Lost article retrieving method

Abstract

The invention provides a lost article retrieving method. The method comprises the following steps: acquiring first identification information sent by a lost article party corresponding to a lost article; when second identification information matched with the first identification information exists, prompt information representing successful matching is sent to the lost object side, and the second identification information is sent by an object picking side. According to the invention, the lost article retrieving efficiency is improved.

Description

Technical Field

[0001] The present invention relates to the field of big data, and in particular to a method for recovering lost property. Background Art

[0002] Lost items are common in schools, businesses, and institutions. A common practice is for the person who lost the item to register it, and then for the administrator to post the information on a display terminal. The hope is that the finder will see the information and return it. However, this method relies on the timeliness of information registration. If the finder doesn't see the information, the efficiency of recovering the lost item is low. Summary of the Invention

[0003] The object of the present invention is to provide a lost property recovery method for solving the problem of low efficiency of physical property recovery caused by the prior art.

[0004] In order to achieve the above object, the technical solution adopted by the present invention is as follows:

[0005] A method for recovering lost property, comprising:

[0006] Obtaining first identification information issued by the lost party corresponding to the lost property;

[0007] When there is second identification information that matches the first identification information, a prompt message indicating that the match is successful is sent to the party who lost the item, and the second identification information is sent by the party who found the item.

[0008] Furthermore, after obtaining the first identification information issued by the lost party corresponding to the lost property, the method further includes:

[0009] When there is no second identification information matching the first identification information, acquiring a captured image, where the captured image is obtained based on a visual terminal at a specified position;

[0010] Determining a matching result between the captured image and the first identification information based on a matching strategy;

[0011] Based on the matching result, a prompt message indicating a successful matching is sent to the party who lost the item, or the first identification information is sent to a display terminal.

[0012] Furthermore, the first identification information includes a picture of the lost property;

[0013] The determining, based on the matching strategy, a matching result between the captured image and the first identification information includes:

[0014] Inputting both the photographed picture and the lost item picture into an image recognition model, the image recognition model outputting a first item sub-image of the lost item picture and a second item sub-image of the photographed picture;

[0015] Determine the similarity between the first item sub-image and the second item sub-image. When the similarity is greater than or equal to a similarity threshold, the matching result is that the photographed image matches the lost item image, and execute the sending of a prompt message indicating a successful match to the lost item party; when the similarity is less than the similarity threshold, the matching result is that the photographed image does not match the lost item image, and execute the sending of the first identification information to the display terminal.

[0016] Furthermore, the first identification information includes text information, and determining a matching result between the captured image and the first identification information based on a matching strategy includes:

[0017] When the similarity is less than a similarity threshold, extracting a first character and a second character of the text information based on a large language model, where the first character is used to represent a first type of the lost object, and the second character is used to represent a first color of the lost object;

[0018] sending the captured image to an image recognition system, wherein the image recognition system outputs a second type and a second color of the object in the captured image;

[0019] When the first type and the second type are the same and the first color and the second color are the same, the matching result is that the photographed picture matches the lost property picture; when the first type and the second type are different or the first color and the second color are different, the matching result is that the photographed picture does not match the lost property picture.

[0020] Furthermore, the image recognition system includes an edge detection model, a classification model, and a color recognition model;

[0021] The sending the captured picture to an image recognition system, wherein the image recognition system outputs the second type and the second color of the object in the captured picture, comprises:

[0022] Inputting the captured image into the edge detection model, the edge detection model extracting the outline of the object in the captured image, and extracting the second object sub-image based on the outline;

[0023] inputting the second item subgraph into a classification model, wherein the classification model inputs the second type of the second item subgraph;

[0024] The second item sub-image is input into a color recognition model. The color recognition model extracts the RGB value of each pixel in the second item sub-image, groups pixels with the same RGB value into a cluster, and determines the second color based on the cluster having a number of pixels greater than a preset number.

[0025] Furthermore, the second identification information includes a location where the object was found;

[0026] After sending a prompt message indicating successful matching to the party who lost the item, the method further includes:

[0027] Obtaining and storing the object picking location, and counting the number of times the object picking location storing is performed;

[0028] When the number of times reaches a preset number, the picking-up location is used as the designated location.

[0029] Furthermore, after sending a prompt message indicating successful matching to the party who lost the item, the method further includes:

[0030] The prompt information is sent to a third-party administrator.

