientations to discover potential quality problems. The designated image acquisition devices within the surrounding range can capture the key identification area of the goods from different angles, providing more comprehensive image information.

[0072] In one embodiment, the above - mentioned remote loading instruction is automatically generated based on the image detection result. To ensure that the goods meet the storage constraint conditions, the status of the goods needs to be continuously monitored. During the preset observation period, the number of times the target goods are identified as not meeting the storage constraint conditions is called the occurrence frequency. For example, in a warehouse, the system will periodically check the goods within half an hour. If a certain piece of goods is identified as not meeting the storage constraint conditions 5 times within half an hour, then its occurrence frequency is 5 times. Based on this occurrence frequency, the instruction type of the target goods can be judged. The instruction type is divided into a single instruction and a combined instruction. If the occurrence frequency is low, it indicates that the recognition result of the goods is random, probably due to image quality problems such as image resolution, resulting in the identified anomalies. At this time, the anomaly problem is not serious, and it can be automatically identified and processed through a single instruction, that is, the remote loading instruction. If the occurrence frequency is high, it indicates that there are relatively serious anomalies in the goods. At this time, to further improve the goods recognition efficiency, the image acquisition device can be jointly loaded and controlled through the auxiliary loading instruction sent by the user.

[0073] When the instruction type of the target goods is a combined instruction, identify the auxiliary loading instruction sent by the user. The auxiliary loading instruction is a voice instruction issued by a person, which can provide additional information for the operation of the system. For example, the user may tell the system to conduct a more detailed inspection of specific goods through a voice instruction. The voice instruction is converted into text or an instruction signal through voice recognition technology, and the system can understand the user's intention. Conduct semantic analysis on the auxiliary loading instruction, aiming to extract the key instruction information therein. For example, the user's voice instruction may be "Conduct high-precision detection on the packaging box of the target goods on Shelf A". By conducting semantic analysis on this sentence, key instruction information such as "Shelf A", "packaging box", and "high-precision detection" can be extracted. Next, it is necessary to determine whether this key instruction information is valid information. Valid information refers to information that can have an actual impact on the detection or operation of the goods. For example, "packaging box" is valid information, which tells which areas of the target goods to conduct high-precision identification on next, while information such as "Check casually" is invalid because it does not provide specific guidance.

[0074] If the key instruction information is valid, correct the auxiliary loading instruction according to the matching degree between the key instruction information and the detection requirements of the key identification area. The higher the matching degree, the more the instruction information conforms to the detection requirements. For example, if the key identification area needs to detect the quality of the goods, and the user's instruction information is about size detection, then the matching degree is relatively low, and the system may need to correct the instruction to make it more in line with the quality detection requirements. The corrected auxiliary loading instruction will be jointly loaded with the remote loading instruction. Joint loading means that the two instructions will be executed simultaneously to jointly complete the identification and detection of the key identification area of the target goods, which can handle the problems of the goods more comprehensively and effectively, improve the efficiency and quality of warehouse management, better adapt to different goods and different detection requirements, and provide more accurate services.

[0075] S103: Based on the remote loading instruction, control the image acquisition device to identify the key identification area, so as to generate goods evaluation information for the target goods according to the obtained identification result, and conduct early warning prompts according to the goods evaluation information.

[0076] Through the remote loading instruction, the remote control system can control the image acquisition device to identify the key identification area, and generate goods evaluation information for the target goods according to the identification result. The remote control system will conduct early warning prompts according to the generated goods evaluation information. If the goods evaluation information indicates that there are quality problems or incorrect storage locations of the goods, the system will issue an early warning in a timely manner to remind the staff to take corresponding measures. For example, if it is found that the storage location of the goods does not meet the regulations, an early warning will be issued to prompt the staff to move the goods to the correct location.

