Animal identification method, system, device and storage medium based on tourism images

Abstract

The present invention provides an animal identification method, system, device and storage medium based on tourist images, the method comprising: classifying the animals in the image through a trained animal classification network to obtain the animal information of the image; and comparing the animal information of the image with a preset animal set, executing the operation information corresponding to the animal species hit by the animal information, the animal set including multiple animal species, and each animal species corresponds to at least one operation information. The present invention can realize a mechanism for automatic identification of open scenes, reduce labor costs, improve efficiency, and, by using detection and classification, can identify the animal species in the image and perform corresponding operations.

Description

Background Art

[0002] In the current era of big data, everyone has a mobile phone or camera, and more and more pictures are taken. Ctrip's photo library also has more pictures. How to organize and apply so many pictures has become a problem. Pictures are of various shapes and sizes, involving various scenes and angles, with uneven quality and huge numbers. Humans do not have so much energy and resources to organize these pictures. In recent years, the rise of deep learning has replaced artificial intelligence with intelligence. Most machine vision is to serve portrait applications, and there are very few solutions for identifying animals.

[0003] Therefore, the present invention provides an animal identification method, system, device and storage medium based on tourist images. Summary of the invention

[0004] In response to the problems in the prior art, the purpose of the present invention is to provide an animal recognition method, system, device and storage medium based on tourist images, which overcomes the difficulties of the prior art, can realize a mechanism for automatic recognition of open scenes, reduce labor costs, improve efficiency, and, through detection and classification, can identify the animal species in the image and perform corresponding operations.

[0005] An embodiment of the present invention provides an animal recognition method based on tourist images, comprising the following steps:

[0006] S110, classifying the image into animals using a trained animal classification network to obtain animal information of the image; and

[0007] S120, comparing the animal information of the image with a preset animal set, and executing operation information corresponding to the animal species matched by the animal information, wherein the animal set includes multiple animal species, and each animal species corresponds to at least one operation information.

[0008] Preferably, the step S110 includes the following steps:

[0009] S111, mining online pictures;

[0010] S112, establishing a target detection model, performing image recognition on the image, and when the probability of containing an animal image is greater than a preset threshold, outputting at least one detection frame containing an animal image;

[0011] S113. Classify and label the animals in the frame using an animal classification model.

[0012] Preferably, the step S113 includes the following steps:

[0013] S1131, unify the training set images into a preset image size, and preprocess the images, including rotation, horizontal position translation, vertical position translation, projection, scaling, and horizontal flipping;

[0014] S1132. Build the corresponding deep convolutional neural network. The basic model selects the xception deep learning model, including two fully connected layers, and adds a bn layer and a dropout layer during training; the dropout layer randomly discards 50% of the nodes to prevent the model from overfitting. The bn layer is added to unify the mean and variance of the pictures in the same batch to speed up the training and improve the model accuracy.

[0015] S1133. Input the image in the detection frame into the classification model, and rotate, translate, scale, and flip the image. The model pre-training adopts the imagenet pre-training model, freezes the training layer, and only trains the subsequent fully connected layer. The optimization algorithm adopts the RMSProp optimization algorithm to output the animal information of the image.

[0016] Preferably, the preset image size is 299 pixels×299 pixels.

[0017] Preferably, the step S110 further includes the following steps:

[0018] S114, obtaining positioning information when the image is generated;

[0019] In the step S120, when the animal information matches the operation information corresponding to the animal species, the operation information is preset according to the animal species and the positioning information of the image.

[0020] Preferably, the step S120 further comprises providing a real-time updated regional electronic map, wherein the regional electronic map collects images of all image acquisition devices within a range and positioning information of the image acquisition devices when generating images.

[0021] Preferably, the animal collection includes a list of dangerous animals and a preset safety radius. When an image of a dangerous animal is obtained along with the electronic map of the area, a warning range is generated based on the positioning information of the image acquisition device that obtained the dangerous animal and the preset safety radius, and the positioning information of the image of the dangerous animal is used to send a warning message to all mobile terminals within the warning range.

[0022] Preferably, an evacuation path is obtained by obtaining the shortest distance path out of the warning range based on the current positioning information of the mobile terminal within the warning range, and the warning information includes the evacuation path.