[0031] Furthermore, the acquiring of the captured image includes:

[0032] Get several arrival locations;

[0033] Based on all the arrival locations and the designated area map, a reachable path map is obtained, wherein the reachable path map includes a plurality of paths such that the reachable path map passes through all the arrival locations;

[0034] Based on all the paths, an inflection point is obtained, where the inflection point represents an intersection point of any two of the paths;

[0035] Obtaining the cluster center points of all the inflection points;

[0036] Obtaining a maximum radius according to the cluster center point and all the inflection points, wherein the maximum radius represents the maximum value of the straight-line distance between the cluster center point and the inflection point;

[0037] Based on the cluster center point and the maximum radius, a possible reachable range is obtained, where the possible reachable range represents a corresponding area on the designated area map of a circle formed by the cluster center point and the maximum radius;

[0038] All visual terminals within the possible reachable range are turned on to acquire the captured images.

[0039] Beneficial effects of the present invention:

[0040] The present invention matches the identification information provided by the party who lost the property (such as the owner) with the identification information provided by the party who found the property (such as the finder), thereby quickly associating the lost property with the finder. When the finder fails to obtain the first identification information, a prompt message is immediately sent to the party who lost the property after the match is successful, thereby reducing manual query and waiting time and shortening the lost property claiming cycle.

[0041] When identification information fails to match, the present invention automatically triggers a visual terminal at a designated location to capture an image, expanding static identification matching into a dynamic scene search. By automating the capture and matching process, the workload of manual inspection and screening is reduced, the system's autonomous object-finding capabilities are enhanced, and object-finding efficiency is improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Attachment Figure 1 This is a flow chart of S100-S200 of the present invention;

[0043] Attachment Figure 2 This is the flow chart of S100, S300-S500 of the present invention;

[0044] Attachment Figure 3 It is the process sub-graph (S410-S420) of S400;

[0045] Attachment Figure 4 It is the process sub-graph of S400 (S430-S450);

[0046] Attachment Figure 5 Flowcharts of S100-S200, S600-S700;

[0047] Attachment Figure 6 A schematic diagram of mapping reachable locations to a map of a designated area;

[0048] Attachment Figure 7 To generate a path diagram;

[0049] Attachment Figure 8 Schematic diagram of cluster center J obtained by clustering;

[0050] Attachment Figure 9 To obtain a straight line diagram of all cluster centers and inflection points;

[0051] Attachment Figure 10 A schematic diagram of the possible reach range is provided. DETAILED DESCRIPTION

[0052] The following describes the embodiments of the present invention with reference to the accompanying drawings and preferred embodiments. Those skilled in the art will readily appreciate the other advantages and benefits of the present invention from the disclosure herein. The present invention may also be implemented or applied through various other specific embodiments, and the various details in this specification may be modified or altered based on different viewpoints and applications without departing from the spirit of the present invention. It should be understood that the preferred embodiments are intended only to illustrate the present invention and are not intended to limit the scope of protection of the present invention.

[0053] It should be noted that the illustrations provided in the following embodiments are merely schematic illustrations of the basic concept of the present invention. Therefore, the illustrations only show components related to the present invention and are not drawn according to the number, shape, and size of components in actual implementation. In actual implementation, the type, quantity, and proportion of each component may be changed arbitrarily, and the component layout may also be more complex.

[0054] This embodiment proposes a lost property recovery method, which can be applied to the server side. Specifically, a lost property recovery platform can be deployed on the server. The platform can be presented in the form of a web page on the PC side or in the form of an APP on the mobile side, so that the lost property recovery method can be run on the platform to improve the efficiency of lost property recovery.

[0055] Therefore, in at least one embodiment, if Figure 1 As shown, the lost property recovery method may include the following steps:

[0056] S100: Obtaining first identification information corresponding to the lost property and issued by the lost property party;

[0057] S200: When there is second identification information that matches the first identification information, a prompt message indicating that the match is successful is sent to the party that lost the item, and the second identification information is sent by the party that found the item.

[0058] The lost and found parties correspond to the person who lost the lost property and the person who found it respectively. The lost person publishes the first identification information about the lost property on the platform, and the finder publishes the second identification information of the found property on the platform. When the first identification information matches the second identification information, the platform issues a prompt message.

[0059] In S100-S200, the finder does not need to actually observe the information sent by the loser. He only needs to register the found item on the platform. The platform will perform matching work. When the first identification information and the second identification information match, the lost item can be recovered.

[0060] In this embodiment, the first identification information may include first text information and / or first picture information, and the second identification information may include second text information and / or second picture information.

