Pedestrian recognition method, device, electronic device and storage medium

Abstract

The embodiments of the present invention disclose a pedestrian recognition method, device, electronic device and storage medium. The method includes: determining a first image captured by a first image acquisition device and a second image captured simultaneously by a second image acquisition device; wherein the first image acquisition device and the second image acquisition device have a common visual field; if there are similar pedestrians in the first image and the second image, then triggering the determination of the position information to be identified and the proportion information to be identified of the similar pedestrians in the first image and the second image; if the triggered position information to be identified and the proportion information to be identified reach a preset quantity condition, then determining the position information to be identified and the proportion information to be identified as credible information. Based on the position information and proportion information of the pedestrians in the image, the accuracy of determining the similarity of the pedestrians is improved, and based on the credible information, the auxiliary judgment conditions for pedestrian re-identification can be increased, thereby improving the accuracy of pedestrian re-identification.

Description

Technical Field

[0001] Embodiments of the present invention relate to the field of image recognition technology, and in particular to a pedestrian recognition method, device, electronic device and storage medium. Background Art

[0002] For a camera group with a common visual field, pedestrians will appear in the camera group's screen at the same time. How to determine whether the target pedestrians in the camera group's screen are the same person is a current research hotspot. Pedestrian re-identification (also called pedestrian re-ID) technology can be used to determine whether there is a target pedestrian in an image or video sequence, and then determine whether the target pedestrians in the camera group's screen are the same person.

[0003] However, the accuracy of pedestrian re-identification has always been a big problem. Currently, it can only be trained through a large number of fixed pedestrian data sets, but the training of fixed pedestrian data sets only improves the accuracy of specific data sets, and the recognition accuracy is often not high enough in practical applications. Summary of the invention

[0004] Embodiments of the present invention provide a pedestrian recognition method, device, electronic device and storage medium, which improve the accuracy of determining pedestrian similarity based on the position information and proportion information of the pedestrian in the image, and increase the auxiliary judgment conditions of pedestrian re-identification based on credible information, thereby improving the accuracy of pedestrian re-identification.

[0005] In a first aspect, an embodiment of the present invention provides a pedestrian recognition method, comprising:

[0006] Determine a first image captured by a first image acquisition device and a second image captured simultaneously by a second image acquisition device; wherein the first image acquisition device and the second image acquisition device have a common visual field;

[0007] If there are similar pedestrians in the first image and the second image, triggering determination of to-be-identified position information and to-be-identified ratio information of the similar pedestrians in the first image and the second image;

[0008] If the triggered position information to be identified and the ratio information to be identified reach a preset quantity condition, the position information to be identified and the ratio information to be identified are determined to be credible information.

[0009] In a second aspect, an embodiment of the present invention further provides a pedestrian recognition device, including:

[0010] An image acquisition module, configured to determine a first image acquired by a first image acquisition device and a second image acquired simultaneously by a second image acquisition device; wherein the first image acquisition device and the second image acquisition device have a common visual field;

[0011] a similar pedestrian information determination module, configured to trigger determination of to-be-identified position information and to-be-identified ratio information of the similar pedestrian in the first image and the second image if there are similar pedestrians in the first image and the second image;

[0012] The credible information determination module is used to determine that the position information to be identified and the ratio information to be identified are credible information if the triggered position information to be identified and the ratio information to be identified reach a preset quantity condition.

[0013] In a third aspect, an embodiment of the present invention further provides an electronic device, including:

[0014] one or more processors;

[0015] a storage device for storing one or more programs,

[0016] When the one or more programs are executed by the one or more processors, the one or more processors implement the pedestrian recognition method as described in any embodiment of the present invention.

[0017] In a fourth aspect, an embodiment of the present invention further provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the pedestrian recognition method as described in any embodiment of the present invention.