[0023] Preferably, the mobile terminal inputs the target animal based on the regional electronic map, and the regional electronic map generates a tour route according to the positioning information of the mobile terminal and the positioning information of the closest photographed target animal.

[0024] An embodiment of the present invention further provides an animal identification system based on tourist images, which is used to implement the above-mentioned animal identification method based on tourist images. The animal identification system based on tourist images includes:

[0025] An animal classification module, which classifies the image into animals through a trained animal classification network to obtain the animal information of the image; and

[0026] The operation information module compares the animal information of the image with a preset animal set, and executes the operation information corresponding to the animal species hit by the animal information, wherein the animal set includes multiple animal species, and each animal species corresponds to at least one operation information.

[0027] An embodiment of the present invention further provides an animal identification device based on tourist images, comprising:

[0028] processor;

[0029] a memory storing executable instructions of the processor;

[0030] Wherein, the processor is configured to perform the steps of the above-mentioned animal identification method based on tourist images by executing the executable instructions.

[0031] An embodiment of the present invention further provides a computer-readable storage medium for storing a program, which, when executed, implements the steps of the above-mentioned animal identification method based on tourist images.

[0032] The purpose of the present invention is to provide an animal recognition method, system, device and storage medium based on tourist images, which can realize a mechanism for automatic recognition of open scenes, reduce labor costs, improve efficiency, and, by detection and classification, can identify the animal species in the image and perform corresponding operations. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] Other features, objectives and advantages of the present invention will become more apparent from a reading of the detailed description of non-limiting embodiments made with reference to the following accompanying drawings.

[0034] Figure 1 It is a flow chart of the animal identification method based on tourist images of the present invention.

[0035] Figures 2 to 6 It is a schematic diagram of the process of implementing the animal identification method based on tourist images of the present invention.

[0036] Figure 7 It is a module schematic diagram of the animal identification system based on tourist images of the present invention.

[0037] Figure 8 It is a structural schematic diagram of the animal identification device based on tourist images of the present invention.

[0038] Fig. 9 It is a schematic diagram of the structure of a computer-readable storage medium according to an embodiment of the present invention. DETAILED DESCRIPTION

[0039] Example embodiments will now be described more fully with reference to the accompanying drawings. However, example embodiments can be implemented in a variety of forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided so that the present invention will be comprehensive and complete and fully convey the concepts of the example embodiments to those skilled in the art. The same reference numerals in the figures represent the same or similar structures, and thus their repeated description will be omitted.

[0040] Figure 1 Flowchart of the animal identification method based on tourist images of the present invention. Figure 1 As shown, an embodiment of the present invention provides an animal recognition method based on tourist images, comprising the following steps:

[0041] S110, classifying the image into animals using a trained animal classification network to obtain animal information of the image. And

[0042] S120, comparing the animal information of the image with a preset animal set, and executing the operation information corresponding to the animal type matched by the animal information, wherein the animal set includes multiple animal types, and each animal type corresponds to at least one operation information.

[0043] The problem to be solved by the present invention is: in view of the current situation that a large number of pictures are unlabeled, an algorithm model is established to filter out a large number of non-animal pictures, identify pictures containing animals, and determine whether they are specific animals, so as to facilitate the subsequent processing of other models.

[0044] In a preferred embodiment, step S110 includes the following steps:

[0045] S111. Mining online pictures.

[0046] S112, establishing a target detection model, performing image recognition on the image, and when the probability of containing an animal image is greater than a preset threshold, outputting at least one detection frame containing an animal image.

[0047] S113. Classify and label the animals in the frame using an animal classification model.

[0048] In a preferred embodiment, step S113 includes the following steps:

[0049] S1131, unify the training set images into a preset image size, and preprocess the images, including rotation, horizontal position translation, vertical position translation, projection, scaling, and horizontal flipping.

[0050] S1132. Build the corresponding deep convolutional neural network. The basic model selects the xception deep learning model, including two fully connected layers, and adds the bn layer and dropout layer during training. The dropout layer randomly discards 50% of the nodes to prevent the model from overfitting. The bn layer is added to speed up the training and improve the model accuracy by unifying the mean and variance of the pictures in the same batch.

