Image acquisition method and system based on dynamic anti-crawling and dual verification

Abstract

The invention relates to the technical field of image data acquisition, in particular to an image method and system based on dynamic anti-crawling and dual verification. According to the image acquisition method, a target disease image is accurately positioned by constructing a semantic interface and utilizing keyword matching, so that the retrieval accuracy is improved; a dynamic anti-crawling strategy is adaptively designed by adopting a triple protection strategy, so that the limitation of a search engine is effectively avoided; image links are analyzed and extracted through a regular expression, and the image is ensured to be clear and complete by using dual quality verification, so that the proportion of invalid data is reduced, and the quality of the crawled image is improved; in addition, ordered management is realized by designing structured storage. According to the method, a high-quality disease image library can be quickly constructed, the data acquisition cost is remarkably reduced, the training effect of a subsequent intelligent diagnosis model is improved, a reliable data basis is provided for agricultural disease recognition and early warning, and the method has important agricultural informatization application value.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of image data collection, and particularly relates to an image collection method and system based on dynamic anti-crawling and double verification. BACKGROUND

[0002] In recent years, intelligent diagnosis technology based on computer vision has been gradually applied to the agricultural field. Disease images are crawled through the crawler technology of a general image search engine.

[0003] However, in actual application, the existing technology lacks semantic understanding ability in the agricultural field, and the search engine strictly blocks high-frequency requests (such as a threshold of 10 times per minute). The traditional crawler is easy to trigger the anti-crawling mechanism, a large number of irrelevant images are mixed in the search results, the retrieval accuracy is low, the proportion of invalid data is high, and about 35% of the crawled images have quality problems such as insufficient resolution, local blur, or format damage. In addition, the existing technology has a common structural defect in the disease image storage scheme. The extensive management of simple time stamps or random naming (such as "IMG_20230701.jpg") leads to a lack of strong semantic association between disease types and images, which seriously restricts the precision retrieval ability. SUMMARY

[0004] Based on the above problems existing in the prior art, the purpose of the embodiments of the present application is to provide an image collection method and system based on dynamic anti-crawling and double verification. The image collection method optimizes semantic search, adopts intelligent anti-crawling strategy, and innovatively integrates multi-level quality verification and structured storage architecture, which significantly improves the efficiency, quality and practicability of disease image collection, and provides high-quality data support for intelligent agriculture.

[0005] The technical scheme adopted by the present application to solve its technical problems is: In a first aspect, the present application provides an image collection method based on dynamic anti-crawling and double verification. The image collection method comprises the following steps: Step S1, obtaining a target disease name, building a semantic search interface based on the target disease name, and generating a disease image crawling page; Step S2, fusing a triple protection strategy to adaptively design a dynamic anti-crawling scheduling request of the search engine, wherein the triple protection strategy integrates a random delay algorithm, a UserAgent rotation technology and a session maintenance technology; Step S3, analyzing the disease image crawling page based on a regular expression, and selecting disease images conforming to the target disease according to a preset optimization strategy; Step S4, verifying the disease images through a double verification mechanism to filter invalid image data; Step S5: Use a three-level directory system to systematically store the disease images that have passed the verification.

[0006] Preferably, the step S1 includes: Step S11: designing a search strategy based on keyword expansion for the professional name of the target rice disease; Step S12: Encode the keywords using a URL encoding function to dynamically generate the crawled disease image pages.

[0007] Preferably, the step S2 further includes: Step S21: Simulating manual operation based on the random delay algorithm, inserting a random interval of a preset time period into each anti-climbing scheduling request; Step S22: dynamically switching access to multiple search engines using the UserAgent rotation technology; Step S23: Maintaining a stable connection state of the search engine through the session persistence technology.

[0008] Preferably, the step S3 further includes: Step S31: accurately parsing the crawled disease image page through the regular expression matching pattern; Step S32: adopting a dual-level optimization strategy, giving priority to selecting the crawled disease image page of medium-sized images, and secondarily selecting the crawled disease image page of thumbnail images; Step S33: eliminating duplicate pages of the crawled disease image through set deduplication processing to obtain the disease image.

[0009] Preferably, the dual verification mechanism in step S4 includes: Step S41: Real-time detection of the resolution of the disease image, wherein the resolution is greater than or equal to 200×200 pixels, to filter thumbnails; Step S42: calling the verify function to verify the integrity of the damaged image and automatically deleting the damaged image.

