Real-scene map processing method and device, equipment and storage medium

Abstract

The invention provides a live-action map processing method and device, equipment and a storage medium, and relates to the technical field of computers, in particular to the technical field of internet map services. The specific implementation scheme is as follows: acquiring a live-action image of a live-action map; carrying out image identification on the live-action image; based on the image identification result and a recall strategy, identifying geographical sensitive information in the live-action image; wherein the geographical sensitive information comprises units and facilities related to the live-action image to be subjected to technical processing, and attribute information related to the live-action image to be subjected to technical processing; and marking the live-action image according to the geographical sensitive information. According to the invention, sensitive information in the live-action map can be identified.

Description

Technical Field

[0001] The present disclosure relates to the field of computer technology, and in particular to the field of Internet map service technology. Background Art

[0002] A streetscape map is an electronic map that associates real-life images with two-dimensional electronic map elements through geographic reference, that is, a streetscape map. A streetscape map can provide 360-degree real-life images of urban built-up areas, scenic spots, tourist attractions and other areas that are closely related to the daily lives of the public, providing a richer and more intuitive Internet map service for the public's life. Users can get an immersive map browsing experience through this service. With the rapid development of Internet map services, the user population is growing, the amount of information is increasing, and the desensitization of real-life maps is becoming more and more important. Accurately grasping the desensitization review standards and improving the automation and intelligence level of desensitization review are important issues facing the desensitization of real-life data. How to achieve accurate targeting and efficient desensitization of sensitive information in real-life maps, so as to better ensure data security, reliability and availability, is a technical problem that needs to be solved. Summary of the invention

[0003] The present invention provides a method, device, equipment and storage medium for processing a real-life map.

[0004] According to one aspect of the present disclosure, a method for processing a real-life map is provided, comprising:

[0005] Obtain the real-life image of the real-life map;

[0006] Performing image recognition on the real scene image;

[0007] Based on the image recognition results and the recall strategy, the geographically sensitive information in the real-life image is identified; the geographically sensitive information includes the units and facilities involved in the real-life image that should be technically processed, and the attribute information involved in the real-life image that should be technically processed;

[0008] The real scene image is marked according to the geographically sensitive information.

[0009] According to another aspect of the present disclosure, a real-view map processing device is provided, comprising:

[0010] An acquisition module is used to acquire the real scene image of the real scene map;

[0011] A first recognition module is used to perform image recognition on the real scene image;

[0012] The second recognition module is used to identify geographically sensitive information in the real-scene image based on the image recognition result and the recall strategy; wherein the geographically sensitive information includes units and facilities involved in the real-scene image that should be technically processed, and attribute information involved in the real-scene image that should be technically processed;

[0013] The marking module is used to mark the real scene image according to the geographical sensitive information.

[0014] According to another aspect of the present disclosure, there is provided an electronic device, comprising:

[0015] at least one processor; and

[0016] a memory communicatively connected to the at least one processor; wherein,

[0017] The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to execute any method in the embodiments of the present disclosure.

[0018] According to another aspect of the present disclosure, a non-transitory computer-readable storage medium storing computer instructions is provided, wherein the computer instructions are used to enable the computer to execute any method according to the embodiments of the present disclosure.

[0019] According to another aspect of the present disclosure, a computer program product is provided, including a computer program, which implements any method according to the embodiments of the present disclosure when executed by a processor.

[0020] The present disclosure performs image recognition and text recognition on the real-scene images of the real-scene map, identifies sensitive information in the real-scene images according to the recognition results and a preset recall strategy, and can mark the sensitive information in the real-scene map.

[0021] It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present disclosure, nor is it intended to limit the scope of the present disclosure. Other features of the present disclosure will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] The accompanying drawings are used to better understand the present solution and do not constitute a limitation of the present disclosure.

[0023] Figure 1 is a flowchart of a method for processing a real-life map according to an embodiment of the present disclosure;

[0024] Figure 2 is a functional schematic diagram of each processing module according to an embodiment of the present disclosure;

[0025] Figure 3It is a schematic diagram of image features of public safety facilities;

[0026] Figure 4 It is a schematic diagram of the characteristics of images of civil facilities;

[0027] Figure 5 It is a schematic diagram of image features of sensitive attribute information;

[0028] Figure 6 is a schematic diagram of image processing;

[0029] Figure 7 It is the sensitive information image before image processing;

[0030] Figure 8 It is a sensitive information image after image processing;

[0031] Fig. 9 is a structural schematic diagram of a real scene map processing device 900 according to an embodiment of the present disclosure;

[0032] Fig.10 A schematic block diagram of an example electronic device 1000 is shown that can be used to implement embodiments of the present disclosure. DETAILED DESCRIPTION

[0033] The following is a description of exemplary embodiments of the present disclosure in conjunction with the accompanying drawings, including various details of the embodiments of the present disclosure to facilitate understanding, which should be considered as merely exemplary. Therefore, it should be recognized by those of ordinary skill in the art that various changes and modifications may be made to the embodiments described herein without departing from the scope of the present disclosure. Similarly, for the sake of clarity and conciseness, the description of well-known functions and structures is omitted in the following description.

[0034] The “and / or” in the embodiments of the present disclosure indicates that three relationships may exist. For example, A and / or B may indicate the following three situations: A exists alone, A and B exist at the same time, and B exists alone. The term “at least one” herein indicates any combination of at least two of any one or more of a plurality of types. For example, at least one of A, B, and C may indicate any one or more elements selected from the set consisting of A, B, and C. The terms “first” and “second” herein refer to and distinguish between multiple similar technical terms, and do not mean to limit the order or to limit the meaning to only two. For example, the first feature and the second feature refer to two types / two features. The first feature may be one or more, and the second feature may also be one or more.

[0035] A streetscape map is an electronic map that associates real-life images with two-dimensional electronic map elements through geographic reference, namely a streetscape map.

[0036] Streetscape image refers to a near-ground image that is consistent with human visual perception and reflects the real spatial relationship, time, and cultural and social environment information of the geographical scene.