[0018] The embodiment of the present invention is based on a first image captured by a first image acquisition device and a second image captured by a second image acquisition device having a common visual field. When the number of triggers for determining similar pedestrians having the same position information to be identified and the same proportion information to be identified in the first image and the second image reaches a preset number condition, the position information to be identified and the proportion information to be identified are determined as credible information. The similarity of similar pedestrians is determined based on the position information and proportion information of the pedestrians in the image based on statistical ideas, thereby improving the accuracy of determining the similarity of pedestrians. Moreover, based on credible information, auxiliary judgment conditions for pedestrian re-identification can be increased, thereby improving the accuracy of pedestrian re-identification. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 is a flow chart of a pedestrian recognition method in Embodiment 1 of the present invention;

[0020] Figure 2 A camera layout map provided by the first embodiment of the present invention;

[0021] Figure 3 This is a schematic diagram of determining the position information and height information of similar pedestrians in the image captured by Cam1 and Cam2 provided in the first embodiment of the present invention;

[0022] Figure 4is a flow chart of a pedestrian recognition method in Embodiment 2 of the present invention;

[0023] Figure 5 is a schematic diagram of the structure of a pedestrian recognition device in Embodiment 3 of the present invention;

[0024] Figure 6 It is a schematic diagram of the structure of an electronic device in Embodiment 4 of the present invention. DETAILED DESCRIPTION

[0025] The present invention will be further described in detail below in conjunction with the accompanying drawings and embodiments. It is to be understood that the specific embodiments described herein are only used to explain the present invention, rather than to limit the present invention. It should also be noted that, for ease of description, only parts related to the present invention, rather than all structures, are shown in the accompanying drawings.

[0026] Embodiment 1

[0027] Figure 1 is a flow chart of a pedestrian recognition method in Embodiment 1 of the present invention. This embodiment can be applied to determine whether pedestrians that appear simultaneously in cameras with a common visual field are the same person. The method can be executed by a pedestrian recognition device, which can be implemented in software and / or hardware and can be configured in an electronic device, such as a background server or other device with communication and computing capabilities. Figure 1 As shown, the method specifically includes:

[0028] Step 101: Determine a first image captured by a first image acquisition device and a second image captured simultaneously by a second image acquisition device; wherein the first image acquisition device and the second image acquisition device have a common visual field.

[0029] The image acquisition device is used to acquire monitoring images of pedestrians, such as a video camera or a camera set in the pedestrian activity area. The first image and the second image are acquired simultaneously, and the first image and the second image have a common visual domain, that is, when a pedestrian in the first image appears within the common visual domain, the pedestrian will also be acquired in the second image, which is conducive to determining the pedestrian's actions from various angles.

[0030] Specifically, the camera information with overlapping visual fields is determined, and the cameras with the same visual fields are marked in pairs, such as Figure 2The figure shows the layout map of the cameras. According to the visible range information of the cameras in the map, it can be obtained that Cam1 and Cam2 have a common visible range, and Cam3, Cam4 and Cam5 have a common visible range. If the common visible range of more than three cameras is determined, the visible range will be reduced. Therefore, pairwise binding is adopted, Cam1 and Cam2 are paired, Cam3 and Cam4 are paired, Cam3 and Cam5 are paired, and Cam4 and Cam5 are paired; cameras without overlapping visible ranges are not considered.

[0031] Exemplarily, taking the images captured by Cam1 and Cam2 that are successfully paired as an example, the image captured by Cam1 is determined to be the first image, the image captured by Cam2 is determined to be the second image, and the first image and the second image are captured at the same time.

[0032] Step 102: If there are similar pedestrians in the first image and the second image, trigger the determination of the to-be-identified position information and to-be-identified ratio information of the similar pedestrians in the first image and the second image.

[0033] Among them, similar pedestrians refer to targets that may be the same person. For example, if there is a pedestrian target in each of the two images, judging them as similar pedestrians means that the two pedestrian targets may be the same person. Similar pedestrians can be determined based on a pedestrian re-identification algorithm, that is, after the target pedestrian is determined in the first image acquisition device, pedestrians with a certain degree of similarity are determined from the second image acquisition device based on the characteristics of the target pedestrian. Alternatively, similar pedestrians can be determined based on other features of pedestrians, such as facial features, height features, clothing features, etc., and the method for determining similar pedestrians is not limited here.

[0034] The position information is used to characterize the screen positions of similar pedestrians in each image, and the ratio information is used to characterize the height characteristics between similar pedestrians.

[0035] If there are similar pedestrians in the first image and the second image, that is, pedestrians with a certain degree of similarity appearing simultaneously in the pair of cameras are similar pedestrians, then the position information and proportion information of the pedestrian in the first image and the second image are triggered to be recorded, and the information is stored in the database and determined as the position information to be identified and the proportion information to be identified, so that the reliable information can be determined later based on the trigger after other similar pedestrians are determined.