[0077] In the embodiments of the present application, when only identifying goods through images, it may be interfered by factors such as the surrounding environment and other goods, resulting in inaccurate identification. Based on the preliminary image recognition result, by controlling the image acquisition device to perform a secondary detection of the key recognition area on the goods within a certain range, it is possible to more accurately locate the parts that need attention and improve the accuracy of remote goods recognition.

[0078] The above is the method embodiment proposed in the present application. Based on the same idea, some embodiments of the present application also provide the corresponding device and non-volatile computer storage medium for the above method.

[0079] Figure 2 It is a schematic structural diagram of a remote warehouse goods recognition device based on artificial intelligence provided by an embodiment of the present application. As Figure 2 shown, it includes:

[0080] At least one processor; and,

[0081] A memory communicatively connected to at least one processor; wherein,

[0082] The memory stores instructions executable by at least one processor. The instructions are executed by at least one processor so that at least one processor can:

[0083] Collect the goods images of the goods to be recognized in the warehouse, and detect the goods images to determine whether the goods to be recognized meet the storage constraint conditions corresponding to their respective goods types;

[0084] If not, screen out the target goods that do not meet the storage constraint conditions, and generate a remote loading instruction for the image acquisition device according to the distribution relationship between the key recognition area in the target goods and the image acquisition devices set in the warehouse;

[0085] Based on the remote loading instruction, control the image acquisition device to recognize the key recognition area, so as to generate goods evaluation information for the target goods according to the obtained recognition result, and perform a warning prompt according to the goods evaluation information.

[0086] The processor can also:

[0087] In the case where the target goods do not meet the quality constraint conditions, screen out the designated image acquisition devices within the coverage area of the target goods from the image acquisition devices in the warehouse according to the abnormal risk information corresponding to each key recognition area in the target goods;

[0088] Generate a first remote loading instruction for the image acquisition device according to the distribution relationship of the spatial positions between the image acquisition device and the target goods, and the concentration of the target goods around the designated image acquisition device.

[0089] The processor is also capable of:

[0090] Obtain the occurrence frequency of the target goods within a preset observation period, and determine the instruction type corresponding to the target goods according to the occurrence frequency; wherein, the instruction types include single instructions and combined instructions;

[0091] When the instruction type of the target goods is a combined instruction, identify the auxiliary loading instruction sent by the user, and correct the auxiliary loading instruction according to the detection requirements of the key recognition area, so as to jointly load the corrected auxiliary loading instruction and the remote loading instruction when identifying the key recognition area.

[0092] The embodiment of the present application provides a non-volatile computer storage medium storing computer-executable instructions, and the computer-executable instructions are set as: a method for identifying goods in a remote warehouse based on artificial intelligence as described in any one of the above.

[0093] Each embodiment in the present application is described in a progressive manner. The same or similar parts among the embodiments can be referred to each other, and the differences between each embodiment and other embodiments are emphasized. In particular, for the device and medium embodiments, since they are basically similar to the method embodiments, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiments.

[0094] The device and medium provided by the embodiment of the present application correspond one by one to the method. Therefore, the device and medium also have beneficial technical effects similar to those of the corresponding method. Since the beneficial technical effects of the method have been described in detail above, the beneficial technical effects of the device and medium will not be elaborated here.

[0095] Those skilled in the art should understand that the embodiments of the present application can be provided as a method, a system, or a computer program product. Therefore, the present application can adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present application can adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0096] This application is described with reference to the flowcharts and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the application. It should be understood that each flow and / or block in the flowcharts and / or block diagrams, and combinations of flows and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing device to produce a machine, such that the instructions executed by the processor of the computer or other programmable data processing device generate means for implementing the functions specified in the Figure 1 one or more flows and / or blocks Figure 1 one or more blocks

[0097] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a particular manner, such that the instructions stored in the computer-readable memory produce a manufacture including instruction means that implement the functions specified in the Figure 1 one or more flows and / or blocks Figure 1 one or more blocks

[0098] These computer program instructions can also be loaded onto a computer or other programmable data processing device, such that a series of operation steps are executed on the computer or other programmable device to produce a computer-implemented process, and thus the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in the Figure 1 one or more flows and / or blocks Figure 1 one or more blocks

[0099] In a typical configuration, a computing device includes one or more processors (CPUs), an input / output interface, a network interface, and memory.

