Personnel trajectory search method, device, equipment and readable storage medium
Patent Information
- Application Number
- CN202210126462.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2022-02-10
- Publication Date
- 2024-12-27
- Estimated Expiration
- 2042-02-10
AI Technical Summary
In the existing surveillance technology, the trajectory search method of target personnel relies on subjective judgment of clues and video player, which can easily lead to misjudgment and it is difficult to accurately determine the whereabouts of target personnel.
By obtaining the monitoring image matching the characteristic information of the target object, the image collection is formed and displayed, in response to user selection, the feature information is fused, and the trajectory is generated until the user inputs the generated trajectory instruction. The method includes summing the confidence level and averaging calculation of feature information, filtering out images with high confidence levels, and generating a trajectory of the target object.
The accuracy of target personnel search is improved, and more accurate trajectories are generated through multiple rounds of filtering and feature information updates.
Smart Images
Figure CN115049971B_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of monitoring technology, and more specifically, to a method, device, equipment, and readable storage medium for finding a person's trajectory. Background Art
[0002] With the continuous development of surveillance technology, surveillance cameras are now deployed in all directions within public places. By overlaying the fields of view of all cameras, the entire public space can be reconstructed. Therefore, theoretically, the whereabouts of a target person can be monitored without blind spots. Currently, surveillance departments rely on clues to retrieve surveillance footage to identify suspected targets. The specific whereabouts of the target person can then be determined based on the playback time of the footage.
[0003] However, the current search method only performs a round of search for the provided clues, and the subsequent judgment is made by the subjective judgment of the video viewers, which may lead to misjudgment of the target person and make it difficult to determine the target person's whereabouts. Summary of the Invention
[0004] In view of the above problems, the present application is proposed to provide a method, device, equipment and readable storage medium for searching a person's trajectory, so as to improve the accuracy of searching a person.
[0005] In order to achieve the above objectives, the following specific plans are proposed:
[0006] A method for finding a person's trajectory, comprising:
[0007] Acquire several monitoring images of the target object that match the characteristic information of the target object, form an image set, and display it;
[0008] In response to a user selecting a target object monitoring image from an image set, determining feature information of a target object image in the target object monitoring image, adding the feature information to a feature information set, and saving the target object monitoring image selected by the user;
[0009] The feature information in the feature information set is fused, and the fused feature information is used as the feature information of the new target object. The step of obtaining a plurality of target object monitoring images that match the feature information of the target object to form an image set is returned to, until a trajectory generation instruction input by the user is received, and the trajectory of the target object is generated according to the saved target object monitoring images.
[0010] Optionally, the feature information includes a plurality of target object image features and a confidence level corresponding to each target object image feature;
[0011] The feature information in the feature information set is fused, and the fused feature information is used as the feature information of the new target object, including:
[0012] For each target object image feature, summing the confidences of the target object image feature in each feature information in the feature information set to obtain a total confidence of the target object image feature;
[0013] For each target object image feature, dividing the total confidence by the number of target object image features whose confidence is not 0 in each feature information to obtain an average confidence of the target object image feature;
[0014] The image features of each target object and the average confidence of each target object image feature are summarized to obtain the feature information of the new target object.
[0015] Optionally, the feature information includes a plurality of target object image features and a confidence level corresponding to each target object image feature;
[0016] After determining the feature information of the target object image in the target object monitoring image, the method further includes:
[0017] In response to a user selecting a specified target object image feature from a plurality of target object image features, the specified target object image feature and the corresponding confidence level are used as new feature information;
[0018] The process of adding the characteristic information to the characteristic information set includes:
[0019] The new feature information is added to the feature information set.
[0020] Optionally, after forming the image set, the method further includes:
[0021] Deleting the saved monitoring images of the same target object from the image set to obtain a processed image set;
[0022] or,
[0023] The target object monitoring image in the image set that is identical to the saved target object monitoring image is marked to inform the user that the target object monitoring image corresponding to the mark has been selected.
[0024] Optionally, the feature information includes a plurality of target object image features and a confidence level corresponding to each target object image feature;
[0025] Obtain several surveillance images of the target object that match the feature information of the target object, including:
[0026] Calculate the matching degree between each image in the target object monitoring image library and each target object image feature in the target object feature information;
[0027] Calculate the average matching degree of the matching degrees of each target object image feature;
[0028] An image with an average matching degree higher than a preset matching threshold is acquired and used as a target object monitoring image.