[0061] The first image information and the second image information, representing images of the lost and found items, respectively, can be captured by the parties who lost and found the items using a camera. A match between the first identification information and the second identification information can indicate that the items corresponding to the first and second image information are highly similar, i.e., a similarity threshold is reached. Specifically, an edge detection algorithm can be used to extract the outlines of the items in the first and second image information. The brightness of pixels within the outline is enhanced, while the brightness of pixels outside the outline is reduced, thereby extracting the items in the two images. The similarity between the two item images is then tested to determine whether the similarity exceeds the similarity threshold.

[0062] The method for determining similarity may include the following steps:

[0063] 1. Image preprocessing

[0064] Unify the size: Scale both images to the same resolution (e.g., 256×256 pixels) to eliminate the effect of size differences.

[0065] Grayscale conversion: Convert color images into grayscale images to avoid color information interfering with similarity judgment.

[0066] Contrast enhancement: Improves image uneven lighting through histogram equalization technology.

[0067] 2. Similarity calculation

[0068] Mean Square Error (MSE): Calculates the average square difference of the corresponding pixel values ​​of two images. The smaller the value, the more similar they are.

[0069] Structural Similarity (SSIM): takes into account brightness, contrast, and structural similarity and outputs a standardized score between 0 and 1.

[0070] It can also be:

[0071] 1. Feature extraction:

[0072] Load a pre-trained convolutional neural network (such as ResNet50, VGG16) and remove the classification layer.

[0073] After the input image is preprocessed (cropped and normalized), the feature vector before the fully connected layer is extracted.

[0074] 2. Similarity calculation

[0075] Compare the feature vectors of two images using cosine similarity or Euclidean distance.

[0076] Optional: Accelerate large-scale vector retrieval through dimensionality reduction (such as PCA) or hash encoding.

[0077] When the finder or either party does not have image information, it can be determined whether the first identification information and the second identification information match by matching text information with image information.

[0078] Assume that the first identification information released by the lost party only contains text information, for example:

[0079] Type of lost property: bag,

[0080] Color: Yellow,

[0081] Length: 30cm-40cm,

[0082] Height: 10cm-30cm.

[0083] Therefore, this embodiment can identify the type of the found object in the picture provided by the finder based on the classification model, then extract the RGB value of the found object based on the image detection algorithm, and finally obtain the size of the found object. When all correspond to the text, it can be considered that the first identification information and the second identification information match successfully, otherwise the match is unsuccessful.

[0084] In this embodiment, the classification model can be trained based on a deep learning model, wherein the training images can be obtained by surveying items that are frequently lost in the area, and then pictures of various easily lost items are collected on the Internet, and the type of each picture is labeled, thereby realizing the training of the classification model.

[0085] In at least one embodiment, Figure 2 As shown, the method may further include the following steps:

[0086] S100: Obtaining first identification information corresponding to the lost property and issued by the lost property party;

[0087] S300: When there is no second identification information matching the first identification information, acquiring a captured image, where the captured image is obtained based on a visual terminal at a specified position;

[0088] S400: Determine a matching result between the captured image and the first identification information based on a matching strategy;

[0089] S500: Based on the matching result, a prompt message indicating a successful matching is sent to the party who lost the item, or the first identification information is sent to a display terminal.

[0090] In S300, it indicates that no finder has found the lost property, that is, the lost property may not have been picked up by other people, so a captured image can be obtained, and the captured image is obtained based on the visual terminal at the designated location. The designated location can be a place where items are often lost, such as a cafeteria, a stadium, etc. The places where items are often lost can be obtained based on street questionnaires. In order to improve the efficiency of recovering lost items, this embodiment can install visual terminals, such as cameras, at designated locations. Through these visual terminals, it can be determined whether there are lost items in easily lost locations, that is, by actively turning on the visual terminals, the server actively searches for lost items. In this embodiment, the first identification information includes a picture of the lost item and text information.

[0091] In S400, if Figure 3 As shown, the following steps may be specifically included:

[0092] S410: Inputting the photographed picture and the lost property picture into an image recognition model, the image recognition model outputting a first item sub-image of the lost property picture and a second item sub-image of the photographed picture;

[0093] S420: Determine the similarity between the first item sub-image and the second item sub-image. When the similarity is greater than or equal to a similarity threshold, the matching result is that the photographed image matches the lost item image. When the similarity is less than the similarity threshold, the matching result is that the photographed image does not match the lost item image.

[0094] In this embodiment, S400 can be specifically implemented by the following method:

[0095] Step S410: The image recognition model generates an item sub-graph

[0096] Core Goals

[0097] From the captured images and lost items images, local sub-images containing the target items (such as wallets, keys, etc.) are extracted respectively, eliminating background interference.