[0037] The disclosed embodiment proposes a real-life map processing method, which is capable of identifying sensitive information in the real-life map, marking the sensitive information in the real-life image of the real-life map, and thereby desensitizing the real-life map according to the marking result.

[0038] Figure 1 is a flowchart of a method for processing a real scene map according to an embodiment of the present disclosure, including:

[0039] S110, obtaining a real scene image of the real scene map;

[0040] S120, performing image recognition on the real scene image;

[0041] S130, based on the image recognition result and the recall strategy, identifying geographically sensitive information in the real scene image; wherein the geographically sensitive information includes units and facilities involved in the real scene image that should be technically processed, and attribute information involved in the real scene image that should be technically processed;

[0042] S140: Mark the real scene image according to the geographic sensitive information.

[0043] Through the above process, sensitive elements in real-life images and the coordinates of sensitive elements can be automatically marked, and sensitive real-life images can also be marked. These include text sensitive elements and / or image sensitive elements. In the subsequent process, the sensitive elements and / or sensitive real-life images determined by the program can be manually verified for a second time, and the sensitive type, sensitive area and processing method of the real-life images can be marked; and based on the processing method, the real-life images can be technically processed to remove sensitive information.

[0044] In some implementations, the units and facilities involved in the real-life images that should be technically processed include:

[0045] a) Various military facilities that directly serve military purposes: command organs, ground and underground command projects, combat projects, military airports, ports, docks, camps, training grounds, test sites, military caverns, warehouses, military communication, reconnaissance, navigation, observation stations and surveying, navigation, and navigation aids, military roads, dedicated railway lines, military communication and power transmission lines, and military oil and water pipelines.

[0046] b) Military restricted areas, military management areas and all units and facilities within them.

[0047] c) Key departments or units involving national security.

[0048] d) Units and facilities related to public security: prisons, criminal detention centers, labor education management centers, drug rehabilitation centers (stations) and custody and education centers, storage areas of weapons and ammunition, explosives, highly toxic substances, dangerous (chemical) products, uranium deposits and centralized storage areas for radioactive materials.

[0049] e) Civilian facilities that involve the national economic lifeline and have a significant impact on people's production and life: large-scale water conservancy facilities, power facilities, communication facilities, oil and gas (natural gas, coal gas, etc.) facilities, important strategic material reserves (grain warehouses, cotton warehouses, etc.), meteorological stations, rainfall radar stations and hydrological observation stations (networks).

[0050] f) Dedicated railways and train lines within stations, railway marshalling yards, and dedicated roads.

[0051] In some implementations, the attribute information involved in the real-scene image to be technically processed includes:

[0052] a) The height limit, width limit, clearance, load capacity and slope attributes of bridges; the height and width attributes of tunnels; and the pavement material attributes of highways.

[0053] b) The navigability, water depth, flow rate, bottom quality and bank quality of rivers; the storage capacity of reservoirs; the height of dams; the nature of water sources; the water depth and mud depth of swamps and their boundary contours; and the properties of ferry crossings.

[0054] c) Publicize the airport's transport capacity attributes.

[0055] Methods for technically processing real-scene images include:

[0056] 1. Delete the entire real-life image:

[0057] If the real-life image contains close-up images of the various units and facilities listed above, the entire image should be deleted. Among them, close-up images refer to images in which the field of view is focused on the subject of the shooting, and the environmental space where the subject of the shooting is located is basically excluded from the picture.

[0058] 2. Local technical processing of real scene images:

[0059] If the real-life image contains the above-listed various units and facilities, clearly visible signs in the side view or distant view, partial facility bodies and other characteristic identification areas, as well as the image area containing the above-listed attribute information signs, local technical processing should be performed, and the technical processing area cannot be clearly identified or have obvious signs of abnormal processing. Among them, the side view refers to the image of the environmental space where the subject is not presented in the front and contains part of the subject; the distant view refers to the image with a deep and open field of vision, which does not highlight a certain type of subject.

[0060] In some embodiments, geographic sensitive information in real-life images is identified based on image recognition results, confidence levels of the image recognition results, and a recall strategy.

[0061] In some embodiments, the image recognition result includes a text recognition result and an image recognition result.

[0062] In some embodiments, in step S120, an image recognition method (such as an image recognition model) may be used to identify geographic sensitive information, and information such as identification category, identification element, element coordinates, and confidence level may be returned. The geographic sensitive information may include image features of units and facilities involved in the real-life image that should be technically processed, and image features of attribute information involved in the real-life image that should be technically processed. For example, it may be identified that a real-life image contains an oil reserve, the coordinates of the oil reserve in the real-life image, and the confidence level.

[0063] In some embodiments, in step S120, a text recognition method, such as an optical character recognition (OCR) method, may be used to recognize text in the real-scene image to determine the terms contained in the real-scene image, the image coordinates of the terms, and the confidence level of the terms. For example, the terms containing the content "First Hospital" in the real-scene image, the coordinates of the image of the terms in the real-scene image, and the confidence level of the terms are identified. Based on the identified terms and the pre-set matching word library, it can be determined whether the terms are sensitive words.

[0064] Personal sensitive information and sensitive attribute information identified by image recognition methods can be directly used in subsequent image processing processes, that is, processing images containing personal sensitive information and sensitive attribute information, where personal sensitive information can include faces, license plates, etc.; for unit and facility attribute information identified by image recognition methods, and sensitive words identified by text recognition methods and pre-set matching word libraries, the sensitive elements (including text sensitive elements and image sensitive elements) contained in the real-life image can be determined based on a pre-set recall strategy. Afterwards, the determined sensitive elements or sensitive real-life images can be manually verified, and then the sensitive elements or sensitive real-life images determined by manual verification can be processed. The aforementioned processing methods may include:

[0065] (1) Delete the entire real-life image.

[0066] (2) Performing local technical processing on real-scene images, so that the technically processed area cannot be clearly identified or has obvious signs of abnormal processing.