[0051] S1133, input the image in the detection frame into the classification model, and perform rotation, translation, scaling, and flipping on the image. The model pre-training adopts the imagenet pre-training model, freezes the training layer, and only trains the subsequent fully connected layer. The optimization algorithm adopts the RMSProp optimization algorithm to output the animal information of the image.

[0052] In a preferred embodiment, the preset image size is 299 pixels×299 pixels.

[0053] In a preferred embodiment, step S110 further includes the following steps:

[0054] S114: Obtain positioning information when the image is generated.

[0055] In step S120, when the animal information matches the operation information corresponding to the animal species, the operation information is preset according to the animal species and the positioning information of the image.

[0056] In a preferred embodiment, step S120 also includes providing a real-time updated regional electronic map, which collects images of all image acquisition devices within the range and positioning information of the image acquisition devices when generating images.

[0057] In a preferred embodiment, the animal collection includes a list of dangerous animals and a preset safety radius. When images of dangerous animals are obtained along with an electronic map of the area, a warning range is generated based on the positioning information of a device that obtains the image of the dangerous animal and the preset safety radius, and the positioning information of the image of the dangerous animal is used as the center to send a warning message to all mobile terminals within the warning range.

[0058] In a preferred embodiment, the evacuation path is obtained based on the current positioning information of the mobile terminal within the warning range to obtain the shortest distance path out of the warning range, and the warning information includes the evacuation path.

[0059] In a preferred embodiment, the mobile terminal inputs the target animal based on the regional electronic map, and the regional electronic map generates a tour route according to the positioning information of the mobile terminal and the positioning information of the closest photographed target animal.

[0060] like Figures 2 to 6 As shown, an example is given to illustrate the implementation process of the animal recognition method based on tourist images of the present invention. Figures 2 to 6 As shown, the animal identification method based on tourist images of the present invention comprises the following steps:

[0061] like Figure 2 As shown, in a national zoo, multiple users 30, 31, 32, 33, 34, 35, 36, 37, 38 are playing in a wildlife park, where Figure 3 As shown, user 30 takes a photo of a tiger with mobile phone 11. The photo has GPS information of the shooting location recorded by mobile phone 11 and uploaded to the server. Figure 4 As shown in FIG. 1 , the surveillance camera 2 set up by the National Zoo captured a photo of the rhino, which has the preset GPS information of the surveillance camera 2 and is uploaded to the server. Figure 5 As shown, user 37 took a photo of an antelope with mobile phone 13. The photo has GPS information of the shooting location recorded by mobile phone 13 and uploaded to the server. The server classifies the image by animals through a trained animal classification network to obtain animal information of the image.

[0062] In this embodiment, the animal identification and classification method specifically includes the following steps:

[0063] S01: Mining online pictures, including but not limited to attractions, entertainment, etc. For S01, in the implementation of this case, different types of online pictures are mined, some scenes where animals may appear are selected, and the animals in the pictures are classified to ensure that there are no less than 500 pictures in each category. Specific categories include tigers, lions, giant pandas, monkeys, flamingos, kangaroos, koalas, elephants, deer, bears, horses / zebras, alpacas, cats and dogs, poultry, penguins, aquatic animals, other animals, and no animals.

[0064] S02: Establish a target detection model, perform image recognition on the picture, and determine whether it contains animals. The target detection model can filter out most non-animal pictures and control the input of the subsequent classification model to be no longer messy. For the target detection model in S02, since it is time-consuming to label animal detection data, an open source target detection model is adopted. After testing, the fasterrcnn-inceptionresnetv2 model has better effect. The training data of the detection model is based on the open dataset of open image, with a total of 500-600 categories. More than 100 animal categories are selected. If the picture prediction contains more than 100 animal category labels and the prediction probability is greater than 0.3, it is considered that the picture may contain animals, and the detection box corresponding to the category is cropped and output. If the picture prediction label does not contain animal categories, it is considered that there are no animals, and the classification model of S03 is not entered.