[0029] Optionally, displaying a plurality of target object surveillance images in the image set includes:
[0030] The target object monitoring images are displayed from front to back according to the average matching degree of each target object monitoring image. Among them, for target object monitoring images with the same average matching degree, the target object monitoring image generated earlier is displayed first.
[0031] A device for finding a person's trajectory, comprising:
[0032] A target object image display unit is used to obtain a plurality of target object monitoring images that match the characteristic information of the target object, form an image set and display it;
[0033] a feature information determining unit, configured to respond to a user's operation of selecting a target object monitoring image from an image set, determine feature information of a target object image in the target object monitoring image, add the feature information to a feature information set, and save the target object monitoring image selected by the user;
[0034] The feature information fusion unit is used to fuse each feature information in the feature information set, and use the fused feature information as the feature information of the new target object, and return to the step of obtaining a plurality of target object monitoring images that match the feature information of the target object to form an image set, until a trajectory generation instruction input by the user is received, and then the trajectory of the target object is generated according to each saved target object monitoring image.
[0035] Optionally, the feature information includes several target object image features and the confidence level corresponding to each target object image feature.
[0036] The feature information fusion unit includes:
[0037] a confidence summing unit, configured to sum, for each target object image feature, the confidences of the target object image features in each piece of feature information in the feature information set to obtain a total confidence of the target object image feature;
[0038] a confidence averaging unit, configured to divide, for each target object image feature, the sum of the confidences by the number of target object image features whose confidences are not 0 in each feature information, to obtain an average confidence of the target object image feature;
[0039] The feature information updating unit is used to summarize the image features of each target object and the average confidence of each target object image feature to obtain new feature information of the target object.
[0040] Optionally, the feature information includes a plurality of target object image features and a confidence level corresponding to each target object image feature;
[0041] The device further comprises:
[0042] a target object image feature selection unit configured to, after determining feature information of a target object image in the target object monitoring image, respond to a user's operation of selecting a specified target object image feature from a plurality of target object image features, and use the specified target object image feature and the corresponding confidence level as new feature information;
[0043] The feature information determining unit includes:
[0044] The selected information adding unit is used to add the new feature information to the feature information set.
[0045] Optionally, the device further includes:
[0046] An image reduction element is used to delete, from the image set, the monitoring images that are identical to the saved target object images, to obtain a processed image set;
[0047] or,
[0048] The image marking unit is used to mark the target object monitoring image in the image set that is the same as the saved target object monitoring image, so as to inform the user that the target object monitoring image corresponding to the mark has been selected.
[0049] Optionally, the feature information includes a plurality of target object image features and a confidence level corresponding to each target object image feature;
[0050] The target object image display unit includes:
[0051] A matching degree calculation unit, used to calculate the matching degree between each image in the target object monitoring image library and each target object image feature in the target object feature information;
[0052] an average matching degree calculation unit, used to calculate the average matching degree of the matching degrees of the image features of each target object;
[0053] The target object monitoring image determining unit is configured to obtain an image with an average matching degree higher than a preset matching threshold value and use the image as the target object monitoring image.
[0054] Optionally, the target object image display unit includes:
[0055] The display sorting unit is used to display the target object monitoring images from front to back according to the average matching degree of each target object monitoring image, wherein, for several target object monitoring images with the same average matching degree, the target object monitoring image generated earlier is displayed in front.
[0056] A device for finding a person's trajectory, comprising: a memory and a processor;
[0057] The memory is used to store programs;
[0058] The processor is used to execute the program to implement the various steps of the aforementioned personnel trajectory search method.
[0059] A readable storage medium stores a computer program, which, when executed by a processor, implements the various steps of the aforementioned method for finding a person's trajectory.