[0010] Preferably, the step S5 further includes: Step S51: Create a three-level directory system and construct a classification storage matrix according to the characteristics of the disease image; Step S52: Obtain a hash value corresponding to the disease image feature using the MD5 hash algorithm; Step S53: Construct a naming convention for storing the diseased image according to the name, serial number and hash value of the diseased image.

[0011] Preferably, the image acquisition method further comprises the following steps: monitoring the running status of the acquisition step S2 in real time and making dynamic adjustments.

[0012] In a second aspect, the present application provides an image acquisition system based on dynamic anti-crawling and double verification. The image acquisition system comprises: a semantic acquisition module, configured to acquire a target disease name, build a semantic search interface based on the target disease name, and generate a disease image page to be crawled; an anti-crawling execution module, configured to fuse a triple protection strategy to adaptively design a dynamic anti-crawling scheduling request of a search engine, wherein the triple protection strategy integrates a random delay algorithm, a UserAgent rotation technology, and a session maintenance technology; an analysis and selection module, configured to analyze the disease image page to be crawled based on a regular expression, and select a disease image conforming to the target disease according to a preset optimization strategy; a double verification module, configured to verify the disease image by configuring a double verification mechanism to filter invalid image data; an intelligent storage module, configured to store the disease image passing the verification by using a three-level directory system architecture.

[0013] Preferably, the image acquisition system further comprises an intelligent control module, configured to monitor a running state of the anti-crawling execution module in real time and dynamically adjust.

[0014] In a third aspect, the present application provides a computer readable storage medium, which stores a computer program, and the computer program is executed by a processor to realize the image method based on dynamic anti-crawling and double verification according to any one of the first aspect.

[0015] The beneficial effects of the present application are: the present application provides an image acquisition method, system and medium based on dynamic anti-crawling and double verification. The image verification method comprises the following steps: obtaining a target disease name, building a semantic search interface, and generating a disease image page; fusing a triple protection strategy to adaptively design a dynamic anti-crawling scheduling request of a search engine; analyzing the disease image page based on a regular expression, and selecting a disease image corresponding to the target disease according to a preset optimization strategy; verifying the disease image by configuring a double verification mechanism to filter invalid image data; and using a three-level directory system to structure the verified disease image. The image acquisition method of the present application can accurately locate the target disease image by constructing a semantic interface and using keyword matching, thereby improving the search accuracy. The triple protection strategy is used to adaptively design a dynamic anti-crawling strategy, which effectively avoids the restrictions of the search engine. The regular expression analysis extracts the image link and the double quality verification ensures the clarity and integrity of the image, reduces the proportion of invalid data, and improves the quality of the crawled images. In addition, the structured storage is designed to realize orderly management. The present application can quickly build a high-quality disease image library, significantly reduce the data acquisition cost, improve the training effect of the subsequent intelligent diagnosis model, provide a reliable data basis for agricultural disease identification and early warning, and has important agricultural informatization application value. BRIEF DESCRIPTION OF DRAWINGS

[0016] The present application will be further described below in conjunction with the drawings and examples.

[0017] Figure 1 is a flowchart of the image acquisition method based on dynamic anti-crawling and double verification of the present application embodiment 1; Figure 2 is a sub-step method flowchart in step S1 of the present application embodiment 1; Figure 3 is a sub-step method flowchart in step S2 of the present application embodiment 1; Figure 4 is a sub-step method flowchart in step S3 of the present application embodiment 1; Figure 5 is a sub-step method flowchart in step S4 of the present application embodiment 1; Figure 6 is a sub-step method flowchart in step S5 of the present application embodiment 1; Figure 7 is a module diagram of the image acquisition system based on dynamic anti-crawling and double verification of the present application embodiment 2. DETAILED DESCRIPTION

[0018] In order to make the purposes, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions of the present application will be described clearly and completely below with reference to the drawings. Obviously, the described embodiments are some of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative work fall within the scope of protection of the present application.