[0067] In some embodiments, step S130 may include: based on the image recognition result and confidence, the text recognition result and confidence, and the recall strategy, identifying the geographically sensitive information in the real-life image. For example, the geographically sensitive information here includes geographically sensitive information of units and facilities. Confidence can be understood as the reliability of the recognition result. The disclosed embodiment only performs sensitive information recognition on the text and / or image content of the reliability of the recognition result, thereby improving the reliability and accuracy of sensitive information recognition.

[0068] In some embodiments, for image recognition and text recognition, when the confidence of the recognition result is greater than or equal to a preset threshold, the geographically sensitive information in the real-life image is determined based on the recognition result and the recall strategy, that is, whether the recognition result is sensitive information; when the confidence of the recognition result is less than the preset threshold, the recognition result is considered unreliable, and the subsequent determination of whether the recognition result is sensitive information is no longer made. The disclosed embodiment assumes that the recognition result is not sensitive information. In one example, the preset threshold is 90%. For example, when the OCR method is used to identify the existence of the term "First Hospital" in the real-life image, and the confidence of the recognition result is 80%, which is less than the preset threshold, the term is not identified as sensitive information, and the image of the term is marked as insensitive.

[0069] In some embodiments, the recall strategy includes at least one of the following:

[0070] (1) when the matching degree between the text recognition result and the detection word library is greater than or equal to a first threshold, and the matching degree between the text recognition result and the exemption word library is less than a second threshold, the text recognition object is determined as a text sensitive element;

[0071] (2) When the image recognition result corresponds to geographically sensitive information, the object of the image recognition is determined as an image sensitive element;

[0072] (3) When there are sensitive elements in the real scene image, objects within a range with the trajectory point corresponding to the real scene image as the center and the first preset length as the radius are determined as sensitive elements;

[0073] (4) When the recognition frame of the image recognition or text recognition is smaller than or equal to the first size, the object of the image recognition or text recognition is determined as a non-sensitive element.

[0074] For the recall strategy in item (1) above, detection word libraries and exemption word libraries can be pre-set. Among them, the detection word library can include military detection word libraries, such as those containing words such as "troops", "navy", "military doctor", etc., and also include other types of detection word libraries, such as those containing "Ankang Hospital", "Mental Health Center", etc.; the detection word library is used to identify sensitive content in images. The exemption word library can include military exemption word libraries and other categories of exemption word libraries, which are mainly used to exclude those words that appear in the image but are reasonable expressions in specific scenarios, and the image should not be marked as sensitive.

[0075] In some embodiments, the first threshold is greater than or equal to 60%. For example, the first threshold is 75%. Setting the first threshold can determine entries that partially match words in the detection word library as sensitive words, thereby improving the accuracy of sensitive word recall.

[0076] In some implementations, the second threshold is 100%. By setting the second threshold, entries that completely match words in the exemption word library can be determined as non-sensitive words, thereby avoiding misidentification of reasonable expressions in some specific scenarios as sensitive words.

[0077] For example, using the OCR method, it is recognized that the real-life image contains "Navy Medical University", and the confidence of the recognition result is 95%, which is greater than the preset threshold (such as 90%); then the detection word library and the exemption word library are used to match the entry. The matching degree of the entry with "Navy" and "Military Doctor" in the detection word library is 100%, which is greater than the preset first threshold; and the matching degree of the entry with the relevant entry in the exemption word library is 0%, which is less than the second threshold, then it is determined that the object of the text recognition is a sensitive element.

[0078] For another example, using the OCR method, it is recognized that the real-life image contains "Ankang City Hospital", and the confidence of the recognition result is 95%, which is greater than the preset threshold (such as 90%); then the detection word library and the exemption word library are used to match the entry. The matching degree of this entry with "Ankang Hospital" in the detection word library is 80% (4 out of 5 characters match), which is greater than the preset first threshold; and the matching degree of this entry with the related entry of "Ankang City Hospital" in the exemption word library is 100%, which is equal to the second threshold, so it is determined that the object of the text recognition is not a sensitive element.

[0079] For another example, using the OCR method, it is recognized that the real scene image contains "Navy Medical University", and the confidence of the recognition result is 80%, which is less than a preset threshold (such as 90%), then it is determined that the object of the text recognition is not a sensitive element.

[0080] By using the detection and exemption word library to detect sensitive words, we can not only identify words that are not allowed to appear in public real-life images, but also exclude those reasonable expressions that appear in the image but belong to specific scenarios, thereby improving the accuracy of sensitive word recognition.

[0081] In addition, the recall strategy only performs sensitive word recognition on entries whose confidence level of text recognition is greater than or equal to a preset threshold, thus avoiding the problem of sensitive word recognition errors caused by low confidence level of text recognition.

[0082] For the recall strategy of item (2) above, when the confidence of the image recognition result is greater than or equal to the preset threshold, if the image recognition result corresponds to geographic sensitive information, the object of image recognition is determined as an image sensitive element.

[0083] For example, for the recognition results of radar, the August 1st Army emblem, optical cables, substation facilities, cooling towers, cylindrical tanks, spherical tanks, etc., the preset threshold can be set to 0%, that is, if radar, the August 1st Army emblem, optical cables, substation facilities, cooling towers, cylindrical tanks or spherical tanks are recognized from the real-life image, and the confidence of the recognition result is greater than 0%, the object of the image recognition is determined as an image sensitive element.

[0084] For example, for recognition results such as guard rails and camouflaged houses, the preset threshold can be set to 80%. That is, if a guard rail or a camouflaged house is recognized from a real-life image, and the confidence level of the recognition result is greater than or equal to 80%, the object recognized in the image is determined to be an image sensitive element.

[0085] This recall strategy only recognizes image sensitive elements for objects whose image recognition confidence is greater than or equal to a preset threshold, thus avoiding the problem of incorrect recognition of image sensitive elements due to low confidence in image recognition.

[0086] For the recall strategy of item (3) above, the trajectory points of the collection vehicle can be calculated according to the element size and the actual model recognition situation, so that objects within a certain range around the recognition image can be recalled according to the same sensitive type. That is, when there are sensitive elements (including text sensitive elements and / or image sensitive elements) in the real scene image, the objects within the range with the trajectory point corresponding to the real scene image as the center and the first preset length as the radius are determined as sensitive elements. The first preset length can be determined according to the type of the recognized object.