[0098] Implementation Logic

[0099] Model selection

[0100] Use pre-trained object detection models (such as YOLO, Faster R-CNN) or semantic segmentation models (such as Mask R-CNN, DeepLab).

[0101] The object detection model outputs the bounding box of the object, and the semantic segmentation model outputs the pixel-level mask.

[0102] Input preprocessing

[0103] Scale the image to the input size required by the model (such as 416×416 or 640×640).

[0104] Normalize pixel values ​​to the range of [0, 1] to match the data distribution during model training.

[0105] Item location and retrieval

[0106] Target detection process:

[0107] The model predicts the categories and confidence levels of all possible objects in the image.

[0108] Filter out the item with the highest confidence (assuming the lost item image contains only a single target).

[0109] According to the bounding box coordinates, the object sub-image is cropped from the original image.

[0110] Semantic segmentation process:

[0111] The model generates pixel-level classification results, distinguishing between foreground (objects) and background.

[0112] Refine the mask edges through morphological operations (such as erosion and dilation).

[0113] Extract the object sub-image from the original image based on the mask area.

[0114] Sub-image post-processing:

[0115] Unify the sub-graph size (such as 224×224) to facilitate subsequent similarity calculation.

[0116] Standardize the subgraphs (e.g., subtract the mean, divide by the standard deviation) to improve model generalization.

[0117] Output:

[0118] The first item sub-image corresponding to the lost item image (such as the front image of the lost wallet).

[0119] The second item sub-image corresponding to the captured image (such as the wallet area detected in the captured scene).

[0120] Step S420: Similarity calculation and matching decision core goal: quantify the similarity between two item subgraphs and determine whether they are the same item by using a threshold. Implementation logic:

[0121] Feature extraction

[0122] Use pre-trained feature extraction models (such as ResNet, VGG) to process subgraphs.

[0123] Remove the classification layer at the top of the model and extract the feature vector (e.g., a 2048-dimensional vector) before the fully connected layer. The feature vector encodes high-level semantic information such as the texture, shape, and color of the object.

[0124] Similarity calculation

[0125] Cosine similarity: Calculates the cosine of the angle between two vectors, ranging from -1 to 1. The larger the value, the more similar they are.

[0126] Euclidean distance: Calculates the straight-line distance in vector space. The smaller the value, the more similar it is.

[0127] Structural Similarity (SSIM): Directly compares the brightness, contrast, and structural similarity of sub-images (suitable for pixel-level comparison).

[0128] Threshold decision

[0129] Static threshold: A fixed threshold is set based on scenario experience (e.g., a cosine similarity ≥ 0.85 is considered a match).

[0130] Dynamic threshold: Combined with historical data distribution, the threshold is automatically adjusted through statistical methods (such as confidence interval).

[0131] Multi-threshold strategy: Set a high threshold (exact match) and a low threshold (candidate match) to process the results in a hierarchical manner.

[0132] Matching result output

[0133] Similarity ≥ threshold → the captured image matches the lost item image.

[0134] Similarity < threshold → the captured image does not match the lost item image.

[0135] In at least one embodiment, Figure 4 As shown, the method can also be:

[0136] S430: When the similarity is less than a similarity threshold, extracting a first character and a second character of the text information based on the large language model, where the first character is used to represent the first type of the lost object, and the second character is used to represent the first color of the lost object;

[0137] S440: Sending the captured image to an image recognition system, wherein the image recognition system outputs a second type and a second color of the object in the captured image;

[0138] S450: When the first type and the second type are the same and the first color and the second color are the same, the matching result is that the photographed picture matches the lost property picture; when the first type and the second type are different or the first color and the second color are different, the matching result is that the photographed picture does not match the lost property picture.

[0139] Wherein: the image recognition system includes an edge detection model, a classification model, and a color recognition model;

[0140] The sending the captured picture to an image recognition system, wherein the image recognition system outputs the second type and the second color of the object in the captured picture, comprises:

[0141] Inputting the captured image into the edge detection model, the edge detection model extracting the outline of the object in the captured image, and extracting the second object sub-image based on the outline;

[0142] inputting the second item subgraph into a classification model, wherein the classification model inputs the second type of the second item subgraph;

[0143] The second item sub-image is input into a color recognition model. The color recognition model extracts the RGB value of each pixel in the second item sub-image, groups pixels with the same RGB value into a cluster, and determines the second color based on the cluster having a number of pixels greater than a preset number.