[0036] Exemplarily, based on the above example, taking the images captured by the successfully paired Cam1 and Cam2 as an example, the first pedestrian target is determined in Cam1, and the features of the first pedestrian target are used as a template to match it in the image captured simultaneously by Cam2. If it is determined that there is a second pedestrian target with a certain degree of similarity to the first pedestrian target, then the first pedestrian target and the second pedestrian target are similar pedestrians, and the first position information and first height information of the first pedestrian target in the first image, as well as the second position information and second height information of the second pedestrian target in the second image are determined, and the proportion information to be identified is determined based on the ratio of the first height information to the second height information, and the first position information, the second position information and the proportion information to be identified are stored in the database.

[0037] In a feasible embodiment, triggering to determine the to-be-identified position information and to-be-identified ratio information of similar pedestrians in the first image and the second image includes:

[0038] Determine the position information of the reference point of the similar pedestrian in the first image and the second image as the position information to be identified; wherein the reference point includes the lowest point or the highest point of the complete human body frame of the similar pedestrian in the image;

[0039] A first height of a complete human frame of a similar pedestrian in the first image and a second height of a complete human frame in the second image are determined, and the ratio information to be identified is determined according to the first height and the second height.

[0040] If a similar pedestrian is not fully displayed in the image, it will cause inaccurate determination of pedestrian features based on height ratio. Therefore, the similar pedestrian needs to have a complete body frame in the image. If not, the position information and ratio information of the similar pedestrian will not be triggered.

[0041] Exemplarily, after similar pedestrians are determined, the human body frames of the similar pedestrians in each picture are determined and the integrity of the human body frames is determined. If they are complete, it triggers the determination of the position information to be identified and the proportion information to be identified of the similar pedestrians in the first image and the second image; if they are incomplete, it does not trigger and the identification of similar pedestrians continues.

[0042] When triggering the determination of the to-be-identified position information and to-be-identified proportion information of a similar pedestrian in the first image and the second image, a complete human body frame in which the human body is located is determined in the first image and the second image, and the lowest point or the highest point of the complete human body frame is determined as a reference point. The position of the reference point in the image screen is used as the position of the similar pedestrian in the image screen, and the first height and the second height of the complete human body frame in the first image and the second image are determined, and the ratio of the first height to the second height is the proportion information to be identified.

[0043] For example, based on the above example, take the image captured by Cam1 and Cam2 which are successfully paired as an example. Figure 3 The figure shows a schematic diagram of determining the position information and height information of similar pedestrians in the image captured by Cam1 and Cam2. According to different position accuracy requirements, the video screen captured by the camera can be divided into 100 equal parts, and the position information of the upper left corner of the image screen is (0, 0), and the position information of the lower right corner is (100, 100). The setting of the equal division ratio can be modified according to actual needs, and there is no restriction here.

[0044] When a first pedestrian target is detected in the first image captured by Cam1, a comparison is simultaneously made to see whether there is a second pedestrian target whose similarity reaches a threshold in the second image captured by Cam2. If there is a second pedestrian target, the position information and the proportion information are recorded.

[0045] A human body is selected in the first image captured by Cam1, and the lowest point of the human body is selected as the reference point. The proportional position of the reference point in the picture is measured, and L1 and W1, that is, the horizontal proportional coordinate and the vertical proportional coordinate of the first pedestrian target, are recorded. The first position information is determined by L1 and W1, and the first height H1 of the human body selection is recorded; similarly, a human body is selected in the second image captured by Cam2, and the proportional position of the reference point in the picture is measured, and L2 and W2 are recorded. The second position information is determined by L2 and W2, and the second height H2 of the human body selection is recorded; a position data table is generated for each group of similar pedestrians, and the position information (L1, W1), (L2, W2) and the proportional information H1 / H2 are recorded in the data table. The recording results are shown in Table 1.

[0046] Table 1

[0047] First location information Second location information Scale information quantity (L1, W1) (L2, W2) H1 / H2 1

[0048] Step 103: If the triggered position information to be identified and the ratio information to be identified reach a preset quantity condition, the position information to be identified and the ratio information to be identified are determined to be credible information.