[0087] For example, when the sensitive elements existing in the real-scene image are image sensitive elements, the first preset length is 40 meters; for example, when sensitive elements such as the August 1st Army Emblem, radar, guardrail or camouflage house are identified in the real-scene image, objects within a circular area with a radius of 40 meters and the trajectory point corresponding to the real-scene image as the center will be recalled with the same sensitive type.

[0088] For example, when the sensitive element existing in the real-scene image is a text sensitive element, the first preset length is 25 meters; for example, when a sensitive word is identified in the real-scene image, objects within a circular area with a radius of 25 meters and the trajectory point corresponding to the real-scene image as the center will be recalled with the same sensitive type.

[0089] Since the image content corresponding to image sensitive elements is generally larger in volume, and the image content corresponding to text sensitive elements is generally smaller in volume, the disclosed embodiment sets different first preset lengths for text sensitive elements and image sensitive elements, respectively, so as to realize sensitive element recall according to different situations.

[0090] When a sensitive element appears in a real-life image, there is a high probability that other sensitive elements of the same type exist in its vicinity. Therefore, by adopting this recall strategy, objects around the sensitive element in the real-life image can be recalled. The recalled content can be used for subsequent manual reconfirmation, thereby reducing the amount of calculation when automatically identifying sensitive elements and improving recall efficiency.

[0091] The recall strategy of item (4) above belongs to the ignore box strategy, that is, if the recognition box is too small, the feature of the recognition element is not clear and it may not be marked as sensitive. Specifically, when the object of image recognition or text recognition is a cooling tower or an oil storage tank, the first size is set to 20 pixels wide and 60 pixels high; or 60 pixels wide and 20 pixels high. That is, if the recognition box of the cooling tower or the oil storage tank is less than or equal to 20 pixels * 60 pixels, or less than or equal to 60 pixels * 20 pixels, it will not be marked as a sensitive element.

[0092] For another example, when the object of image recognition or text recognition is neither a cooling tower nor an oil storage tank, the first size is set to 10 pixels in width and 10 pixels in height. That is, for other recognition objects except cooling towers and oil storage tanks, if the recognition box is less than or equal to 10 pixels * 10 pixels, it is not marked as a sensitive element.

[0093] Since the element features of very small objects in real-life images are unclear, such objects generally cannot carry effective information. By setting the ignore box strategy, objects with unclear features can be excluded and not set as sensitive elements, thereby improving the accuracy of sensitive element recognition.

[0094] In some implementations, the recall strategy proposed in the embodiments of the present disclosure may also include a continuous / multi-factor strategy, including at least one of the following:

[0095] (1) When the same sensitive element exists in M ​​real-scene images that are continuous in time sequence, determine the middle real-scene image among the M real-scene images;

[0096] Other real scene images within a range with the trajectory point corresponding to the intermediate real scene image as the center and the second preset length as the radius are marked as sensitive real scene images.

[0097] In one example, M is 5.

[0098] In one example, the second preset length is 40 meters.

[0099] For example, if the same text sensitive element / image sensitive element is matched in five consecutive images, it means that there is a high possibility that sensitive information exists here. In this case, the trajectory points can be calculated, and images within 40 meters around the middle image can be taken and marked as sensitive.

[0100] When the same sensitive element appears continuously in multiple real-life images, it means that the range or volume of the sensitive element is large. In this case, the possibility of the sensitive element appearing in the adjacent range is also greater; therefore, by adopting the above recall strategy, multiple real-life images where sensitive elements may appear can be recalled at one time, thereby reducing the amount of calculation when identifying sensitive elements and improving recall efficiency.

[0101] It is easy to understand that the larger M is, the less likely it is that a one-time recall will occur; the larger the second preset length is, the more content will be recalled at one time. The specific values ​​of M and the second preset length can be set according to specific circumstances.

[0102] (2) When the number of sensitive elements in a real-scene image is greater than or equal to N, the real-scene image is marked as a sensitive real-scene image, and other real-scene images within a range with a trajectory point corresponding to the real-scene image as the center and a third preset length as the radius are marked as sensitive real-scene images.

[0103] In one example, N is greater than or equal to 2. For example, N is 3.

[0104] In one example, the third preset length is 40 meters.

[0105] When multiple sensitive elements appear in a real-life image, it means that there is a greater possibility that sensitive elements appear within the range of the real-life image and in the vicinity. Therefore, by adopting the above recall strategy, multiple real-life images where sensitive elements may appear can be recalled at one time, thereby reducing the amount of calculation when identifying sensitive elements and improving recall efficiency.

[0106] It is easy to understand that the larger N is, the less the one-time recall is; the larger the third preset length is, the more the one-time recall is. The specific values ​​of N and the third preset length can be set according to the specific situation.

[0107] For example, if a picture matches multiple elements / vocabulary (such as 3 or more), it means that the picture is likely to contain sensitive information. In this case, the trajectory points can be calculated, and pictures within 40 meters around the picture can be taken and marked as sensitive.

[0108] (3) When there are text sensitive elements in a real-scene image, if the same text sensitive elements do not exist in the X real-scene images before and after the real-scene image in the time sequence of the real-scene images, the real-scene image is marked as insensitive.

[0109] In one example, X is 5.

[0110] For example, if a sensitive word in the detection vocabulary is matched in an image, and the sensitive word is not matched in the five images before and after the image, the image will not be marked as sensitive.

[0111] This method can reduce the problem of mislabeling caused by text recognition errors. For example, among multiple consecutive real-life images, if only one real-life image contains a sensitive word, and the sensitive word does not appear in the real-life images before and after it, it is very likely that the real-life image with the detected sensitive word has a recognition error. Therefore, the real-life image can be marked as insensitive to correct the recognition error.

[0112] It is easy to understand that the larger the value of X, the stricter the evaluation criteria for text recognition errors. For example, the larger the value of X, the more it is considered a text recognition error when a sensitive word appears only once in a longer time range; the smaller the value of X, the more it is considered a text recognition error when a sensitive word appears only once in a shorter time range, so the evaluation criteria for text recognition errors are looser. The value of X can be determined according to the specific situation. For example, if the criteria for determining sensitive information are stricter, the value of X is larger; if the criteria for determining sensitive information are looser, the value of X is smaller.