[0144] Through S430-S450, it is determined whether the lost property is in the designated location by detecting whether the text information of the first identification information matches the photographed image.

[0145] S500 indicates that when the first identification information matches the photographed image, a prompt message indicating a successful match is sent to the party who lost the item. If there is no match, the lost item does not exist at the designated location, and no finder has picked up the lost item. Therefore, the first identification information is sent to the display terminal, and people can obtain relevant information about the lost item through the display terminal, thereby improving the probability and efficiency of recovering the lost item.

[0146] In at least one embodiment, Figure 5 As shown, the method may further include the following steps:

[0147] S100: Obtaining first identification information corresponding to the lost property and issued by the lost property party;

[0148] S200: When there is second identification information that matches the first identification information, a prompt message indicating a successful match is sent to the party who lost the item;

[0149] S600: Acquire and store the object picking location, and count the number of times the object picking location storing is performed;

[0150] S700: When the number of times reaches a preset number, the picking-up location is used as the designated location.

[0151] In this embodiment, when the number of the found locations is large enough, it indicates that the found locations are locations where items are easily lost. Therefore, the easily designated locations can be dynamically updated in this way, thereby further improving the efficiency of lost item search.

[0152] In this embodiment, after issuing the prompt information indicating successful matching, the successful matching information needs to be synchronized to the administrator.

[0153] In at least one embodiment, S300 can determine the visual terminal that needs to be turned on in the following manner.

[0154] First, obtain several arrival location points. This step can be achieved by obtaining the arrival location points filled in by the person who lost the item on the server. For example, this embodiment is based on a school scenario. On the webpage, a prompt is given to the person who lost the item: Please provide the locations you have reached in the school, and provide options such as the cafeteria, library, first teaching building, and administrative building. Then, based on the name of the arrival location point selected by the person who lost the item, the name is mapped to the coordinate location point on the school map, thereby obtaining the coordinates of the arrival location point on the school map. The school map described in this step can be the designated area map of the following step. The school map includes the location coordinates of all locations and all road information.

[0155] In this embodiment, Figure 6 As shown, the locations A, B, C, and D selected by the lost party are mapped to the school map. The school map is Figure 6 The above is represented by ZONE, and A, B, C, and D respectively represent the coordinates on the school map of the four arrival locations selected by the lost party.

[0156] Second, based on all the arrival location points and the designated area map, a reachable path map is obtained, where the reachable path map includes a plurality of paths, so that the reachable path map passes through all the arrival location points.

[0157] like Figure 7 As shown, according to the path planning algorithm, based on the coordinates of the four location points A, B, C, and D, and the road information on the school map, a reachable path map can be obtained. The reachable path map represents the path of the lost party through A, B, C, and D.

[0158] B, C, D strategies, wherein the reachable path graph in the embodiment is represented by Figure 7 The graph is composed of AE, EI, IB, BF, FD, DG, GC, CH and HA, where AE, EI, IB, BF, FD, DG, GC, CH and HA all represent paths in the reachable path graph.

[0159] Third, based on all the paths, an inflection point is obtained, where the inflection point represents the intersection of any two paths. AI represents an inflection point, which is the intersection of any two of AE, EI, IB, BF, FD, DG, GC, CH, and HA.

[0160] Fourth, obtain the cluster center point of the position set, where the points of the position set include all the arrival position points and all the inflection points. Figure 8 As shown in Figure 1, each point AI represents the coordinates of a location point on the school map. According to the K-means clustering algorithm, the cluster center point J of the point AI can be obtained.

[0161] Fifth, according to the cluster center point and the inflection point, a maximum radius is obtained, wherein the maximum radius represents the maximum value of the straight-line distance between the cluster center point and the inflection point. Figure 9 As shown, a straight line from point J to point AI can be obtained, and then the length of each straight line is calculated, and the largest length is selected as the maximum radius. In this embodiment, it is assumed that JA is selected as the maximum radius.

[0162] Sixth, based on the cluster center point and the maximum radius, a possible reach range is obtained, where the possible reach range represents the corresponding area on the designated area map of the circle formed by the cluster center point and the maximum radius. Figure 10 As shown, JA is used as the radius and J as the center to form a circle.

[0163] Seventh, all visual terminals within the possible reach are turned on to obtain the captured images. Figure 10 The formed circle corresponds to the school map, and the possible reachable range is obtained, and the visual terminal of the possible reachable range is opened.