[0019] Embodiment 1 As shown in the figure, the present embodiment provides an image acquisition method based on dynamic anti-crawling and double verification, which comprises the following steps: Figures 1-6 Step S1, obtaining a target disease name, building a semantic search interface based on the target disease name, and generating a disease image page for crawling. Step S2, fusing a triple protection strategy to adaptively design a dynamic anti-crawling scheduling request of the search engine, and the triple protection strategy integrates a random delay algorithm, a UserAgent rotation technology and a session maintenance technology.

[0020] As an optional embodiment, as shown in the figure, step S1 further comprises the following steps: Figure 2 Step S11, designing a search strategy based on keyword expansion for the professional name of the target disease of rice diseases; for example, “rice blast” is expanded to “rice blast leaf spot” and “early symptoms of rice leaf blast”, etc., to improve the relevance of the search results.

[0021] Step S12, encoding the keywords by a URL encoding function to dynamically generate a disease image page for crawling. Specifically, the keywords are encoded by urllib.parse.quote (a function for converting a string to URL encoding) to dynamically generate a search engine (such as Baidu) image page URL (such as pn={page*30}, 30 images per page), to realize automatic page turning for batch image grabbing and realize accurate conversion from disease semantics to disease images.

[0022] Step S2, fusing a triple protection strategy to adaptively design a dynamic anti-crawling scheduling request of the search engine, and the triple protection strategy integrates a random delay algorithm, a UserAgent rotation technology and a session maintenance technology.

[0023] As an optional embodiment, as shown in the figure, step S2 further comprises the following steps: Figure 3As shown, step S2 further includes: step S21, simulating manual operation based on a random delay algorithm, inserting a random interval of a preset time period in each anti-crawling scheduling request; specifically, pausing for a random interval of 1.5-3.5 seconds before each request, disrupting the anti-crawling request time regularity, and avoiding being identified as a robot due to fixed frequency; for example, delaying 2.1 seconds for the first request, 3.0 seconds for the second request, and 1.8 seconds for the third request.

[0024] Step S22, dynamically switching the access of multiple search engines through UserAgent rotation technology; the purpose is to disguise as multiple browsers to avoid being identified as a crawler due to fixed UA (User Agent). Specifically, first, a dynamic UA pool is constructed, including mainstream browser identifiers (Chrome / Firefox / Safari / Edge); second, using the "safe time window" created by the random delay algorithm, randomly and dynamically switching the UA in the dynamic UA pool (such as Chrome→Firefox→Safari).

[0025] Step S23, maintaining the stable connection state of the search engine through session maintenance technology; the purpose is to reuse TCP (Transfer Control Protocol) connection, reduce request overhead, and maintain stable access state. Specifically, first, create a persistent session, under the disguise of UA rotation, reduce the overhead of frequent handshaking, maintain cookies and other session states, and ensure that the simulation is a real user browsing behavior.

[0026] As an optional embodiment, the image acquisition method further includes the following steps: monitoring the running state of step S2 in real time and dynamically adjusting. Specifically, by monitoring the random interval of step S21, when the random interval reaches 5-8 seconds, dynamically adjust to a random interval of 1.5-3.5 seconds; control the UA switching frequency in S22, which can be set to change UA every 1-3 requests; monitor the cookies and other session states in real time, and re-establish the session when abnormal.

[0027] Step S3, analyzing the disease image page based on regular expression, and selecting the disease image that meets the target disease according to the preset preferred strategy. In this embodiment, taking the search engine Baidu as an example, Baidu image search uses AIAX dynamic loading technology, and traditional web page HTML parsing method cannot directly obtain the real image link. Because the real image link data is requested asynchronously through JavaScript (AJAX), the initial HTML only contains an empty container, and the image URL is hidden in a complex JSON format, making it difficult to extract.

[0028] As an optional embodiment, as shown in FIG. 4, the image acquisition method further includes the following steps: Figure 4As shown in the figure, step S3 further includes: step S31, accurately parsing the crawled disease image page through regular expression matching mode; specifically, the regular expression is used to match the complex JSON format, and the thumbURL and middleURL two data formats are extracted therefrom.

[0029] Step S32, a two-stage preferred strategy is adopted to preferentially select the crawling disease image page of the medium size image, and secondarily select the crawling disease image page of the thumbnail; the two-stage preferred strategy is adopted to preferentially select the middleURL medium size image, and secondarily select the thumbURL thumbnail image.