[0049] During the process of continuous image acquisition by the first image acquisition device and the second image acquisition device, a large number of similar pedestrians will be identified, and a large amount of position information to be identified and ratio information to be identified will be triggered. If the number of occurrences of the target position information to be identified and the target ratio information to be identified reaches a preset number condition, the target position information to be identified and the target ratio information to be identified are determined to be credible information.

[0050] Exemplarily, based on the above example, the position information to be identified and the proportion information to be identified of similar pedestrians are stored in the database. When a new similar pedestrian appears, triggering the acquisition of new position information to be identified and the proportion information to be identified, it is determined whether there is a consistent record of the position information to be identified and the proportion information to be identified in the database. If so, the quantity value of the record is increased by one. If not, a new data record is created in the database for the position information to be identified and the proportion information to be identified. When the accumulated quantity value of the record reaches a preset number, the data in the record is considered to be credible data. For example, when the accumulated quantity value reaches 10, that is, the position information to be identified and the proportion information to be identified appear ten times, then the identified position information and the proportion information to be identified are credible information. Exemplarily, the credible information is added to the auxiliary information table to achieve credible information in the auxiliary information table.

[0051] Therefore, if the target pedestrian is detected in both the first image acquisition device and the second image acquisition device, and the position information and proportion information of the target pedestrian in the first image and the second image are consistent with the credible information, it means that the target pedestrian is the same person. That is, the credible information improves the accuracy of human body recognition from a statistical perspective. In addition, the recognition accuracy can be adjusted by adjusting the preset quantity conditions.

[0052] The embodiment of the present invention is based on a first image captured by a first image acquisition device and a second image captured by a second image acquisition device having a common visual field. When the number of triggers for determining similar pedestrians having the same position information to be identified and the same proportion information to be identified in the first image and the second image reaches a preset number condition, the position information to be identified and the proportion information to be identified are determined as credible information. The similarity of similar pedestrians is determined based on the position information and proportion information of the pedestrians in the image based on statistical ideas, thereby improving the accuracy of determining the similarity of pedestrians. Moreover, based on credible information, auxiliary judgment conditions for pedestrian re-identification can be increased, thereby improving the accuracy of pedestrian re-identification.

[0053] Embodiment 2

[0054] Figure 4 is a flow chart of a pedestrian recognition method in Embodiment 2 of the present invention. Embodiment 2 is further optimized on the basis of Embodiment 1. Figure 4 As shown, the method includes:

[0055] Step 401: Determine a first image captured by a first image acquisition device and a second image captured simultaneously by a second image acquisition device; wherein the first image acquisition device and the second image acquisition device have a common visual field.

[0056] Step 402: Determine the similarity of pedestrians in the first image and the second image through a pedestrian re-identification model; and determine similar pedestrians based on the pedestrian similarity.

[0057] The pedestrian re-identification model refers to a model for re-identifying pedestrians based on a deep learning algorithm, and the specific model architecture is not limited in the embodiment of the present invention. Exemplarily, the pedestrian re-identification model is a convolutional neural network model.

[0058] Specifically, a target pedestrian in the first image is determined, and the similarity between each pedestrian target and the target pedestrian in the second image is determined by a preset pedestrian re-identification model. If the similarity reaches a similarity threshold, the pedestrian is determined to be a similar pedestrian.

[0059] Step 403: If there are similar pedestrians in the first image and the second image, trigger the determination of the to-be-identified position information and to-be-identified ratio information of the similar pedestrians in the first image and the second image.

[0060] Step 404: If the triggered position information to be identified and the ratio information to be identified reach a preset quantity condition, the position information to be identified and the ratio information to be identified are determined to be credible information.

[0061] In a feasible embodiment, the method further includes: identifying pedestrians in the first image and the second image based on the credible information to determine credible similar pedestrians.

[0062] After determining the credible information, the pedestrians in the first image and the second image can be identified based on the credible information. When the position information and proportion information of the target pedestrian appearing in both the first image and the second image are consistent with the credible information, it can be considered that the target pedestrian in the first image and the second image are credible similar pedestrians, that is, the same person.

[0063] In a feasible embodiment, pedestrians in the first image and the second image are identified according to the credible information to determine credible similar pedestrians, including:

[0064] If the to-be-identified position information and to-be-identified ratio information of the similar pedestrian in the first image and the second image successfully match the credible information, the similar pedestrian is determined to be a credible similar pedestrian.