[0113] (4) When there is a single image-sensitive element corresponding to a camouflaged house or a guardrail in the real-scene image, the real-scene image is marked as insensitive.

[0114] Since camouflaged houses or guard rails are also common in scenes of non-sensitive units and facilities, setting this recall strategy can avoid mislabeling caused by camouflaged houses or guard rails.

[0115] After the above identification, sensitive real-life images and sensitive elements in the real-life images can be identified. After that, the images that are determined to contain sensitive information of units and facilities can be rechecked to mark the sensitive categories, sensitive areas and processing methods of the images.

[0116] According to the program recall and the results of manual judgment, the real scene image can be deleted, photoshopped or blurred. Specifically:

[0117] (1) Deletion means deleting the entire real-scene image and not displaying it.

[0118] (2) PS processing, including using blur processing tools to automatically match pixels around the picture frame based on the sensitive information coordinate frame and form a background area.

[0119] (3) Blurring: This may include mosaic processing of sensitive information boxes.

[0120] Figure 2 Schematic diagram of the functions of each processing module according to an embodiment of the present disclosure. Figure 2 As shown, the present disclosure uses an image recognition module 210 to perform image recognition on a picture; and uses a text recognition module 220 to perform text recognition on a picture, such as using an OCR method to perform text recognition on a picture. The picture may be a picture collected during the driving process of the vehicle, and each picture corresponds to a track point of the vehicle; the picture may be used as a real-scene image in a real-scene map.

[0121] Specifically, the image recognition module 210 can recognize the following elements and return the recognition category, recognition element, element coordinates and confidence level.

[0122] (1) Geographically sensitive information image elements: Geographically sensitive information may include units and facilities involved in real-life images that should be technically processed as specified in relevant regulations, including the following types:

[0123] a. Sensitive unit image features, such as image features of military unit gates (such as badges).

[0124] b. Sensitive facility image features, such as public safety facility image features "oil storage depot", such as Figure 3 As shown; the image feature of civil facilities "cooling tower", such as Figure 4 As shown;

[0125] c. Sensitive attribute information image features, such as the height limit, width limit, clearance, load and slope attributes of bridges; the height and width attributes of tunnels; and the pavement material attributes of highways. Figure 5 shown.

[0126] (2) Personal sensitive information image elements:

[0127] The text recognition module 220 can recognize the text in the image and return the recognized term, the image coordinates of the term and the confidence level. The text recognition module 220 can use OCR technology for recognition.

[0128] like Figure 2 As shown, the real-life map processing framework proposed in the embodiment of the present disclosure also includes a sensitive word matching module 230, which is used to determine whether the entry recognized by the text recognition module 220 is a sensitive word. A matching word library can be set in the sensitive word matching module 230, and the matching word library can include a detection word library and an exemption word library. Among them, the detection word library is mainly used to identify sensitive content in an image or picture. If the match is successful, the image is marked as sensitive. Specifically, the following detection word libraries can be set:

[0129] (1) Military detection word library, such as "troops", "navy", "military doctor", etc.;

[0130] (2) Other detection word libraries, such as "Ankang Hospital", "Mental Health Center", etc.

[0131] The exemption word library is mainly used to exclude those words that appear in the image but are reasonable expressions in specific scenarios, and the image should not be marked as sensitive. The exemption word library can include the following:

[0132] (1) Setting up military exemption word libraries, such as “Navy.*(store|supermarket|hotel|memorial|(Navy.*(store|supermarket|hotel|memorial|company|business|old site|city|wholesale|barbecue|place|shop|store|equipment|building|pavilion|building materials|hospital|building|clinic|intersection|martyr)|Navy(garden|museum|hotel|tobacco and alcohol|road|farmhouse|decoration|street|supermarket|dental|electrical appliances))company|business|old site|city|wholesale|barbecue|place|shop|store|equipment|building|pavilion|building materials|hospital|building|clinic|intersection|martyr)|Navy(garden|museum|hotel|tobacco and alcohol|road|farmhouse|decoration|street|supermarket|dental|electrical appliances)”, “Military hot pot”;

[0133] (2) Other categories of exemption vocabulary, such as "Ankang Central Hospital", "Ankang People's Hospital", "Ankang Traditional Chinese Medicine Hospital", etc.

[0134] Using the above detection word library and exemption word library, the sensitive word matching module 230 can adopt a matching strategy to match sensitive words.

[0135] The matching strategy can be set as follows: if the detection word library is hit and the exemption word library is not hit, it is considered sensitive.

[0136] In one example, if the match between the detected word and the detection word library is greater than or equal to 75% (the threshold is configurable), and the confidence of the detected word is greater than or equal to 90% (the threshold is configurable), the detected word is marked as a sensitive type.

[0137] In another example, if the match degree between the detected word and the exemption rule vocabulary is equal to 100% (the threshold is configurable), and the confidence of the detected word is greater than or equal to 90% (the threshold is configurable), the detected word will not be marked as a sensitive type and will proceed to the next stage.

[0138] In another example, if the match between the detected word and the detection word library is less than 75% (this threshold is configurable), that is, there is no match to the content in the detection word library, the detected word will not be marked as a sensitive type and will enter the next stage.

[0139] For example, if the confidence of the term "Navy Medical University" is greater than a preset threshold (such as 90%), and the term 100% matches the military detection terms "navy" and "military doctor", and the term does not match the exemption vocabulary, then the term is marked as military sensitive.

[0140] like Figure 2 As shown, the real-life map processing framework proposed in the embodiment of the present disclosure also includes a strategy fusion module 240, which is used to determine whether the image is sensitive and the sensitive type based on the sensitive words identified by the sensitive word matching module 230 and the unit and facility image sensitive elements identified by the image recognition module 210.