[0030] Step S33, repeated crawling disease image pages are eliminated through set de-duplication processing to obtain disease images. After the crawler obtains the target image link, the Set feature of Python is used to automatically remove duplicate URLs, so that each unique link is processed only once. Because the same image may appear in different pages in the Baidu image search results.

[0031] Step S4, the disease images are verified through a double verification mechanism to filter invalid image data. The double verification of the embodiment significantly reduces the invalid image rate, and the effective image rate is as high as 93%.

[0032] As an optional embodiment, as shown in the figure, Figure 5 Step S4 further includes: step S41, detecting the resolution of the disease image in real time, and filtering the thumbnail if the resolution is greater than or equal to 200*200 pixels; and step S42, calling the verify function to check the integrity of the disease image, and automatically deleting the damaged image.

[0033] In the embodiment, about 35% of the disease images crawled by the traditional crawler have problems such as insufficient resolution (<200*200 pixels), blur, or format damage, which need to be manually screened, and the efficiency is low. Therefore, the double verification mechanism is proposed in the embodiment to solve these problems. Specifically, the resolution of the disease image is detected in real time, and images that do not meet the resolution requirement and thumbnails are filtered out; then, the verify function in the PIL library is called to deeply check the integrity of the disease image that meets the resolution requirement, including the format and file structure of the image, and the damaged image is automatically deleted.

[0034] Step S5, the verified disease images are stored in a three-level directory structure. In the embodiment, a directory tree system driven by disease image characteristics is established to improve the retrieval efficiency. The URL hash value binding naming mechanism is applied to ensure data traceability.

[0035] As an optional embodiment, as shown in the figure, Figure 6As shown, step S5 further includes: step S51, creating a three-level directory system, and constructing a classification storage matrix according to the characteristics of the disease image.

[0036] Step S52: Obtain the hash value corresponding to the disease image feature through the MD5 hash algorithm; generate a unique identifier through the MD5 hash algorithm to avoid duplicate storage and support accurate data management.

[0037] Step S53: Construct a naming convention for storing the disease images based on the names, serial numbers, and hash values ​​of the disease images. Specifically, the naming convention is disease name_serial number_MD5 hash; for example, rice blast_1_5f3a8b.

[0038] The present embodiment provides an image acquisition method based on dynamic anti-crawl and double verification. The image acquisition method of this embodiment improves the retrieval accuracy by constructing a semantic interface and using keyword matching to accurately locate the target disease image; adopts a triple protection strategy to adaptively design a dynamic anti-crawl strategy to effectively circumvent the limitations of search engines; extracts image links through regular expression analysis and uses double quality verification to ensure that the image is clear and complete, reduces the proportion of invalid data, and improves the quality of crawled images; in addition, orderly management is achieved by designing structured storage. The invention can quickly build a high-quality disease image library, significantly reduce data acquisition costs, improve the training effect of subsequent intelligent diagnosis models, provide a reliable data foundation for agricultural disease identification and early warning, and has important agricultural informatization application value.

[0039] Example 2 like Figure 7 As shown, this embodiment provides an image acquisition system based on dynamic anti-crawl and double verification. The image acquisition system 10 includes: a semantic acquisition module 11, an anti-crawl execution module 12, an analysis and selection module 13, a double verification module 14, and an intelligent storage module 15.

[0040] The semantic collection module 11 is used to obtain the target disease name, build a semantic search interface based on the target disease name, and generate a crawled disease image page.

[0041] The anti-crawling execution module 12 is configured to fuse the dynamic anti-crawling scheduling request of the search engine designed by the adaptive triple protection strategy, and the adaptive triple protection strategy integrates the random delay algorithm, the UserAgent rotation technology and the session maintenance technology. In the embodiment, the anti-crawling execution module 12 specifically avoids the anti-crawling mechanism of the search engine (such as Baidu, and the search engine in the following is taken as an example) by simulating manual operation, inserts an interval of 1.5-3.5 seconds between each request by using the random delay algorithm to avoid high-frequency triggering of anti-crawling, selects the browser identifier from the dynamic pool (Chrome / Firefox / Safari) at random by using the UserAgent rotation technology to avoid being identified as a single operation, and reduces the resource consumption by using the session maintenance technology by reusing the TCP connection through requests.Session().