[0065] Exemplarily, when a human body is detected in the images of Cam1 and Cam2, the first position information and the second position information collected by Cam1 and Cam2 are matched in the auxiliary information table. If the match is successful, that is, the current first position information and the second position information are the same as the first position information and the second position information in the target trusted information in the auxiliary information table, and the proportion information is also consistent, it is determined that it is a trusted similar pedestrian, that is, the human body in Cam1 and Cam2 is the same person.

[0066] The embodiment of the present invention first needs to determine the camera information with a common visual field, and mark them in pairs. If pedestrians with a certain similarity appear in the pair of cameras at the same time, the position of the pedestrian in the picture and the height ratio of the framed human body are recorded, and the information is recorded in the database. If the position information and ratio information of pedestrians with a certain similarity appear again, the record count of the position information and ratio information is increased by one. When the cumulative count reaches a certain threshold, the position information and ratio information are used in actual scenes to realize auxiliary judgment of pedestrian re-identification. When a pedestrian with the same position and the same ratio appears again, even if the similarity does not reach the set threshold, it is considered to be the same person, thereby increasing the accuracy and data volume of pedestrian re-identification.

[0067] Step 405: Update and train the pedestrian re-identification model based on the credible information.

[0068] Since a credible similar pedestrian can be obtained through credible information, the pedestrian re-identification model is updated and trained based on the credible similar pedestrian. Exemplarily, after the credible similar pedestrian is determined through credible information, the first image and the second image corresponding to the credible similar pedestrian are used as training samples to update the pedestrian re-identification model to adjust the parameters of the pedestrian re-identification model.

[0069] Since the probability that the credible similar pedestrians determined by credible information are the same person is relatively high, the first image and the second image corresponding to the feasible similar pedestrians are used as training samples, which not only increases the sample size of the training set, but also ensures the accuracy of the training samples, thereby improving the accuracy of the model update training.

[0070] Exemplarily, if a non-similar pedestrian whose pedestrian similarity is less than a similarity threshold appears in the first image and the second image, but the position information and proportion information of the non-similar pedestrian are consistent with the trusted information, the first image and the second image of the non-similar pedestrian are input into the pedestrian re-identification model for training and learning, and the similarity calculation result of the sample is improved by adjusting the model parameters, and then the accuracy of the pedestrian re-identification model is improved through repeated training and learning.

[0071] In a feasible embodiment, step 405 includes:

[0072] Determining credible similar pedestrians in the first image and the second image based on the credible information;

[0073] If the pedestrian similarity of the credible similar pedestrian is less than the similarity threshold, the pedestrian re-identification model is updated and trained based on the credible similar pedestrian.

[0074] Specifically, after the credible information is determined, the pedestrians in the first image and the second image can be identified based on the credible information. When the position information and proportion information of the target pedestrian in both the first image and the second image are consistent with the credible information, the target pedestrian in the first image and the second image can be considered to be a credible similar pedestrian, that is, the same person. After the credible similar pedestrian is determined based on the credible information, the similarity of the credible similar pedestrian is determined based on the pedestrian re-identification model. If the similarity is greater than the similarity threshold, it means that the detection results of the two are consistent, and there is no need to update the training of the pedestrian re-identification model; if the pedestrian similarity is less than the similarity threshold, a positive sample data set is established based on the credible similar pedestrian, and the pedestrian re-identification model is updated and trained based on the positive sample data set to adjust the model parameters and improve the accuracy of the model.

[0075] In the embodiment of the present invention, the pedestrian re-identification model is continuously trained in actual combat, and the location information and proportion information data of a large number of similar pedestrians are used to determine the credible information. The credible similar pedestrians obtained by the credible information are used to train the pedestrian re-identification model, and the trained model is applied again in actual combat. This process is repeated to improve the accuracy of human body re-identification.

[0076] The embodiments of the present invention utilize pedestrian recognition data collected by a large number of cameras with a common visual field to assist in determining whether similar pedestrians are the same person. This can be used to adjust the parameters of a pedestrian re-identification model for training, and also realizes the addition of determination conditions for similar pedestrians in actual combat, thereby improving the accuracy of pedestrian recognition.