[0065] S03: If an animal is detected in the image, the animal is selected and the animal frame with the highest probability is output. An animal classification model is established to classify the animal in the frame. If no animal is detected, the classification model is not entered. For the implementation of the classification model in S03, the specific flow chart is as follows: Figure 2 Specifically, the following steps are included:

[0066] Step 1: Dataset preprocessing. The training set images are uniformly fixed to 299×299, and the images are preprocessed, including rotation, horizontal position translation, vertical position translation, projection, scaling, horizontal flipping, etc. Different images can be obtained through transformation operations, the number of training sets can be increased, and the overfitting phenomenon can be effectively reduced. Finally, normalization operations are performed to eliminate the influence of unit and scale differences between features.

[0067] Step 2: Network construction. In the present invention, according to the target characteristics, the corresponding deep convolutional neural network is built, and the basic model selects the xception deep learning model, which is faster and more accurate. Modify the last fully connected layer, replace the last fully connected layer in the original model with two fully connected layers, and add bn layer and dropout layer during training. Dropout randomly discards 50% of the nodes to prevent the model from overfitting, and adds a bn layer to speed up the training and improve the model accuracy by unifying the mean and variance of the pictures ordered in the same batch. Finally, add the classification layer for classification, including animals and others, a total of 18 categories.

[0068] Step 3: model training. In the present invention, the classification model input is the detection box in S02. Then the picture is rotated, translated, scaled, flipped, etc. to enhance the robustness of the model. The model pre-training adopts the imagenet pre-training model. Since the animal data set and the imagenet data set have certain similarities, the model is not fully trained during the training process, but the training layer is frozen, and only the fully connected layer behind is trained. The optimization algorithm adopts the RMSProp optimization algorithm. Due to the extreme imbalance of online training samples, some categories have a large number, while some categories have a small number. Different weights are adopted for different numbers of categories during training. The categories with a small number have a larger weight during training, so when iterating the loss function of the model, the categories with a small number can be optimized emphatically. Finally, the model is iterated until its test effect on the validation set reaches the best.

[0069] Step 4: Model prediction. Use the model trained in step 3 to make forward predictions for online OTA images and output the animal category of the image. If there is no animal, no output is done. Normalize the predicted image without any other preprocessing.

[0070] S04: Manual intervention and confirmation of image classification, and error correction.

[0071] For S04, the moral category predicted in S03 is matched with the manual one. If the model predicts the same category as the manual one, it is correct. Otherwise, it is wrong and further training adjustment is required. For the wrong cases, the error type is summarized. If there is a lack of relevant data in the training set, the crawler is used to add the corresponding data to the training set and repeated multiple rounds of iterations.

[0072] Finally, the animal information of the image is compared with the preset animal set, and the operation information corresponding to the animal type hit by the animal information is executed. The animal set includes multiple animal types, and each animal type corresponds to at least one operation information. A real-time updated regional electronic map is provided. The regional electronic map collects images of all image acquisition devices within the range and the positioning information of the image acquisition devices when generating images. In the regional electronic map, icons of tiger 21, rhino 22, and antelope 23 are generated according to the animal types and positioning information of the photos in other places.

[0073] In this embodiment, the animal set includes a list of dangerous animals and a preset safety radius (200 meters), and the list of dangerous animals includes tigers and rhinos, etc. When an image of a dangerous animal is obtained with the regional electronic map, a circular warning range is generated based on the positioning information of the device that obtains the image of the dangerous animal as the center and the preset safety radius (200 meters), and the positioning information of the image of the dangerous animal is sent as a warning message to all mobile terminals within the warning range.