[0100] The memory may include non-permanent memory in the form of computer-readable media, random access memory (RAM), and / or non-volatile memory such as read-only memory (ROM) or flash memory (flash RAM). The memory is an example of computer-readable media.

[0101] A computer-readable medium includes permanent and non-permanent, removable and non-removable media that can implement information storage by any method or technology. The information can be computer-readable instructions, data structures, program modules, or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, compact disc read-only memory (CD-ROM), digital versatile disc (DVD) or other optical storage, magnetic cassette tapes, magnetic tape magnetic disk storage or other magnetic storage devices, or any other non-transitory medium that can be used to store information that can be accessed by a computing device. As defined herein, a computer-readable medium does not include transitory computer-readable media, such as modulated data signals and carrier waves.

[0102] It should also be noted that the term "comprising", "including" or any other variant thereof is intended to cover non-exclusive inclusion, such that a process, method, article or device comprising a series of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article or device. Without further limitation, an element defined by the statement "comprising an..." does not exclude the presence of additional identical elements in the process, method, article or device comprising the element.

[0103] The above description is only for the embodiments of the present application and is not intended to limit the present application. For those skilled in the art, the present application may have various changes and modifications. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included within the scope of the claims of the present application.

Claims

1. A method for identifying goods in a remote warehouse based on artificial intelligence, characterized in that, The method includes: Collecting a goods image of the goods to be recognized in the warehouse, and detecting the goods image to determine whether the goods to be recognized meet the warehousing constraint conditions corresponding to the goods type to which it belongs; If not, screening out the target goods that do not meet the warehousing constraint conditions, and generating a remote loading instruction for the image acquisition device according to the distribution relationship between the key recognition area in the target goods and the image acquisition devices set in the warehouse; Based on the remote loading instruction, controlling the image acquisition device to recognize the key recognition area, so as to generate goods evaluation information for the target goods according to the obtained recognition result, and giving an early warning prompt according to the goods evaluation information.

2. The method for remotely identifying goods in a warehouse based on artificial intelligence according to claim 1, wherein, The warehousing constraint conditions include quality constraint conditions and space constraint conditions. Generating a remote loading instruction for the image acquisition device according to the distribution relationship between the key recognition area in the target goods and the image acquisition devices set in the warehouse specifically includes: In the case that the target goods do not meet the quality constraint conditions, screening out the designated image acquisition devices within the coverage area of the target goods from the image acquisition devices in the warehouse according to the abnormal risk information corresponding to each key recognition area in the target goods; Generating a first remote loading instruction for the image acquisition device according to the distribution relationship of the spatial positions between the designated image acquisition device and the target goods, and the concentration degree of the target goods around the designated image acquisition device.

3. The method for remotely identifying goods in a warehouse based on artificial intelligence according to claim 1, wherein, After generating the remote loading instruction for the image acquisition device, the method further includes: Obtaining the occurrence frequency of the target goods within a preset observation period, and determining the instruction type corresponding to the target goods according to the occurrence frequency; wherein, the instruction type includes a single instruction and a combined instruction; In the case that the instruction type of the target goods is the combined instruction, identifying the auxiliary loading instruction sent by the user, and correcting the auxiliary loading instruction according to the detection requirements required by the key recognition area, so as to jointly load the corrected auxiliary loading instruction and the remote loading instruction when recognizing the key recognition area.