[0060] By means of the above technical solution, the present application obtains several target object surveillance images that match the characteristic information of the target object, forms an image set, and displays the image set. In response to a user selecting a target object surveillance image from the image set, the application determines the characteristic information of the target object image in the target object surveillance image, adds the characteristic information to the characteristic information set, and saves the target object surveillance image selected by the user. The characteristic information in the characteristic information set is fused, and the fused characteristic information is used as the characteristic information of the new target object. The application then returns to the step of obtaining several target object surveillance images that match the characteristic information of the target object and forming the image set, until a user input is received for generating a trajectory, at which point the trajectory of the target object is generated according to each of the saved target object surveillance images. Thus, using the characteristic information of the target object as an initial reference, by updating the characteristic information and the image set, and by providing a preferred selection to the user, multiple rounds of screening are performed to obtain multiple target object surveillance images, thereby making the object found more precise, and on this basis, making the generated trajectory of the target object more accurate. BRIEF DESCRIPTION OF THE DRAWINGS
[0061] Various other advantages and benefits will become apparent to those skilled in the art upon reading the detailed description of the preferred embodiment below. The accompanying drawings are for illustration purposes only and are not to be considered as limiting the present application. The same reference symbols are used throughout the drawings to represent the same components. In the drawings:
[0062] Figure 1 A flowchart of a method for finding a person's trajectory provided in an embodiment of the present application;
[0063] Figure 2-Figure 3 A schematic diagram of a user clicking to select a target object image feature provided in an embodiment of the present application;
[0064] Figure 4-Figure 5 A schematic diagram of selecting a monitoring image of a target object provided in an embodiment of the present application;
[0065] Figure 6 A schematic diagram of the structure of a device for finding a person's trajectory provided in an embodiment of the present application;
[0066] Figure 7 A schematic diagram of the structure of a device for finding a person's trajectory provided in an embodiment of the present application. DETAILED DESCRIPTION
[0067] The following will be combined with the drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are only part of the embodiments of this application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.
[0068] The present application solution can be implemented based on a terminal with data processing capabilities, which can be a mobile phone, computer, server, cloud, etc.
[0069] Next, combine Figure 1 The method for finding a person's track in the present application may include the following steps:
[0070] Step S110: Acquire several monitoring images of the target object that match the characteristic information of the target object, compose an image set, and display the image set.
[0071] Specifically, the source for acquiring the target object monitoring image may be a target object monitoring image library, and the target object monitoring image library may be composed of a plurality of target object images captured from each monitoring snapshot.
[0072] By inputting the target object's characteristic information, the user can match and display surveillance images of the target object with similar characteristics. For example, if the user enters the following characteristic information: red top, yellow hair, female, surveillance images of the target object with similar characteristics will be matched.
[0073] It is understandable that there may be other objects with similar feature information to the target object, and they may be captured and stored in the target object monitoring image library, so that several target object monitoring images may be output.
[0074] Step S120 , in response to the user selecting a target object monitoring image in an image set, determining feature information of the target object image in the target object monitoring image, adding the feature information to a feature information set, and saving the target object monitoring image selected by the user.
[0075] Specifically, in response to the user selecting one of several target object monitoring images, it can indicate that the user has selected the target object monitoring image that is closest to the target object. From this target object monitoring image, all feature information of the target object image is extracted and all the extracted feature information is stored.
[0076] Among them, all the extracted feature information may include all the feature information of the target object, and may also include feature information that is not included in the feature information of the target object. For example, all the extracted feature information: light red top, light yellow hair, female, gray-black pants, middle-aged, among which gray-black pants and middle-aged are not included in the feature information of the target object.
[0077] Step S130: Fusing each feature information in the feature information set, and using the fused feature information as feature information of a new target object.
[0078] It is understood that the feature information contained in the feature information set may be feature information determined each time the user selects a target object monitoring image. The target object monitoring image selected by the user is the image that is closest to the actual target object among the numerous images, and therefore the feature information determined is more realistic. By fusing the feature information in the feature information set and integrating all the user's historical selections, the resulting fused feature information is more objective and more acceptable to the user.
[0079] Specifically, the feature information set may include several feature information, and the fusion method may be to fuse the same type of features of two feature information to obtain new features of the same type.
[0080] For another example, the set contains feature information 1 (light red top, light yellow hair, female, gray-black pants, middle-aged) and feature information 2 (red top, yellow hair, female, black pants), then the fused feature information can be: reddish top, yellowish hair, female, black pants, middle-aged.
[0081] Step S140 , determining whether a trajectory generation instruction input by the user is received, if so, executing step S150 , if not, returning to executing step S110 .
[0082] Specifically, after obtaining the characteristic information of the new target object, the image in the target object monitoring image library can be matched according to the characteristic information of the new target object, that is, the process automatically returns to execute step S110 until a trajectory generation instruction input by the user is received, indicating that the user believes that there are sufficient target object monitoring images to generate a trajectory, and there is no need to continue selecting target object monitoring images, and then step 150 is executed.
[0083] Step S150: Generate a trajectory of the target object according to the stored monitoring images of each target object.
[0084] It can be understood that after multiple cycles of steps S110 to S140, the user selects and saves several target object monitoring images, which may have generation time and specific location. Then, the spatiotemporal trajectory of the target object can be generated based on the generation time and corresponding location of each target object monitoring image.