[0042] The analysis and selection module 13 is configured to analyze the disease image page based on a regular expression, and select the disease image of the target disease according to a preset optimization strategy. In the embodiment, the search engine Baidu is taken as an example. The Baidu image search adopts the AIAX dynamic loading technology, and the traditional web page HTML parsing method cannot directly obtain the real image link because the real image link data is requested asynchronously through JavaScript (AJAX), the initial HTML only contains an empty container, and the image URL is hidden in a complex JSON format, and it is troublesome to extract. Specifically, in the embodiment, the analysis and selection module 13 first matches the complex JSON format by using a regular expression, and extracts two data formats of thumbURL and middleURL from the complex JSON format. The analysis and selection module 13 adopts a two-stage optimization strategy, preferentially selects the middleURL image of medium size, and secondarily selects the thumbURL thumbnail image. After the analysis and selection module 13 obtains the target image link through the crawler, the analysis and selection module 13 automatically removes the repeated URL by using the Set feature of Python to ensure that each unique link is processed only once. Because the same image may appear in different pages in the Baidu image search result.

[0043] The double verification module 14 is configured to verify the disease images by configuring a double verification mechanism to filter invalid image data. In this embodiment, about 35% of the images crawled by the traditional crawling method have problems such as insufficient resolution (<200x200 pixels), blurring or format damage, and need to be manually screened, which is inefficient. Therefore, the double verification mechanism is proposed in this embodiment to solve these problems. Specifically, the double verification module 14 first detects the resolution of the disease image in real time, and requires the resolution to be greater than or equal to 200x200 pixels. Images and thumbnails that do not meet the resolution are filtered out. Then, the double verification module 14 calls the verify function in the PIL library to deeply check the integrity of the disease images that meet the resolution, including the format and file structure of the images, and automatically deletes damaged images. The double verification of this embodiment significantly reduces the rate of invalid images, and the rate of valid images is as high as 93%.

[0044] The intelligent storage module 15 is configured to store the verified disease images using a three-level directory structure. In this embodiment, the intelligent storage module 15 improves the retrieval efficiency by establishing a directory tree system driven by disease image features. The URL hash value binding naming mechanism is used to ensure data traceability.

[0045] The intelligent control module 16 is configured to monitor the running state of the anti-crawling execution module 12 in real time and make dynamic adjustments. In this embodiment, specifically, the random interval of the random delay in the anti-crawling execution module 12 is monitored in real time. When the random interval reaches 5-8 seconds, the random interval is dynamically adjusted to 1.5-3.5 seconds. The frequency of UA switching in the anti-crawling execution module 12 is controlled, and the UA can be changed every 1-3 requests. The session state such as cookies in the anti-crawling execution module 12 is monitored in real time, and the session is re-established when it is abnormal.

[0046] The embodiment provides an image collection system based on dynamic anti-crawling and double verification. The image collection system is realized on the basis of the image collection method based on dynamic anti-crawling and double verification provided in the above embodiment 1. The image collection system uses a semantic interface, keyword matching, accurate positioning of target disease images, and improved retrieval accuracy. A triple protection strategy is used to design a dynamic anti-crawling strategy to effectively avoid the restrictions of search engines. Regular expression analysis is used to extract image links, and double quality verification is used to ensure that the images are clear and complete, reduce the proportion of invalid data, and improve the quality of the crawled images. In addition, structured storage is designed to achieve orderly management. The invention can quickly build a high-quality disease image library, significantly reduce the data collection cost, improve the training effect of the subsequent intelligent diagnosis model, provide a reliable data basis for agricultural disease identification and early warning, and has important agricultural informatization application value.

[0047] Embodiment 3 This embodiment relates to a computer readable storage medium, which stores a computer program and is installed in a warming table. The computer program, when executed by a processor, implements the above-mentioned embodiment 1.

[0048] That is, those skilled in the art can understand that all or part of the steps in the above-mentioned embodiment method can be completed by instructing the relevant hardware through a program, which is stored in a storage medium and includes a plurality of instructions for causing a device (which can be a single-chip microcomputer, a chip, etc.) or a processor to execute all or part of the steps of the method described in each embodiment of the present application. The foregoing storage medium includes a U disk, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, and various media that can store program codes.