[0141] In one example, a recall strategy is pre-set based on different features of the image and OCR sensitive information, and the strategy fusion module 240 can comprehensively calculate whether the image is sensitive and the sensitive category based on the recognition result and the recall strategy. The recall strategy includes at least one of the following:

[0142] 1. Recognition confidence strategy: When the confidence of the recognized element is greater than or equal to the set threshold, the recognition result is considered correct.

[0143] (1) The confidence threshold of text recognition is 80%; when the confidence of text recognition is greater than or equal to 80%, the text recognition result is considered to be credible. Then, it is further determined whether the text recognition result is a sensitive word.

[0144] (2) Setting the confidence threshold for image feature recognition:

[0145] In one example, for radar, August 1st Army Emblem, optical cable, substation facility, cooling tower, cylindrical tank, and spherical tank, the confidence threshold is set to 0%;

[0146] In one example, a guard railing, a camouflaged house, and the confidence threshold is set to 80%.

[0147] For example, if an image recognition method is used to identify the presence of radar, the August 1st Army Emblem, optical cables, substation facilities, cooling towers, cylindrical tanks, or spherical tanks in an image, and the confidence level of the recognition result is greater than or equal to 0%, then the recognition result is considered correct, meaning that sensitive elements exist in the image.

[0148] For another example, if an image recognition method is used to identify the presence of a guardrail or a camouflaged house in an image, and the confidence level of the recognition result is greater than or equal to 80%, then the recognition result is considered correct, i.e., there are sensitive elements in the image.

[0149] 2. Spatial recall strategy: according to the size of the elements and the actual model recognition situation, the trajectory points are calculated, and the x (x can be set) meters around the recognized image are recalled according to the same confidential type;

[0150] (1) Image recognition elements (August 1st Army emblem, radar, guardrail, camouflage house) are recalled according to the same sensitive type within 40 meters;

[0151] (2) OCR method: recall the same sensitive type within a 25m radius (this range is configurable).

[0152] 3. Ignore the box strategy. If the recognition box is too small, the features of the recognition elements are not clear and may not be marked as sensitive.

[0153] (1) When the recognition box is less than or equal to 10 pixels × 10 pixels, the element is judged as insensitive;

[0154] (2) When the identification frame of the cooling tower or oil storage tank is less than or equal to 20 pixels × 60 pixels / 60 pixels × 20 pixels, this element is judged to be insensitive.

[0155] 4. Continuous / multi-factor strategy:

[0156] (1) If the same element / word is matched in five consecutive images, it means that there is a high possibility of sensitive information. The trajectory points are calculated, and images within 40m (the range is configurable) around the middle image are selected and marked as sensitive;

[0157] (2) If a picture matches multiple elements / vocabulary (for example, greater than or equal to 3, the threshold is configurable), it means that the picture is likely to contain sensitive information. The trajectory points are calculated, and pictures within 40m (the range is configurable) around the picture are taken and marked as sensitive;

[0158] (3) If an image matches a sensitive word in the detection vocabulary, and multiple images before and after the image (e.g., 5 images) do not match the sensitive word in chronological order, then the image will not be marked as sensitive;

[0159] (4) When the camouflage house or guard railing appears alone, the image is not sensitive to the mark.

[0160] After the strategy fusion module 240 uses the above recall strategy to mark the sensitive map or the sensitive elements in the sensitive map, the manual operation module 250 can be used to perform a secondary check on the content marked as the sensitive map or sensitive element. After that, the result processing module 260 combines the recall strategy and the result of manual judgment to finally determine the sensitive map or sensitive element, and processes the image based on the result, including deletion, PS processing or blurring.

[0161] Among them, PS processing can use fuzzy processing tools to automatically match the pixels around the picture frame based on the sensitive information coordinate frame and form a background area, such as Figure 6 As shown in an example, the images before and after processing are as follows Figure 7 , 8 shown.

[0162] The present disclosure also provides a real-scene map processing device. Fig. 9 is a structural diagram of a real scene map processing device 900 according to an embodiment of the present disclosure, comprising:

[0163] An acquisition module 910 is used to acquire a real scene image of a real scene map;

[0164] A first recognition module 920 is used to image the real scene image;

[0165] The second recognition module 930 is used to recognize geographically sensitive information in the real-scene image based on the image recognition result and the recall strategy; wherein the geographically sensitive information includes units and facilities involved in the real-scene image that should be technically processed, and attribute information involved in the real-scene image that should be technically processed;

[0166] The marking module 940 is used to mark the real scene image according to geographical sensitive information.

[0167] In some implementations, the second identification module 930 is further configured to:

[0168] Based on the image recognition results, the confidence of the image recognition results and the recall strategy, the geographically sensitive information in the real-life image is identified.

[0169] In some embodiments, the image recognition result includes a text recognition result and an image recognition result.

[0170] In some embodiments, the recall strategy includes at least one of the following:

[0171] When the matching degree between the result of text recognition and the detection word library is greater than or equal to the first threshold, and the matching degree between the result of text recognition and the exemption word library is less than the second threshold, determining the object of text recognition as a text sensitive element;

[0172] When the result of image recognition corresponds to geographically sensitive information, the object of image recognition is determined as an image sensitive element;

[0173] In the case where there are sensitive elements in the real scene image, objects within a range with the trajectory point corresponding to the real scene image as the center and the first preset length as the radius are determined as sensitive elements;

[0174] When the recognition frame of the image recognition or text recognition is smaller than or equal to the first size, the object of the image recognition or text recognition is determined as a non-sensitive element.

[0175] In some embodiments, the first threshold is greater than or equal to 60%.

[0176] In some embodiments, the first threshold is 75%.

[0177] In some embodiments, the second threshold is 100%.

[0178] In some embodiments,

[0179] When the sensitive element existing in the real scene image is an image sensitive element, the first preset length is 40 meters;

[0180] When the sensitive element in the real scene image is a text sensitive element, the first preset length is 25 meters.

[0181] In some embodiments, when the object of image recognition or text recognition is a cooling tower or an oil storage tank, the first size is 20 pixels wide and 60 pixels high; or 60 pixels wide and 20 pixels high;

[0182] When the object of image recognition or text recognition is neither a cooling tower nor an oil storage tank, the first size is 10 pixels in width and 10 pixels in height.