[0085] The method for finding a person's trajectory, provided in an embodiment of the present application, updates the target object's feature information based on each user selection. Each update of the target object's feature information updates the image set, ensuring that the target object's surveillance images in the image set obtained with each update better meet the user's needs. This continues until the user no longer selects the target object surveillance image and enters a trajectory generation command, generating the target object's trajectory. This demonstrates that, using the target object's features as an initial reference, through its structured information and the user's preferred selections, multiple rounds of screening are used to identify multiple images of the target object, resulting in a more precise search for the target object and, on this basis, a more accurate generation of the target object's trajectory.
[0086] The feature information mentioned in the above embodiments may include a plurality of target object image features and a confidence level corresponding to each target object image feature. In some embodiments of the present application, the process of fusing the feature information in the feature information set in step S130 and using the fused feature information as the new target object feature information is introduced. This process may include:
[0087] S1. For each target object image feature, sum the confidences of the target object image features in each feature information in the feature information set to obtain the total confidence of the target object image feature.
[0088] Specifically, the confidence level may represent the credibility of the corresponding target object image feature.
[0089] For example, a feature information set may include feature information A and feature information B. Feature information A may include: hair (yellow hair, 100% confidence), top (red top, 100% confidence), gender (female, 100% confidence), and glasses (wearing glasses, 100% confidence). Feature information B may include hair (yellow hair, 80% confidence, white hair, 20% confidence), top (red top, 70% confidence, pink top, 30% confidence), gender (female, 60% confidence, male, 40% confidence), glasses (wearing glasses, 100% confidence, not wearing glasses, 0% confidence), bottoms (black pants, 90% confidence, gray pants, 10% confidence), and age (teenagers, 5%, middle-aged, 85%, elderly, 10%). Then, the total confidence of the target object image feature of hair is (yellow 180%, white 20%), the total confidence of the top is (red 170%, pink 30%), the total confidence of gender is (female 160%, male 40%), the total confidence of glasses is (wearing glasses 200%, not wearing glasses 0%), the total confidence of bottoms is (black pants 90%, gray pants 10%), and the total confidence of age is (teenagers 5%, middle-aged 85%, elderly 10%).
[0090] S2. For each target object image feature, divide the total confidence by the number of target object image features whose confidence is not 0 in each feature information to obtain an average confidence of the target object image feature.
[0091] Specifically, the average confidence may represent the average likelihood of integrating the corresponding target object image features.
[0092] For example, the average confidence of the target object image feature of hair is (yellow 90%, white 10%), the average confidence of tops is (red 85%, pink 15%), the average confidence of gender is (female 80%, male 20%), the average confidence of glasses is (wearing glasses 100%, not wearing glasses 0%), the average confidence of bottoms is (black pants 90%, gray pants 10%), and the average confidence of age is (teenagers 5%, middle-aged 85%, elderly 10%).
[0093] S3. Summarize the image features of each target object and the average confidence of each target object image feature to obtain feature information of a new target object.
[0094] For example, the feature information of the new target object can be the hair and the average confidence of the hair, the top and the average confidence of the top, the gender and the average confidence of the gender, the glasses and the average confidence of the glasses, the bottoms and the average confidence of the bottoms, the age and the average confidence of the age in the example S2 of this embodiment, or a summary of the above.
[0095] The method for searching for a person's trajectory provided in the embodiment of the present application can reconstruct new feature information of the target object by calculating the comprehensive confidence of the target object image features in each feature information in the feature information set, so that the target object image features represented by the new feature information are more in line with the user's intention.
[0096] The feature information mentioned in the above embodiment may include several target object image features and the confidence level corresponding to each target object image feature. In some embodiments of the present application, considering that the feature information of the target object image in the target object monitoring image may contain many uninteresting target object image features, and these uninteresting target object image features may interfere with the matching selection, based on this, the embodiment of the present application provides another method for finding a person's trajectory. In this embodiment, after determining the feature information of the target object image in the target object monitoring image in the above step S102, a process for the user to screen the target object image features of interest can be further added. Specifically, the process may include:
[0097] In response to a user's operation of selecting a designated target object image feature from a plurality of target object image features, the designated target object image and the corresponding confidence level are used as new feature information.
[0098] It is understandable that if the feature information originally provided only contains several target object image features, such as hair, top, and gender, and the feature information of the target object image in the target object monitoring image may include other target object image features in addition to hair, top, and gender, such as bottoms and age, the user may not be interested in the additional target object image features, then only the target object image features of interest (hair, top, gender) and the corresponding confidence levels can be selected as new feature information.