[0074] like Figure 6 As shown, the mobile phones of users 30 and 31 in the circular warning range 25 with the positioning information of the dangerous animal tiger 21 as the center and a radius of 200 meters receive warning information, which includes the evacuation paths 40 and 41 obtained by obtaining the shortest distance path from the warning range 25 based on their current positioning information, thereby reminding users 30 and 31 to evacuate from the vicinity of the dangerous animal as soon as possible, avoiding the occurrence of dangerous accidents (tiger attacks on users), and improving the personal safety of users 30 and 31. The mobile phone of user 38 in the circular warning range 25 with the positioning information of the dangerous animal rhino 22 as the center and a radius of 200 meters receives warning information, which includes the evacuation path 42 obtained by obtaining the shortest distance path from the warning range 26 based on their current positioning information, thereby reminding user 38 to evacuate from the vicinity of the dangerous animal as soon as possible, avoiding the occurrence of dangerous accidents (rhino attacks on users), and improving the personal safety of user 38. Users 34 and 35 use mobile terminals to input the target animal antelope based on the regional electronic map. The regional electronic map then generates tour routes 60 and 61 respectively according to the positioning information of the mobile terminal and the positioning information of the closest photograph of the target animal antelope 23. Tour route 60 is sent to the mobile phone of user 35, and tour route 61 is sent to the mobile phone of user 34, thereby guiding users 34 and 35 to see the target animal antelope 23 faster.

[0075] The present invention is based on a large number of online pictures, and by using detection and classification, it can filter out most of the pictures without animals, and then further classify the pictures that may contain animals according to our needs, thereby realizing a mechanism for automatic recognition of open scenes, reducing labor costs, improving efficiency, and facilitating the subsequent further use of recognition results. The animal recognition method based on tourist images of the present invention can realize a mechanism for automatic recognition of open scenes, reducing labor costs, improving efficiency, and, by using detection and classification, it can identify the types of animals in the image and perform corresponding operations.

[0076] Figure 7 Schematic diagram of the module of the animal identification system based on tourist images of the present invention. Figure 7 As shown, the animal identification system 5 based on tourist images of the present invention includes:

[0077] The animal classification module 51 classifies the image into animals through a trained animal classification network to obtain the animal information of the image.

[0078] The operation information module 52 compares the animal information of the image with a preset animal set, and executes the operation information corresponding to the animal species matched by the animal information. The animal set includes multiple animal species, and each animal species corresponds to at least one operation information.

[0079] The animal recognition system based on tourist images of the present invention can realize a mechanism for automatically recognizing open scenes, reduce labor costs, and improve efficiency. Moreover, by using detection and classification, it can recognize the types of animals in the images and perform corresponding operations.

[0080] The embodiment of the present invention also provides an animal identification device based on tourist images, comprising a processor and a memory in which executable instructions of the processor are stored, wherein the processor is configured to execute the steps of the animal identification method based on tourist images by executing the executable instructions.

[0081] As shown above, the animal identification system based on tourist images of this embodiment of the present invention can realize a mechanism for automatic identification of open scenes, reduce labor costs, improve efficiency, and, by detection and classification, can identify the types of animals in the image and perform corresponding operations.

[0082] It will be appreciated by those skilled in the art that various aspects of the present invention may be implemented as systems, methods or program products. Therefore, various aspects of the present invention may be specifically implemented in the following forms, namely: complete hardware implementation, complete software implementation (including firmware, microcode, etc.), or a combination of hardware and software implementations, which may be collectively referred to herein as "circuits", "modules" or "platforms".

[0083] Figure 8 Schematic diagram of the structure of the animal identification device based on tourist images of the present invention. Figure 8 The electronic device 600 according to this embodiment of the present invention is described. Figure 8 The electronic device 600 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.

[0084] like Figure 8 As shown, the electronic device 600 is in the form of a general computing device. The components of the electronic device 600 may include, but are not limited to: at least one processing unit 610, at least one storage unit 620, a bus 630 connecting different platform components (including the storage unit 620 and the processing unit 610), a display unit 640, etc.

[0085] The storage unit stores program codes, which can be executed by the processing unit 610, so that the processing unit 610 executes the steps of various exemplary embodiments of the present invention described in the electronic prescription circulation processing method section of this specification. For example, the processing unit 610 can execute the following steps: Figure 1 Follow the steps shown in .

[0086] The storage unit 620 may include a readable medium in the form of a volatile storage unit, such as a random access memory unit (RAM) 6201 and / or a cache memory unit 6202 , and may further include a read-only memory unit (ROM) 6203 .

[0087] The storage unit 620 may also include a program / utility 6204 having a set (at least one) of program modules 6205, such program modules 6205 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.