[0164] This embodiment forms a reachable range based on the first to seventh methods. The reachable range indicates the possible activity range of the lost party in the school based on the arrival location. Therefore, there is no need to turn on all visual terminals, only the visual terminals within the activity range, which reduces resource consumption.

[0165] The above embodiments are only preferred embodiments for fully illustrating the present invention, and the protection scope of the present invention is not limited thereto. Any equivalent substitution or modification made by those skilled in the art based on the present invention is within the protection scope of the present invention.

Claims

1. A method for recovering lost property, characterized by: The method comprises: Obtaining first identification information issued by the lost party corresponding to the lost property; When there is second identification information that matches the first identification information, a prompt message indicating that the match is successful is sent to the party who lost the item, and the second identification information is sent by the party who found the item.

2. The method according to claim 1, wherein: After obtaining the first identification information issued by the lost party corresponding to the lost property, the method further includes: When there is no second identification information matching the first identification information, acquiring a captured image, where the captured image is obtained based on a visual terminal at a specified position; Determining a matching result between the captured image and the first identification information based on a matching strategy; Based on the matching result, a prompt message indicating a successful matching is sent to the party who lost the item, or the first identification information is sent to a display terminal.

3. The method according to claim 2, wherein: The first identification information includes a picture of the lost property; The determining, based on the matching strategy, a matching result between the captured image and the first identification information includes: Inputting both the photographed picture and the lost item picture into an image recognition model, the image recognition model outputting a first item sub-image of the lost item picture and a second item sub-image of the photographed picture; Determine the similarity between the first item sub-image and the second item sub-image. When the similarity is greater than or equal to a similarity threshold, the matching result is that the photographed image matches the lost item image, and execute the sending of a prompt message indicating a successful match to the lost item party; when the similarity is less than the similarity threshold, the matching result is that the photographed image does not match the lost item image, and execute the sending of the first identification information to the display terminal.

4. The method according to claim 3, wherein: The first identification information includes text information, and determining a matching result between the captured image and the first identification information based on a matching strategy includes: When the similarity is less than a similarity threshold, extracting a first character and a second character of the text information based on a large language model, where the first character is used to represent a first type of the lost object, and the second character is used to represent a first color of the lost object; sending the captured image to an image recognition system, wherein the image recognition system outputs a second type and a second color of the object in the captured image; When the first type and the second type are the same and the first color and the second color are the same, the matching result is that the photographed picture matches the lost property picture; when the first type and the second type are different or the first color and the second color are different, the matching result is that the photographed picture does not match the lost property picture.

5. The method according to claim 4, characterized in that : The image recognition system includes an edge detection model, a classification model, and a color recognition model; The sending the captured picture to an image recognition system, wherein the image recognition system outputs the second type and the second color of the object in the captured picture, comprises: Inputting the captured image into the edge detection model, the edge detection model extracting the outline of the object in the captured image, and extracting the second object sub-image based on the outline; inputting the second item subgraph into a classification model, wherein the classification model inputs the second type of the second item subgraph; The second item sub-image is input into a color recognition model. The color recognition model extracts the RGB value of each pixel in the second item sub-image, groups pixels with the same RGB value into a cluster, and determines the second color based on the cluster having a number of pixels greater than a preset number.

6. The method according to claim 2, wherein: The second identification information includes the location of the picked-up object; After sending a prompt message indicating successful matching to the party who lost the item, the method further includes: Obtaining and storing the object picking location, and counting the number of times the object picking location storing is performed; When the number of times reaches a preset number, the picking-up location is used as the designated location.

7. The method according to any one of claims 1 to 5, characterized in that: After sending a prompt message indicating successful matching to the party who lost the item, the method further includes: The prompt information is sent to a third-party administrator.

8. The method according to claim 2, characterized in that , the acquisition of captured images includes, Get several arrival locations; Based on all the arrival locations and the designated area map, a reachable path map is obtained, wherein the reachable path map includes a plurality of paths such that the reachable path map passes through all the arrival locations; Based on all the paths, an inflection point is obtained, where the inflection point represents an intersection point of any two of the paths; Obtaining the cluster center points of all the inflection points; Obtaining a maximum radius according to the cluster center point and all the inflection points, wherein the maximum radius represents the maximum value of the straight-line distance between the cluster center point and the inflection point; Based on the cluster center point and the maximum radius, a possible reachable range is obtained, where the possible reachable range represents a corresponding area on the designated area map of a circle formed by the cluster center point and the maximum radius; All visual terminals within the possible reachable range are turned on to acquire the captured images.

Images