[0077] Embodiment 3

[0078] Figure 5 is a schematic diagram of the structure of a pedestrian recognition device in Embodiment 3 of the present invention. This embodiment is applicable to determining whether pedestrians that appear simultaneously in cameras with a common visual field are the same person. Figure 5 As shown, the device comprises:

[0079] An image acquisition module 510, configured to determine a first image acquired by a first image acquisition device and a second image acquired simultaneously by a second image acquisition device; wherein the first image acquisition device and the second image acquisition device have a common visual field;

[0080] A similar pedestrian information determination module 520, configured to trigger the determination of to-be-identified position information and to-be-identified ratio information of the similar pedestrian in the first image and the second image if there are similar pedestrians in the first image and the second image;

[0081] The credible information determination module 530 is used to determine that the position information to be identified and the ratio information to be identified are credible information if the triggered position information to be identified and the ratio information to be identified reach a preset quantity condition.

[0082] The embodiment of the present invention is based on a first image captured by a first image acquisition device and a second image captured by a second image acquisition device having a common visual field. When the number of triggers for determining similar pedestrians having the same position information to be identified and the same proportion information to be identified in the first image and the second image reaches a preset number condition, the position information to be identified and the proportion information to be identified are determined as credible information. The similarity of similar pedestrians is determined based on the position information and proportion information of the pedestrians in the image based on statistical ideas, thereby improving the accuracy of determining the similarity of pedestrians. Moreover, based on credible information, auxiliary judgment conditions for pedestrian re-identification can be increased, thereby improving the accuracy of pedestrian re-identification.

[0083] Optionally, the similar pedestrian information determination module 520 is specifically configured to:

[0084] Determine the position information of the reference point of the similar pedestrian in the first image and the second image as the position information to be identified; wherein the reference point includes the lowest point or the highest point of the complete human body frame of the similar pedestrian in the image;

[0085] Determine a first height of a complete human frame of the similar pedestrian in the first image and a second height of a complete human frame in the second image, and determine the proportion information to be identified according to the first height and the second height.

[0086] Optionally, the device further includes a credible similar pedestrian determination module, which is used to:

[0087] Pedestrians in the first image and the second image are identified according to the credible information to determine credible similar pedestrians.

[0088] Optional, trusted similar pedestrian determination module, specifically used for:

[0089] If the to-be-identified position information and to-be-identified proportion information of similar pedestrians in the first image and the second image successfully match the credible information, the similar pedestrians are determined to be credible similar pedestrians.

[0090] Optionally, the device further comprises a model-based similar pedestrian determination module, which is used to:

[0091] Determine the similarity of pedestrians in the first image and the second image by using a pedestrian re-identification model;

[0092] Similar pedestrians are determined according to the pedestrian similarities.

[0093] Optionally, the device further includes a model updating module, configured to:

[0094] The pedestrian re-identification model is updated and trained based on the trusted information.

[0095] Optional, model update module, specifically used to:

[0096] Determining credible similar pedestrians in the first image and the second image according to the credible information;

[0097] If the pedestrian similarity of the credible similar pedestrian is less than a similarity threshold, the pedestrian re-identification model is updated and trained based on the credible similar pedestrian.

[0098] The pedestrian recognition device provided in the embodiment of the present invention can execute the pedestrian recognition method provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects for executing the pedestrian recognition method.

[0099] Embodiment 4

[0100] Figure 6 It is a structural schematic diagram of an electronic device provided in Embodiment 4 of the present invention. Figure 6 A block diagram of an exemplary electronic device 12 suitable for use in implementing embodiments of the present invention is shown. Figure 6 The electronic device 12 shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present invention.

[0101] like Figure 6 As shown, the electronic device 12 is in the form of a general purpose computing device. The components of the electronic device 12 may include, but are not limited to: one or more processors or processing units 16, system memory 28, and a bus 18 connecting various system components (including the system memory 28 and the processing unit 16).

[0102] Bus 18 represents one or more of several types of bus structures, including a storage device bus or storage device controller, a peripheral bus, an accelerated graphics port, a processor or a local bus using any of a variety of bus architectures. By way of example, these architectures include, but are not limited to, an Industry Standard Architecture (ISA) bus, a Micro Channel Architecture (MAC) bus, an Enhanced ISA bus, a Video Electronics Standards Association (VESA) local bus, and a Peripheral Component Interconnect (PCI) bus.

[0103] The electronic device 12 typically includes a variety of computer system readable media. These media can be any available media that can be accessed by the electronic device 12, including volatile and non-volatile media, removable and non-removable media.