[0099] It is understandable that when the user selects only the target object image features of interest as new feature information, what is added to the feature information set is not the feature information containing the target object image features of interest and non-interest, but new feature information.
[0100] Examples include Figure 2 The optional target object image features are all target object image features that determine the characteristic information of the target object image in the target object monitoring image, which can include hair, top, bottom, glasses, age, gender, and mask. Based on these target object image features, 13 images with a relatively high degree of matching can be selected from the target object monitoring image library. On this basis, the user can select only the target object image features of interest: hair, top, gender, such as Figure 3As shown, since the user selects the features of the designated target object image, the feature information is also updated accordingly, so that 8 images that are more consistent with the updated feature information are selected from the target object monitoring image library.
[0101] Based on this, the process of adding the feature information to the feature information set mentioned in step S120 of the above embodiment may specifically include:
[0102] The new feature information is added to the feature information set.
[0103] The personnel trajectory search method provided in the embodiment of the present application can provide users with the function of filtering only the target object image features of interest, thereby improving the user experience, optimizing the feature information obtained directly from the target object image in the target object monitoring image, and optimizing the matching effect of the target object monitoring image.
[0104] In some embodiments of the present application, considering the need to distinguish selected target object surveillance images from unselected target object surveillance images to avoid repeated selection by the user, an embodiment of the present application provides another method for finding a person's trajectory. In step S101 of the above embodiment, after forming an image set, a process for allowing the user to distinguish selected and unselected target object surveillance images can be further added. Specifically, the process can include the following two situations, including:
[0105] The first method is to delete the saved monitoring images of the same target object in the image set to obtain a processed image set.
[0106] Specifically, each image in the image set is compared with the saved target object monitoring images, and if there are identical images, they are deleted. Therefore, the images retained after deletion are target object monitoring images that the user has not selected, and the user can then select an image from the target object monitoring images that have not been selected.
[0107] The second method is to mark the target object monitoring image in the image set that is identical to the saved target object monitoring image, so as to inform the user that the target object monitoring image corresponding to the mark has been selected.
[0108] Specifically, each image in the image set is compared with the saved target object monitoring images, and if there are identical ones, they are marked. Therefore, the unmarked images are target object monitoring images that the user has not selected, and the user can clearly select images from the target object monitoring images that have not been selected.
[0109] Examples include Figure 4, the user selects the first target object monitoring image in the second row from the image collection, then the feature information of this target object monitoring image is determined and added to the feature information collection, and this target object monitoring image is marked with an asterisk, the feature information collection is updated and the feature information in the collection is fused to obtain the fused feature information, and then 6 images with a relatively high degree of matching can be selected from the target object monitoring image library and displayed. Among them, the image selected by the user is displayed in front, such as Figure 5 shown.
[0110] The method for searching for a person's trajectory provided in an embodiment of the present application can delete or mark images selected or saved by the user in an image collection, so that the user can clearly select images from surveillance images of target objects that have not been selected, thereby avoiding repeated selection.
[0111] The feature information mentioned in the above embodiments may include several target object image features and the confidence level corresponding to each target object image feature. In some embodiments of the present application, the process of obtaining several target object monitoring images that match the feature information of the target object mentioned in step S110 of the above embodiment is introduced. This process may include:
[0112] S1. Calculate the matching degree between each image in the target object monitoring image library and each target object image feature in the target object feature information.
[0113] Specifically, the matching degree may indicate the degree of agreement between the target object image features that can be determined from one of the images in the target object monitoring image library and the target object image features in the feature information of the target object.
[0114] For example, the target object image features and corresponding confidence levels that can be determined from one of the images in the target object monitoring image library can be (hair: yellow 80%, top: red 70%, gender: female 60%, glasses: wearing glasses 100%), and the target object image features and corresponding confidence levels of the target object's characteristic information can be (hair: yellow 100%, top: red 90%, gender: female 100%). Then, the matching degree of the target object image feature of hair is 80% / 100%=80%, the matching degree of top is 70% / 90%≈78%, and the matching degree of gender is 60% / 100%=60%.
[0115] S2. Calculate the average matching degree of the matching degrees of the image features of each target object.
[0116] Specifically, the average matching degree may represent the degree of similarity between the feature information that can be determined from one of the images in the target object monitoring image library and the feature information of the target object.
[0117] Continuing with the example of this embodiment, the matching degree of hair can be 80%, the matching degree of top can be 78%, and the matching degree of gender can be 60%. Then the average matching degree of the matching degrees of the image features of each target object is (80%+78%+60%) / 3≈72.6%.