[0088] Bus 630 may represent one or more of several types of bus structures, including a memory unit bus or memory unit controller, a peripheral bus, an accelerated graphics port, a processing unit, or a local bus using any of a variety of bus architectures.

[0089] The electronic device 600 may also communicate with one or more external devices 700 (e.g., keyboards, pointing devices, Bluetooth devices, etc.), may also communicate with one or more devices that enable a user to interact with the electronic device 600, and / or communicate with any device that enables the electronic device 600 to communicate with one or more other computing devices (e.g., routers, modems, etc.). Such communication may be performed via an input / output (I / O) interface 650. Furthermore, the electronic device 600 may also communicate with one or more networks (e.g., a local area network (LAN), a wide area network (WAN), and / or a public network, such as the Internet) via a network adapter 660. The network adapter 660 may communicate with other modules of the electronic device 600 via a bus 630. It should be understood that, although not shown in the figure, other hardware and / or software modules may be used in conjunction with the electronic device 600, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage platforms, etc.

[0090] The embodiment of the present invention also provides a computer-readable storage medium for storing a program, and the steps of the animal identification method based on tourist images are implemented when the program is executed. In some possible implementations, various aspects of the present invention can also be implemented in the form of a program product, which includes a program code. When the program product is run on a terminal device, the program code is used to enable the terminal device to execute the steps of various exemplary embodiments of the present invention described in the above-mentioned electronic prescription circulation processing method section of this specification.

[0091] As shown above, the animal identification system based on tourist images of this embodiment of the present invention can realize a mechanism for automatic identification of open scenes, reduce labor costs, improve efficiency, and, by detection and classification, can identify the types of animals in the image and perform corresponding operations.

[0092] Fig. 9 Schematic diagram of the structure of the computer-readable storage medium of the present invention. Fig. 9 As shown, a program product 800 for implementing the above method according to an embodiment of the present invention is described, which can adopt a portable compact disk read-only memory (CD-ROM) and include program code, and can be run on a terminal device, such as a personal computer. However, the program product of the present invention is not limited thereto. In this document, a readable storage medium can be any tangible medium containing or storing a program, which can be used by or in combination with an instruction execution system, an apparatus or a device.

[0093] The program product may use any combination of one or more readable media. The readable medium may be a readable signal medium or a readable storage medium. The readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples of readable storage media (a non-exhaustive list) include: an electrical connection with one or more wires, a portable 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.

[0094] Computer readable storage media may include data signals propagated in baseband or as part of a carrier wave, wherein readable program codes are carried. Such propagated data signals may take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. The readable storage medium may also be any readable medium other than a readable storage medium, which may send, propagate, or transmit a program for use by or in conjunction with an instruction execution system, device, or device. The program codes contained on the readable storage medium may be transmitted using any suitable medium, including but not limited to wireless, wired, optical cable, RF, etc., or any suitable combination thereof.

[0095] Program code for performing the operations of the present invention may be written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Java, C++, etc., and conventional procedural programming languages ​​such as "C" or similar programming languages. The program code may be executed entirely on the user computing device, partially on the user device, as a separate software package, partially on the user computing device and partially on a remote computing device, or entirely on a remote computing device or server. In the case of a remote computing device, the remote computing device may be connected to the user computing device through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computing device (e.g., through the Internet using an Internet service provider).

[0096] In summary, the purpose of the present invention is to provide an animal recognition method, system, device and storage medium based on tourist images. The animal recognition system based on tourist images of the present invention can realize a mechanism for automatic recognition of open scenes, reduce labor costs, improve efficiency, and utilize detection and classification to identify the types of animals in the image and perform corresponding operations.

[0097] The above contents are further detailed descriptions of the present invention in combination with specific preferred embodiments, and it cannot be determined that the specific implementation of the present invention is limited to these descriptions. For ordinary technicians in the technical field to which the present invention belongs, several simple deductions or substitutions can be made without departing from the concept of the present invention, which should be regarded as falling within the protection scope of the present invention.