[0104] The system storage device 28 may include computer system readable media in the form of volatile storage devices, such as random access memory devices (RAM) 30 and / or cache memory devices 32. The electronic device 12 may further include other removable / non-removable, volatile / non-volatile computer system storage media. By way of example only, the storage system 34 may be used to read and write non-removable, non-volatile magnetic media ( Figure 6 not shown, usually called a "hard drive"). Although Figure 6 Not shown in the figure, a disk drive for reading and writing to a removable non-volatile disk (e.g., a "floppy disk"), and an optical disk drive for reading and writing to a removable non-volatile optical disk (e.g., a CD-ROM, a DVD-ROM, or other optical media) may be provided. In these cases, each drive may be connected to the bus 18 via one or more data medium interfaces. The storage device 28 may include at least one program product having a set (e.g., at least one) of program modules that are configured to perform the functions of the various embodiments of the present invention.

[0105] A program / utility 40 having a set (at least one) of program modules 42 may be stored, for example, in the storage device 28, such program modules 42 including, but not limited to, an operating system, one or more application programs, other program modules, and program data, each of which or some combination may include an implementation of a network environment. The program modules 42 generally perform the functions and / or methods of the embodiments described herein.

[0106] The electronic device 12 may also communicate with one or more external devices 14 (e.g., keyboards, pointing devices, displays 24, etc.), one or more devices that enable a user to interact with the device 12, and / or any device that enables the device 12 to communicate with one or more other computing devices (e.g., network cards, modems, etc.). Such communication may be performed via an input / output (I / O) interface 22. Furthermore, the electronic device 12 may also communicate with one or more networks (e.g., local area networks (LANs), wide area networks (WANs), and / or public networks, such as the Internet) via a network adapter 20. Figure 6 As shown, the network adapter 20 communicates with other modules of the electronic device 12 via the bus 18. It should be understood that although Figure 6 Not shown, other hardware and / or software modules may be used in conjunction with the electronic device 12, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage systems.

[0107] The processing unit 16 executes various functional applications and data processing by running the programs stored in the system storage device 28, for example, implementing the pedestrian recognition method provided in the embodiment of the present invention, including:

[0108] Determine a first image captured by a first image acquisition device and a second image captured simultaneously by a second image acquisition device; wherein the first image acquisition device and the second image acquisition device have a common visual field;

[0109] If there are similar pedestrians in the first image and the second image, triggering determination of to-be-identified position information and to-be-identified ratio information of the similar pedestrians in the first image and the second image;

[0110] If the triggered position information to be identified and the ratio information to be identified reach a preset quantity condition, the position information to be identified and the ratio information to be identified are determined to be credible information.

[0111] Embodiment 5

[0112] Embodiment 5 of the present invention further provides a computer-readable storage medium, on which a computer program is stored. When the program is executed by a processor, the pedestrian recognition method provided in the embodiment of the present invention is implemented, including:

[0113] Determine a first image captured by a first image acquisition device and a second image captured simultaneously by a second image acquisition device; wherein the first image acquisition device and the second image acquisition device have a common visual field;

[0114] If there are similar pedestrians in the first image and the second image, triggering determination of to-be-identified position information and to-be-identified ratio information of the similar pedestrians in the first image and the second image;

[0115] If the triggered position information to be identified and the ratio information to be identified reach a preset quantity condition, the position information to be identified and the ratio information to be identified are determined to be credible information.

[0116] The computer storage medium of the embodiment of the present invention can adopt any combination of one or more computer-readable media. The computer-readable medium can be a computer-readable signal medium or a computer-readable storage medium. The computer-readable storage medium can be, for example, but not limited to, a system, device or device of electricity, magnetism, light, electromagnetic, infrared, or semiconductor, or any combination of the above. More specific examples (non-exhaustive list) of computer-readable storage media include: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In this document, a computer-readable storage medium can be any tangible medium containing or storing a program, which can be used by an instruction execution system, a device or a device or used in combination with it.

[0117] Computer-readable signal media may include data signals propagated in baseband or as part of a carrier wave, which carry computer-readable program code. Such propagated data signals may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. Computer-readable signal media may also be any computer-readable medium other than a computer-readable storage medium, which may send, propagate, or transmit a program for use by or in conjunction with an instruction execution system, apparatus, or device.