[0118] S3. Acquire an image whose average matching degree is higher than a preset matching threshold and use it as a target object monitoring image.
[0119] Specifically, the preset matching threshold may represent the minimum average matching degree that can effectively characterize the feature information of the target object, such as 50%. If the average matching degree is not higher than 50%, it is considered that the image does not match the feature information of the target object.
[0120] Continuing with the example of this embodiment, since the average matching degree is 72.6%>50%, it is considered that the image matches the feature information of the target object and is used as the target object monitoring image.
[0121] The personnel trajectory search method provided in the embodiment of the present application can calculate the confidence of each target object image feature of each image from the target object monitoring image library, thereby obtaining the matching degree of the target object image feature with the feature information of the target object, and finally determine whether each image matches the feature information of the target object.
[0122] Based on this, in some embodiments of the present application, the process of displaying several monitoring images of the target object in the image set mentioned in the above embodiment is introduced. The process may include:
[0123] The target object monitoring images are displayed from front to back according to the average matching degree of each target object monitoring image. Among them, for target object monitoring images with the same average matching degree, the target object monitoring image generated earlier is displayed first.
[0124] Specifically, images with higher average matching degrees are displayed at the front so that users can observe them first. The target object monitoring images are displayed from front to back according to the average matching degrees from high to low.
[0125] It is understandable that when the characteristic information of the target object contains fewer target object image features, or the user only cares about a small number of target object image features, there may be multiple target object monitoring images in the target object monitoring image library with the same average matching degree with the characteristic information of the target object. In this case, these images can be sorted and displayed according to the image generation time.
[0126] For example, there may be two target object monitoring images C and D in the target object monitoring image library, so that the average matching degree with the characteristic information of the target object is 80%, and the generation time of image C is "xx / xx / xx 9:00", and the generation time of image D is "xx / xx / xx 10:00", then image C is located in front of image C in the displayed position.
[0127] The personnel trajectory search method provided in the embodiment of the present application can sort images according to the average matching degree and, when the average matching degrees are the same, according to the generation time of the images, so that the display method of the target object monitoring image is more in line with the user's attention.
[0128] The following describes an apparatus for implementing personnel trajectory search provided in an embodiment of the present application. The apparatus for implementing personnel trajectory search described below and the method for implementing personnel trajectory search described above can refer to each other.
[0129] See also Figure 6 , Figure 6 This is a schematic diagram of the structure of a device for realizing personnel trajectory search disclosed in an embodiment of the present application.
[0130] like Figure 6 As shown, the device may include:
[0131] The target object image display unit 11 is used to obtain a plurality of target object monitoring images that match the characteristic information of the target object, form an image set and display it;
[0132] a feature information determining unit 12 configured to, in response to a user selecting a target object monitoring image from an image set, determine feature information of a target object image in the target object monitoring image, add the feature information to a feature information set, and save the target object monitoring image selected by the user;
[0133] The feature information fusion unit 13 is used to fuse the feature information in the feature information set, and use the fused feature information as the feature information of the new target object, and return to the step of obtaining a plurality of target object monitoring images that match the feature information of the target object to form an image set, until receiving a trajectory generation instruction input by the user, and generating the trajectory of the target object according to the saved target object monitoring images.
[0134] Optionally, the feature information includes several target object image features and the confidence level corresponding to each target object image feature.
[0135] The feature information fusion unit 13 includes:
[0136] a confidence summing unit, configured to sum, for each target object image feature, the confidences of the target object image features in each piece of feature information in the feature information set to obtain a total confidence of the target object image feature;
[0137] a confidence averaging unit, configured to divide, for each target object image feature, the sum of the confidences by the number of target object image features whose confidences are not 0 in each feature information, to obtain an average confidence of the target object image feature;
[0138] The feature information updating unit is used to summarize the image features of each target object and the average confidence of each target object image feature to obtain new feature information of the target object.
[0139] Optionally, the feature information includes a plurality of target object image features and a confidence level corresponding to each target object image feature;
[0140] The device further comprises:
[0141] a target object image feature selection unit configured to, after determining feature information of a target object image in the target object monitoring image, respond to a user's operation of selecting a specified target object image feature from a plurality of target object image features, and use the specified target object image feature and the corresponding confidence level as new feature information;
[0142] The feature information determining unit 12 includes:
[0143] The selected information adding unit is used to add the new feature information to the feature information set.