Claims

1. An animal recognition method based on tourist images, characterized in that: The following steps are involved: S110, mining online images; establishing a target detection model, performing image recognition on the images, and when the probability of containing an animal image is greater than a preset threshold, outputting at least one detection frame containing an animal image; classifying and marking the animals in the frame through an animal classification model; and obtaining positioning information when the image is generated; and S120, comparing the animal information of the image with a preset animal set, executing the operation information corresponding to the animal species hit by the animal information, the animal set including multiple animal species, and each animal species corresponds to at least one operation information, when the operation information corresponding to the animal species hit by the animal information, the operation information is preset according to the animal species and positioning information of the image, and a real-time updated regional electronic map is provided, the regional electronic map collects images of all image acquisition devices within the range and the positioning information of the image acquisition devices when generating the images, the animal set includes a list of dangerous animals and a preset safety radius, when the image of the dangerous animal is obtained with the regional electronic map, a warning range is generated based on the positioning information of the image acquisition device that obtains the dangerous animal as the center and the preset safety radius, and the positioning information of the image captured by the dangerous animal is used to send a warning message to all mobile terminals within the warning range, and an evacuation path is obtained based on the current positioning information of the mobile terminals within the warning range, and the warning message includes the evacuation path.

2. The animal identification method based on tourist images according to claim 1, characterized in that: The process of classifying and marking the animals in the frame by using the animal classification model includes the following steps: S1131, unify the training set images into a preset image size, and preprocess the images, including rotation, horizontal position translation, vertical position translation, projection, scaling, and horizontal flipping; S1132. Build the corresponding deep convolutional neural network. The basic model selects the xception deep learning model, including two fully connected layers, and adds a bn layer and a dropout layer during training; the dropout layer randomly discards 50% of the nodes to prevent the model from overfitting. The bn layer is added to unify the mean and variance of the pictures in the same batch to speed up the training and improve the model accuracy. S1133. Input the image in the detection frame into the classification model, and rotate, translate, scale, and flip the image. The model pre-training adopts the imagenet pre-training model, freezes the training layer, and only trains the subsequent fully connected layer. The optimization algorithm adopts the RMSProp optimization algorithm to output the animal information of the image.

3. The animal identification method based on tourist images according to claim 2, characterized in that: The preset image size is 299 pixels×299 pixels.

4. The animal identification method based on tourist images according to claim 1, characterized in that: The mobile terminal inputs a target animal based on the regional electronic map, and the regional electronic map generates a tour route according to the positioning information of the mobile terminal and the positioning information of the closest photographed target animal.

5. An animal identification system based on tourist images, used to implement the animal identification method based on tourist images as claimed in claim 1, characterized in that: include: Animal classification module, mining online pictures; Establishing a target detection model, performing image recognition on the image, and outputting at least one detection frame containing the animal image when the probability of containing the animal image is greater than a preset threshold; classifying and marking the animals in the frame through an animal classification model; and obtaining the positioning information when the image is generated; and An operation information module compares the animal information of the image with a preset animal set, executes the operation information corresponding to the animal species hit by the animal information, the animal set includes multiple animal species, and each animal species corresponds to at least one operation information, when the animal information hits the animal species corresponding to the operation information, the operation information is preset according to the animal species and positioning information of the image, and a real-time updated regional electronic map is provided, the regional electronic map collects images of all image acquisition devices within the range and the positioning information of the image acquisition devices when generating the images, the animal set includes a list of dangerous animals and a preset safety radius, when the image of the dangerous animal is obtained with the regional electronic map, a warning range is generated based on the positioning information of the image acquisition device that obtains the dangerous animal as the center and the preset safety radius, and the positioning information of the image captured by the dangerous animal is used to send a warning message to all mobile terminals within the warning range, and an evacuation path is obtained based on the current positioning information of the mobile terminals within the warning range, and the warning message includes the evacuation path.

6. An animal identification device based on tourist images, characterized in that: include: processor; a memory storing executable instructions of the processor; Wherein, the processor is configured to perform the steps of the animal identification method based on tourism images as claimed in claim 1 by executing the executable instructions.

7. A computer-readable storage medium for storing a program, characterized in that: When the program is executed, the steps of the animal identification method based on tourist images described in claim 1 are implemented.

Images