[0118] The program code embodied on the computer readable medium may be transmitted using any appropriate medium, including but not limited to wireless, wireline, optical fiber cable, RF, etc., or any suitable combination of the foregoing.

[0119] Computer program code for performing the operation of the present invention may be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, C++, and conventional procedural programming languages ​​such as "C" or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, as a separate software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0120] Note that the above are only preferred embodiments of the present invention and the technical principles used. Those skilled in the art will understand that the present invention is not limited to the specific embodiments described herein, and that various obvious changes, readjustments and substitutions can be made by those skilled in the art without departing from the scope of protection of the present invention. Therefore, although the present invention has been described in more detail through the above embodiments, the present invention is not limited to the above embodiments, and may include more other equivalent embodiments without departing from the concept of the present invention, and the scope of the present invention is determined by the scope of the appended claims.

Claims

1. A pedestrian recognition method, characterized in that: include: Determine a first image captured by a first image acquisition device and a second image captured simultaneously by a second image acquisition device; wherein the first image acquisition device and the second image acquisition device have a common visual field; If there are similar pedestrians in the first image and the second image, triggering determination of to-be-identified position information and to-be-identified ratio information of the similar pedestrians in the first image and the second image; If the triggered position information to be identified and the ratio information to be identified reach a preset quantity condition, the position information to be identified and the ratio information to be identified are determined to be credible information; The triggering and determining the to-be-recognized position information and to-be-recognized ratio information of the similar pedestrian in the first image and the second image includes: Determine the position information of the reference point of the similar pedestrian in the first image and the second image as the position information to be identified; wherein the reference point includes the lowest point or the highest point of the complete human body frame of the similar pedestrian in the image; Determine a first height of a complete human frame of the similar pedestrian in the first image and a second height of a complete human frame in the second image, and determine the proportion information to be identified according to the first height and the second height.

2. The method according to claim 1, characterized in that The method further includes: identifying pedestrians in the first image and the second image according to the credible information to determine credible similar pedestrians.

3. The method according to claim 2, characterized in that Identifying pedestrians in the first image and the second image according to the credible information to determine credible similar pedestrians includes: If the to-be-identified position information and to-be-identified proportion information of similar pedestrians in the first image and the second image successfully match the credible information, the similar pedestrians are determined to be credible similar pedestrians.

4. The method according to any one of claims 1 to 3, characterized in that: The method further comprises: Determine the similarity of pedestrians in the first image and the second image by using a pedestrian re-identification model; Similar pedestrians are determined according to the pedestrian similarities.

5. The method according to claim 4, characterized in that The method further comprises: The pedestrian re-identification model is updated and trained based on the trusted information.

6. The method according to claim 5, characterized in that The method further comprises: updating and training the pedestrian re-identification model based on the credible information, comprising: Determining credible similar pedestrians in the first image and the second image according to the credible information; If the pedestrian similarity of the credible similar pedestrian is less than a similarity threshold, the pedestrian re-identification model is updated and trained based on the credible similar pedestrian.

7. A pedestrian recognition device, characterized in that: include: An image acquisition module, configured to determine a first image acquired by a first image acquisition device and a second image acquired simultaneously by a second image acquisition device; wherein the first image acquisition device and the second image acquisition device have a common visual field; a similar pedestrian information determination module, configured to trigger determination of to-be-identified position information and to-be-identified ratio information of the similar pedestrian in the first image and the second image if there are similar pedestrians in the first image and the second image; A credible information determination module, configured to determine that the position information to be identified and the ratio information to be identified are credible information if the triggered position information to be identified and the ratio information to be identified reach a preset quantity condition; The similar pedestrian information determination module is specifically used for: Determine the position information of the reference point of the similar pedestrian in the first image and the second image as the position information to be identified; wherein the reference point is the lowest point or the highest point of the complete human body frame of the similar pedestrian in the image; Determine a first height of a complete human frame of the similar pedestrian in the first image and a second height of a complete human frame in the second image, and determine the proportion information to be identified according to the first height and the second height.

8. An electronic device, characterized in that: include: one or more processors; a storage device for storing one or more programs, When the one or more programs are executed by the one or more processors, the one or more processors implement the pedestrian recognition method as described in any one of claims 1-6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the pedestrian recognition method as described in any one of claims 1 to 6 is implemented.

Images