4. The method for remotely identifying goods in a warehouse based on artificial intelligence according to claim 1, characterized in that, Detecting the goods image to determine whether the goods to be recognized meet the warehousing constraint conditions corresponding to the goods type to which it belongs specifically includes: Detecting the goods image to determine the goods type corresponding to the goods to be recognized, and identifying the position identification area and the quality detection dimension corresponding to the goods to be recognized; Obtaining the warehousing information corresponding to the goods type, and comparing the position identification information carried in the position identification area with the warehousing information to determine whether the position identification information matches the warehousing information; And, based on the quality detection dimension, matching the goods image with the standard image of the goods to be recognized to identify whether there are quality defects in the goods image; In the case that the position identification information does not match the warehousing information and / or there are quality defects in the goods image, it is determined that the goods to be recognized do not meet the warehousing constraint conditions corresponding to the goods type to which it belongs.

5. The method for identifying goods in a remote warehouse based on artificial intelligence according to claim 4, wherein, After matching the goods image with the standard image of the goods to be recognized based on the quality inspection dimension, the method further includes: According to the type of the goods, determine whether the goods to be recognized need to be subjected to linkage detection. If so, determine the linkage device corresponding to the goods to be recognized; Obtain the sensing data collected by the linkage device, so as to identify whether there are quality defects in the quality inspection area through the sensing data and the standard image.

6. The method for remotely identifying goods in a warehouse based on artificial intelligence according to claim 2, wherein, The remote loading instruction further includes a second remote loading instruction. According to the distribution relationship between the key recognition area in the target goods and the image acquisition device set in the warehouse, a remote loading instruction for the image acquisition device is generated, specifically including: When the target goods do not meet the space constraint condition, determine the target image acquisition device with the closest spatial distance to the target goods, and generate a second remote loading instruction for the target image acquisition device according to the distribution relationship between the target image acquisition device and the key recognition area.

7. An artificial intelligence-based remote warehouse goods recognition device, characterized in that, The device includes: At least one processor; And a memory communicatively connected to the at least one processor; Wherein, the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor, so that the at least one processor can: Collect the goods image of the goods to be recognized in the warehouse, and detect the goods image to determine whether the goods to be recognized meet the warehousing constraint conditions corresponding to the type of goods to which it belongs; If not, screen out the target goods that do not meet the warehousing constraint conditions, and generate a remote loading instruction for the image acquisition device according to the distribution relationship between the key recognition area in the target goods and the image acquisition device set in the warehouse; Based on the remote loading instruction, control the image acquisition device to recognize the key recognition area, so as to generate goods evaluation information for the target goods according to the obtained recognition result, and perform a warning prompt according to the goods evaluation information.

8. An intelligent remote warehouse goods recognition device based on artificial intelligence according to claim 7, characterized in that, The warehousing constraint conditions include quality constraint conditions and space constraint conditions, and the processor can also: When the target goods do not meet the quality constraint conditions, according to the abnormal risk information corresponding to each key recognition area in the target goods, screen out the designated image acquisition device within the coverage area of the area where the target goods are located from the image acquisition devices in the warehouse; Generate a first remote loading instruction for the image acquisition device according to the distribution relationship between the spatial positions of the image acquisition device and the target goods, and the concentration of the target goods around the designated image acquisition device.

9. The remote warehouse goods recognition device based on artificial intelligence according to claim 7, characterized in that, The processor can also: Obtain the occurrence frequency of the target goods within a preset observation period, and determine the instruction type corresponding to the target goods according to the occurrence frequency; wherein, the instruction type includes a single instruction and a joint instruction; In the case where the instruction type of the target goods is the combined instruction, identify the auxiliary loading instruction sent by the user, and correct the auxiliary loading instruction according to the detection requirements of the key identification area, so as to jointly load the corrected auxiliary loading instruction and the remote loading instruction when identifying the key identification area.

10. A non-volatile computer storage medium stores computer-executable instructions, characterized in that, The computer-executable instructions are set as follows: A method for identifying goods in a remote warehouse based on artificial intelligence according to any one of claims 1-6.

Images