[0144] Optionally, the device further includes:
[0145] An image reduction element is used to delete, from the image set, the monitoring images that are identical to the saved target object images, to obtain a processed image set;
[0146] or,
[0147] The image marking unit is used to mark the target object monitoring image in the image set that is the same as the saved target object monitoring image, so as to inform the user that the target object monitoring image corresponding to the mark has been selected.
[0148] Optionally, the feature information includes a plurality of target object image features and a confidence level corresponding to each target object image feature;
[0149] The target object image display unit 11 includes:
[0150] A matching degree calculation unit, used to calculate the matching degree between each image in the target object monitoring image library and each target object image feature in the target object feature information;
[0151] an average matching degree calculation unit, used to calculate the average matching degree of the matching degrees of the image features of each target object;
[0152] The target object monitoring image determining unit is configured to obtain an image with an average matching degree higher than a preset matching threshold value and use the image as the target object monitoring image.
[0153] Optionally, the target object image display unit 11 includes:
[0154] The display sorting unit is used to display the target object monitoring images from front to back according to the average matching degree of each target object monitoring image, wherein, for several target object monitoring images with the same average matching degree, the target object monitoring image generated earlier is displayed in front.
[0155] The device for searching for a person's track provided in the embodiment of the present application can be applied to a person's track searching device, such as a terminal: a mobile phone, a computer, etc. Optionally, Figure 7 The hardware structure diagram of the personnel trajectory search is shown in FIG. Figure 7 ,The hardware structure of the personnel trajectory search device may include: at least one processor 1, at least one communication interface 2, at least one memory 3 and at least one communication bus 4;
[0156] In the embodiment of the present application, the number of the processor 1, the communication interface 2, the memory 3, and the communication bus 4 is at least one, and the processor 1, the communication interface 2, and the memory 3 communicate with each other through the communication bus 4;
[0157] The processor 1 may be a central processing unit (CPU), or an application-specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of the present invention;
[0158] The memory 3 may include a high-speed RAM memory, and may also include a non-volatile memory, such as at least one disk memory;
[0159] The memory stores a program, and the processor can call the program stored in the memory, wherein the program is used to:
[0160] Acquire several monitoring images of the target object that match the characteristic information of the target object, form an image set, and display it;
[0161] In response to a user selecting a target object monitoring image from an image set, determining feature information of a target object image in the target object monitoring image, adding the feature information to a feature information set, and saving the target object monitoring image selected by the user;
[0162] The feature information in the feature information set is fused, and the fused feature information is used as the feature information of the new target object. The step of obtaining a plurality of target object monitoring images that match the feature information of the target object to form an image set is returned to, until a trajectory generation instruction input by the user is received, and the trajectory of the target object is generated according to the saved target object monitoring images.
[0163] Optionally, the detailed functions and extended functions of the program may refer to the above description.
[0164] An embodiment of the present application further provides a storage medium, which may store a program suitable for execution by a processor, wherein the program is used to:
[0165] Acquire several monitoring images of the target object that match the characteristic information of the target object, form an image set, and display it;
[0166] In response to a user selecting a target object monitoring image from an image set, determining feature information of a target object image in the target object monitoring image, adding the feature information to a feature information set, and saving the target object monitoring image selected by the user;
[0167] The feature information in the feature information set is fused, and the fused feature information is used as the feature information of the new target object. The step of obtaining a plurality of target object monitoring images that match the feature information of the target object to form an image set is returned to, until a trajectory generation instruction input by the user is received, and the trajectory of the target object is generated according to the saved target object monitoring images.
[0168] Optionally, the detailed functions and extended functions of the program may refer to the above description.
[0169] Finally, it should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or device comprising a series of elements includes not only those elements, but also other elements not explicitly listed, or elements inherent to such process, method, article, or device. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of additional identical elements in the process, method, article, or device comprising the element.
[0170] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The various embodiments can be combined as needed, and the same or similar parts can be referenced to each other.
[0171] The above description of the disclosed embodiments is intended to enable one skilled in the art to implement or use the present application. Various modifications to these embodiments will be readily apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application is not limited to the embodiments shown herein, but is intended to conform to the widest scope consistent with the principles and novel features disclosed herein.