[0049] Based on the above ideal embodiments according to the present application, through the above description, relevant personnel can make various changes and modifications without deviating from the scope of the present application. The technical scope of the present application is not limited to the contents of the specification, and must be determined according to the scope of the claims.

Claims

1. An image acquisition method based on dynamic anti-creeping and double verification, characterized in that: The image acquisition method comprises the following steps: Step S1: Obtain the name of the target disease, build a semantic search interface based on the target disease name, and generate a crawled disease image page; Step S2: Adaptively designing a dynamic anti-crawling scheduling request for a search engine by integrating a triple protection strategy, wherein the triple protection strategy integrates a random delay algorithm, a UserAgent rotation technology, and a session persistence technology; Step S3: analyzing the crawled disease image page based on regular expressions, and selecting disease images that meet the target disease according to a preset optimization strategy; Step S4: verifying the disease image by configuring a double verification mechanism to filter out invalid image data; Step S5: Use a three-level directory system to systematically store the disease images that have passed the verification.

2. The image acquisition method based on dynamic anti-creeping and double verification according to claim 1 is characterized in that: The step S1 comprises: Step S11: designing a search strategy based on keyword expansion for the professional name of the target rice disease; Step S12: Encode the keywords using a URL encoding function to dynamically generate the crawled disease image pages.

3. The image acquisition method based on dynamic anti-creeping and double verification according to claim 1 is characterized in that: The step S2 further includes: Step S21: Simulating manual operation based on the random delay algorithm, inserting a random interval of a preset time period into each anti-climbing scheduling request; Step S22: dynamically switching access to multiple search engines using the UserAgent rotation technology; Step S23: Maintaining a stable connection state of the search engine through the session persistence technology.

4. The image acquisition method based on dynamic anti-creeping and double verification according to claim 1 is characterized in that: The step S3 further includes: Step S31: accurately parsing the crawled disease image page through the regular expression matching pattern; Step S32: adopting a dual-level optimization strategy, giving priority to selecting the crawled disease image page of medium-sized images, and secondarily selecting the crawled disease image page of thumbnail images; Step S33: eliminating duplicate pages of the crawled disease image through set deduplication processing to obtain the disease image.

5. The image acquisition method based on dynamic anti-creeping and double verification according to claim 1 is characterized in that: The dual verification mechanism in step S4 includes: Step S41: Real-time detection of the resolution of the disease image, wherein the resolution is greater than or equal to 200×200 pixels, to filter thumbnails; Step S42: calling the verify function to verify the integrity of the damaged image and automatically deleting the damaged image.

6. The image acquisition method based on dynamic anti-creeping and double verification according to claim 1 is characterized in that: The step S5 further includes: Step S51: Create a three-level directory system and construct a classification storage matrix according to the characteristics of the disease image; Step S52: Obtain a hash value corresponding to the disease image feature using the MD5 hash algorithm; Step S53: Construct a naming convention for storing the diseased image according to the name, serial number and hash value of the diseased image.

7. The image acquisition method based on dynamic anti-creeping and double verification according to claim 3 is characterized in that: The image acquisition method further includes the following steps: monitoring the running status of the acquisition step S2 in real time and making dynamic adjustments.

8. An image acquisition system based on dynamic anti-creeping and double verification, characterized in that: The image acquisition system comprises: A semantic acquisition module is used to obtain the target disease name, build a semantic search interface based on the target disease name, and generate a crawled disease image page; An anti-crawling execution module, which is used to adaptively design dynamic anti-crawling scheduling requests for search engines by integrating a triple protection strategy that integrates a random delay algorithm, UserAgent rotation technology, and session persistence technology; An analysis and selection module, configured to analyze the crawled disease image page based on a regular expression and select disease images that meet the target disease according to a preset optimization strategy; A double verification module, configured to verify the disease image by configuring a double verification mechanism to filter invalid image data; The intelligent storage module is used to adopt a three-level directory system to systematically store the disease images that have passed the verification.

9. The image acquisition system based on dynamic anti-creeping and double verification according to claim 8, characterized in that: The image acquisition system also includes an intelligent control module for collecting the operating status of the anti-creep execution module through real-time monitoring and making dynamic adjustments.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the image acquisition method based on dynamic anti-crawl and double verification according to any one of claims 1 to 7 is implemented.

Images