[0183] In some embodiments, the recall strategy further comprises:

[0184] In the case where the same sensitive element exists in M ​​real-scene images that are continuous in time sequence, determining the middle real-scene image among the M continuous real-scene images;

[0185] Other real-scene images within a range with the trajectory point corresponding to the middle real-scene image as the center and the second preset length as the radius are marked as sensitive real-scene images.

[0186] In some embodiments, M is 5.

[0187] In some embodiments, the second preset length is 40 meters.

[0188] In some embodiments, the recall strategy further comprises:

[0189] When the number of sensitive elements in the real-scene image is greater than or equal to N, the real-scene image is marked as a sensitive real-scene image, and other real-scene images with the trajectory point corresponding to the real-scene image as the center and the third preset length as the radius are marked as sensitive real-scene images.

[0190] In some embodiments, N is greater than or equal to 2.

[0191] In some embodiments, N is 3.

[0192] In some implementations, the third preset length is 40 meters.

[0193] In some embodiments, the recall strategy further comprises:

[0194] In the case where there are text sensitive elements in the real-scene image, according to the time sequence of the real-scene images, if the same text sensitive elements do not exist in the X real-scene images before and after the real-scene image, the real-scene image is marked as insensitive.

[0195] In some embodiments, X is 5.

[0196] In some embodiments, the recall strategy further comprises:

[0197] When there is a single image-sensitive element corresponding to a camouflaged house or a guard railing in the real-scene image, the real-scene image is marked as insensitive.

[0198] In some implementations, the units and facilities involved in the real-life images that should be technically processed include at least one of the following:

[0199] Various military facilities that directly serve military purposes;

[0200] Military restricted areas, military management areas and all units and facilities within them;

[0201] Key departments or units involving national security;

[0202] Units and facilities related to public safety;

[0203] Civilian facilities that involve the national economic lifeline and have a significant impact on people's production and life;

[0204] Dedicated railways and train lines within stations, railway marshalling yards, and dedicated roads.

[0205] In some implementations, the attribute information involved in the real scene image to be technically processed includes at least one of the following:

[0206] The height limit, width limit, clearance, load capacity and slope attributes of bridges; the height and width attributes of tunnels; the pavement material attributes of highways;

[0207] The navigability, water depth, flow rate, bottom quality and bank quality of rivers; the storage capacity of reservoirs; the height of dams; the nature of water sources; the water depth and mud depth of swamps and their boundary contours; the properties of ferry crossings;

[0208] Exposes the airport's transport capacity attributes.

[0209] For the description of specific functions and examples of each module and submodule of the device in the embodiment of the present disclosure, reference can be made to the relevant description of the corresponding steps in the above method embodiment, which will not be repeated here.

[0210] In the technical solution disclosed in the present invention, the acquisition, storage and application of personal information of users involved are in compliance with the provisions of relevant laws and regulations and do not violate public order and good morals.

[0211] According to an embodiment of the present disclosure, the present disclosure also provides an electronic device, a readable storage medium and a computer program product.

[0212] Fig.10 A schematic block diagram of an example electronic device 1000 that can be used to implement an embodiment of the present disclosure is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital assistants, cellular phones, smart phones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present disclosure described and / or required herein.

[0213] like Fig.10 As shown, the device 1000 includes a computing unit 1001, which can perform various appropriate actions and processes according to a computer program stored in a read-only memory (ROM) 1002 or a computer program loaded from a storage unit 1008 into a random access memory (RAM) 1003. In the RAM 1003, various programs and data required for the operation of the device 1000 can also be stored. The computing unit 1001, the ROM 1002, and the RAM 1003 are connected to each other via a bus 1004. An input / output (I / O) interface 1005 is also connected to the bus 1004.

[0214] A number of components in the device 1000 are connected to the I / O interface 1005, including: an input unit 1006, such as a keyboard, a mouse, etc.; an output unit 1007, such as various types of displays, speakers, etc.; a storage unit 1008, such as a disk, an optical disk, etc.; and a communication unit 1009, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 1009 allows the device 1000 to exchange data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0215] The computing unit 1001 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the computing unit 1001 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, digital signal processors (DSPs), and any appropriate processors, controllers, microcontrollers, etc. The computing unit 1001 performs the various methods and processes described above, such as detection methods. For example, in some embodiments, the detection method may be implemented as a computer software program, which is tangibly contained in a machine-readable medium, such as a storage unit 1008. In some embodiments, part or all of the computer program may be loaded and / or installed on the device 1000 via ROM 1002 and / or communication unit 1009. When the computer program is loaded into RAM 1003 and executed by the computing unit 1001, one or more steps of the detection method described above may be performed. Alternatively, in other embodiments, the computing unit 1001 may be configured to perform the detection method in any other appropriate manner (e.g., by means of firmware).

[0216] Various implementations of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0217] The program code for implementing the method of the present disclosure may be written in any combination of one or more programming languages. These program codes may be provided to a processor or controller of a general-purpose computer, a special-purpose computer, or other programmable data processing device, so that the program code, when executed by the processor or controller, enables the functions / operations specified in the flow chart and / or block diagram to be implemented. The program code may be executed entirely on the machine, partially on the machine, partially on the machine and partially on a remote machine as a stand-alone software package, or entirely on a remote machine or server.

[0218] In the context of the present disclosure, a machine-readable medium may be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, device, or equipment. A machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or equipment, or any suitable combination of the foregoing. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0219] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user can provide input to the computer. Other types of devices can also be used to provide interaction with the user; for example, the feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including acoustic input, voice input, or tactile input).

[0220] The systems and techniques described herein may be implemented in a computing system that includes back-end components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes front-end components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such back-end components, middleware components, or front-end components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), and the Internet.

[0221] A computer system may include a client and a server. The client and the server are generally remote from each other and usually interact through a communication network. The relationship of client and server is generated by computer programs running on respective computers and having a client-server relationship with each other. The server may be a cloud server, a server of a distributed system, or a server combined with a blockchain.