Claims
1. A method for finding a person's trajectory, characterized in that: include: Acquire a number of monitoring images of the target object that match the characteristic information of the target object, form an image set and display it; In response to a user selecting a target object monitoring image in an image set, determining feature information of a target object image in the target object monitoring image, adding the feature information to a feature information set, and saving the target object monitoring image selected by the user; The feature information in the feature information set is fused, and the fused feature information is used as the feature information of the new target object, and the step of obtaining a plurality of target object monitoring images matching the feature information of the target object to form an image set is returned to execute, until a trajectory generation instruction input by a user is received, and the trajectory of the target object is generated according to the saved target object monitoring images; The feature information includes a plurality of target object image features and a confidence level corresponding to each target object image feature; The step of acquiring a plurality of monitoring images of the target object that match the characteristic information of the target object includes: Calculate the matching degree between each image in the target object monitoring image library and each target object image feature in the feature information of the target object, wherein, under the same image feature dimension, the matching degree is equal to the value of the lower confidence divided by the higher confidence under the image feature dimension; Calculate the average matching degree of the matching degrees of each target object image feature; An image whose average matching degree is higher than a preset matching threshold is acquired and used as a monitoring image of the target object.
2. The method according to claim 1, characterized in that The feature information includes a plurality of target object image features and a confidence level corresponding to each target object image feature; The feature information in the feature information set is merged, and the merged feature information is used as the feature information of the new target object, including: For each target object image feature, summing the confidences of the target object image features in each feature information in the feature information set to obtain the sum of the confidences of the target object image features; For each target object image feature, the sum of the confidences is divided by the number of target object image features whose confidences are not 0 in each feature information to obtain an average confidence of the target object image feature; The image features of each target object and the average confidence of each target object image feature are summarized to obtain feature information of a new target object.
3. The method according to claim 1, characterized in that The feature information includes a plurality of target object image features and a confidence level corresponding to each target object image feature; After determining the feature information of the target object image in the target object monitoring image, the method further includes: In response to a user selecting a specified target object image feature from a plurality of target object image features, the specified target object image feature and the corresponding confidence level are used as new feature information; The process of adding the characteristic information to the characteristic information set includes: The new feature information is added to the feature information set.
4. The method according to claim 1, characterized in that: After the image set is formed, the method further comprises: Deleting the same target object monitoring images as those that have been saved from the image set to obtain a processed image set; or, The target object monitoring image in the image set that is identical to the one that has been saved is marked to inform the user that the target object monitoring image corresponding to the mark has been selected.
5. The method according to claim 1, characterized in that Displaying a plurality of target object monitoring images in the image set, including: The target object monitoring images are displayed from front to back according to the average matching degree of each target object monitoring image, wherein for target object monitoring images with the same average matching degree, the target object monitoring image generated earlier is displayed first.
6. A device for finding a person's track, characterized in that: include: A target object image display unit, used to obtain a plurality of target object monitoring images matching the characteristic information of the target object, compose an image set and display it; a feature information determination unit, configured to respond to a user's operation of selecting a target object monitoring image in an image set, determine feature information of a target object image in the target object monitoring image, add the feature information to a feature information set, and save the target object monitoring image selected by the user; a feature information fusion unit, configured to fuse each feature information in the feature information set, and use the fused feature information as the feature information of the new target object, and return to execute the step of acquiring a plurality of target object monitoring images matching the feature information of the target object to form an image set, until receiving a trajectory generation instruction input by a user, and generating the trajectory of the target object according to each saved target object monitoring image; The feature information includes a plurality of target object image features and a confidence level corresponding to each target object image feature; The target object image display unit comprises: A matching degree calculation unit, used to calculate the matching degree between each image in the target object monitoring image library and each target object image feature in the feature information of the target object, wherein, under the same image feature dimension, the matching degree is equal to the value of the lower confidence divided by the higher confidence under the image feature dimension; An average matching degree calculation unit, used to calculate the average matching degree of the matching degrees of each target object image feature; The target object monitoring image determining unit is used to obtain the image whose average matching degree is higher than a preset matching threshold and use it as the target object monitoring image.
7. The device according to claim 6, characterized in that The feature information includes a plurality of target object image features and a confidence level corresponding to each target object image feature. The feature information fusion unit comprises: A confidence summing unit, configured to sum, for each target object image feature, the confidences of the target object image features in each feature information in the feature information set to obtain a sum of the confidences of the target object image features; a confidence averaging unit, for each target object image feature, dividing the sum of the confidences by the number of target object image features whose confidences are not 0 in each feature information, to obtain an average confidence of the target object image feature; The feature information updating unit is used to summarize the image features of each target object and the average confidence of each target object image feature to obtain feature information of a new target object.
8. A device for finding a person's track, characterized in that: include: Memory and processor; The memory is used to store programs; The processor is used to execute the program to implement each step of the method for searching a person's trajectory as described in any one of claims 1 to 5.
9. A readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, each step of the method for searching a person's trajectory as claimed in any one of claims 1 to 5 is implemented.