[0222] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps recorded in this disclosure can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solutions disclosed in this disclosure can be achieved, and this document does not limit this.

[0223] The above specific implementations do not constitute a limitation on the protection scope of the present disclosure. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the principles of the present disclosure shall be included in the protection scope of the present disclosure.

Claims

1. A method for processing a real-life map, comprising: Obtain the real-life image of the real-life map; Performing image recognition on the real scene image; Based on the image recognition result and the recall strategy, identify the geographically sensitive information in the real-scene image; wherein the geographically sensitive information includes the units and facilities involved in the real-scene image that should be technically processed, and the attribute information involved in the real-scene image that should be technically processed; The real scene image is marked according to the geographic sensitive information.

2. The method according to claim 1, wherein: The method further comprises: Based on the image recognition result, the confidence of the image recognition result and the recall strategy, the geographically sensitive information in the real-life image is identified.

3. The method according to claim 2, wherein: The image recognition result includes a text recognition result and an image recognition result.

4. The method according to claim 3, wherein: The recall strategy includes at least one of the following: When the matching degree between the text recognition result and the detection word library is greater than or equal to the first threshold, and the matching degree between the text recognition result and the exemption word library is less than the second threshold, determining the object of text recognition as a text sensitive element; In a case where the image recognition result corresponds to the geographically sensitive information, determining the object of image recognition as an image sensitive element; In the case where there are sensitive elements in the real scene image, objects within a range with a trajectory point corresponding to the real scene image as a circle center and a first preset length as a radius are determined as sensitive elements; When the recognition frame of the image recognition or text recognition is smaller than or equal to the first size, the object of the image recognition or text recognition is determined as a non-sensitive element.

5. The method according to claim 4, wherein: The first threshold is greater than or equal to 60%.

6. The method according to claim 5, wherein: The first threshold is 75%.

7. The method according to claim 4, wherein: The second threshold is 100%.

8. The method according to claim 4, wherein: In the case where the sensitive element present in the real scene image is an image sensitive element, the first preset length is 40 meters; When the sensitive element in the real scene image is a text sensitive element, the first preset length is 25 meters.

9. The method according to claim 4, wherein: When the object of the image recognition or text recognition is a cooling tower or an oil storage tank, the first size is 20 pixels in width and 60 pixels in height; or 60 pixels in width and 20 pixels in height; When the object of the image recognition or text recognition is neither a cooling tower nor an oil storage tank, the first size is 10 pixels in width and 10 pixels in height.

10. The method according to any one of claims 4 to 9, wherein: The recall strategy also includes: In the case where the same sensitive element exists in the M real-scene images that are continuous in time sequence, determining an intermediate real-scene image among the M continuous real-scene images; Other real scene images within a range with the trajectory point corresponding to the intermediate real scene image as the center and the second preset length as the radius are marked as sensitive real scene images.

11. The method according to claim 10, wherein: The M is 5.

12. The method according to claim 10, wherein: The second preset length is 40 meters.

13. The method according to any one of claims 4 to 9, wherein: The recall strategy also includes: When the number of sensitive elements in the real-scene image is greater than or equal to N, the real-scene image is marked as a sensitive real-scene image, and other real-scene images with the trajectory point corresponding to the real-scene image as the center and a third preset length as the radius are marked as sensitive real-scene images.

14. The method according to claim 13, wherein: The N is greater than or equal to 2.

15. The method according to claim 14, wherein: The N is 3.

16. The method according to claim 13, wherein: The third preset length is 40 meters.

17. The method according to any one of claims 4 to 9, wherein: The recall strategy also includes: In the case where there are text sensitive elements in the real-scene image, according to the time sequence of the real-scene images, if the same text sensitive elements do not exist in the X real-scene images before and after the real-scene image, the real-scene image is marked as insensitive.

18. The method according to claim 17, wherein: The X is 5.

19. The method according to any one of claims 4 to 9, wherein: The recall strategy also includes: In the case where there is only an image-sensitive element corresponding to a camouflaged house or a guard railing in the real-scene image, the real-scene image is marked as insensitive.

20. The method according to any one of claims 1 to 9, wherein: The units and facilities involved in the real-life images that should be technically processed include at least one of the following: Various military facilities that directly serve military purposes; Military restricted areas, military management areas and all units and facilities within them; Key departments or units involving national security; Units and facilities related to public safety; Civilian facilities that involve the national economic lifeline and have a significant impact on people's production and life; Dedicated railways and train lines within stations, railway marshalling yards, and dedicated roads.

21. The method according to any one of claims 1 to 9, wherein: The attribute information involved in the real scene image to be technically processed includes at least one of the following: The height limit, width limit, clearance, load capacity and slope attributes of bridges; the height and width attributes of tunnels; the pavement material attributes of highways; The navigability, water depth, flow rate, bottom quality and bank quality of rivers; the storage capacity of reservoirs; the height of dams; the nature of water sources; the water depth and mud depth of swamps and their boundary contours; the properties of ferry crossings; Exposes the airport's transport capacity attributes.

22. A real-life map processing device, comprising: An acquisition module is used to acquire the real scene image of the real scene map; A first recognition module, used for performing image recognition on the real scene image; A second recognition module is used to recognize geographically sensitive information in the real-scene image based on the image recognition result and the recall strategy; wherein the geographically sensitive information includes units and facilities involved in the real-scene image that should be technically processed, and attribute information involved in the real-scene image that should be technically processed; The marking module is used to mark the real scene image according to the geographical sensitive information.

23. The device according to claim 22, wherein: The second identification module is further used for: Based on the image recognition result, the confidence of the image recognition result and the recall strategy, the geographically sensitive information in the real-life image is identified.

24. An electronic device comprising: at least one processor; as well as 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 to enable the at least one processor to perform the method according to any one of claims 1 to 21.

25. A non-transitory computer-readable storage medium storing computer instructions, wherein: The computer instructions are used to cause the computer to execute the method according to any one of claims 1-21.

26. A computer program product comprising a computer program, which, when executed by a processor, implements the method according to any one of claims 1 to 21.

Images