Image feature recognition and matching method and device

Abstract

The present invention discloses an image feature recognition and matching method and device, the method comprising: obtaining first image information; performing recognition processing on the first image information to obtain image feature recognition type information; performing fusion processing on the first image information and the image feature recognition type information to obtain first trajectory information; performing calculation processing on the first trajectory information to obtain first target trajectory information. It can be seen that the present application is conducive to improving the recognition accuracy of sliding verification codes so as to smoothly obtain data sources through crawler technology.

Description

Technical Field

[0001] The present invention relates to the field of crawlers, and in particular to an image feature recognition and matching method and device. Background Art

[0002] In recent years, with the rapid development of large model technology, it is necessary to collect massive corpora from the Internet as training materials. However, many data sources have adopted anti-crawling mechanisms for crawlers. These mainly include: setting robots.txt file restrictions, limiting the UserAgent field for access filtering, restricting based on user access behavior, blocking the IP address of crawlers, and taking dynamic verification forms such as pictures or sliding verification codes when logging in. Therefore, it is particularly important to bypass the anti-crawling mechanism of the sliding verification code so that the crawler can successfully obtain the content of the data source. Summary of the invention

[0003] The technical problem to be solved by the present invention is to provide an image feature recognition and matching method and device, which is conducive to improving the recognition accuracy of sliding verification codes so as to smoothly obtain data sources through crawler technology.

[0004] In order to solve the above technical problems, the first aspect of the embodiment of the present invention discloses an image feature recognition and matching method, the method comprising:

[0005] S1, obtaining first image information;

[0006] S2, performing recognition processing on the first image information to obtain image feature recognition type information;

[0007] S3, fusing the first image information and the image feature recognition type information to obtain first trajectory information;

[0008] S4: Calculate and process the first trajectory information to obtain first target trajectory information.

[0009] As an optional implementation manner, in the first aspect of the embodiment of the present invention, the fusing process of the first image information and the image feature recognition type information to obtain the first trajectory information includes:

[0010] S31, performing an extraction operation on the image feature recognition type information to obtain first extraction result information;

[0011] When the first extraction result information is first image feature recognition type information, executing S32;

[0012] When the first extraction result information is second image feature recognition type information, executing S33;

[0013] When the first extraction result information is third image feature recognition type information, executing S34;

[0014] S32, performing calculation processing on the first image information to obtain second trajectory information, and determining the second trajectory information as the first trajectory information;

[0015] S33, performing calculation processing on the first image information to obtain third trajectory information, and determining the third trajectory information as the first trajectory information;

[0016] S34: perform calculation processing on the first image information to obtain fourth trajectory information, and determine the fourth trajectory information as the first trajectory information.

[0017] As an optional implementation manner, in the first aspect of the embodiment of the present invention, the calculating and processing the first image information to obtain the third trajectory information includes:

[0018] S331, obtaining a first calculation model and first starting point information;

[0019] S332, performing model training processing on the first computing model to obtain a second computing model;

[0020] S333, using a second calculation model, performing calculation processing on the first image information to obtain first endpoint information;

[0021] S334: merge the first starting point information and the first end point information to obtain third trajectory information.

[0022] As an optional implementation manner, in the first aspect of the embodiment of the present invention, the calculating and processing the first image information to obtain the fourth trajectory information includes:

[0023] S341, performing calculation processing on the first image information to obtain second starting point information, inflection point information and second end point information; the inflection point information includes one or two second inflection point information;

[0024] When the inflection point information includes one piece of the second inflection point information, executing S342;

[0025] When the inflection point information includes two pieces of the second inflection point information, executing S343;

[0026] S342, calculating and processing the second starting point information, the inflection point information and the second end point information to obtain fourth trajectory information;

[0027] S343, performing calculation processing on the first image information, the second starting point information and the inflection point information to obtain third inflection point information and fourth inflection point information;

[0028] S344: Fusing the second starting point information, the third inflection point information, the fourth inflection point information, and the second end point information to obtain the fourth trajectory information.

[0029] As an optional implementation manner, in the first aspect of the embodiment of the present invention, the calculating and processing the first trajectory information to obtain the first target trajectory information includes:

[0030] S41, performing an extraction operation on preset target trajectory pattern information to obtain second extraction result information;

[0031] When the second extraction result information is the first target trajectory pattern information, executing S42;

[0032] When the second extraction result information is second target trajectory pattern information, executing S43;

[0033] S42, performing calculation processing on the first trajectory information to obtain second target trajectory information, and determining the second target trajectory information as the first target trajectory information;

[0034] S43: Calculate and process the first trajectory information to obtain third target trajectory information, and determine the third target trajectory information as the first target trajectory information.

[0035] As an optional implementation manner, in the first aspect of the embodiment of the present invention, the calculating and processing the first trajectory information to obtain the second target trajectory information includes:

[0036] S421, presetting first speed information, first time information, and second time information;

[0037] S422, using a third calculation model, calculating and processing the first speed information, the first time information, and the second time information to obtain first displacement information;

[0038] Wherein, the third calculation model is:

[0039]

[0040] Wherein, MB(s1) is the first displacement information, v1 is the first velocity information, u0 is the preset initial acceleration information, w1 is the preset first position offset, S′ and S″ are the first time information and the second time information, σ1, σ2, σ3, σ4 and σ5 are the first weight parameter, the second weight parameter, the third weight parameter, the fourth weight parameter and the fifth weight parameter;

[0041] S423: Calculate and process the first trajectory information, the second time information, and the first displacement information to obtain second target trajectory information.

[0042] As an optional implementation manner, in the first aspect of the embodiment of the present invention, the calculating and processing the first trajectory information to obtain the third target trajectory information includes:

[0043] S431, using a fourth calculation model, calculating and processing the first trajectory information to obtain first coefficient information;

[0044] Wherein, the fourth calculation model is:

[0045]

[0046] Wherein, FD is the first coefficient information, JL is the distance from the starting point to the end point in the first trajectory information, T1 is the preset first coefficient parameter, σ6 is the sixth weight parameter, Indicates rounding down;

[0047] S432, using a fifth calculation model, calculating and processing the first coefficient information to obtain second displacement information;

[0048] Wherein, the fifth calculation model is:

[0049]

[0050] Wherein, NB(T2) is the second displacement information, FD is the first coefficient information, JL is the distance from the starting point to the end point in the first trajectory information, σ7 is the seventh weight parameter, Indicates rounding down;

[0051] S433: Calculate and process the first trajectory information and the second displacement information to obtain third target trajectory information.

[0052] A second aspect of an embodiment of the present invention discloses an image feature recognition and matching device, characterized in that the device comprises:

[0053] An acquisition module, used for acquiring first image information;

[0054] A recognition module, used to perform recognition processing on the first image information to obtain image feature recognition type information;

[0055] A first calculation module is used to perform a fusion process on the first image information and the image feature recognition type information to obtain first trajectory information;

[0056] The second calculation module is used to calculate and process the first trajectory information to obtain first target trajectory information.

[0057] A third aspect of an embodiment of the present invention discloses another image feature recognition and matching device, characterized in that the device comprises:

[0058] processor;

[0059] a memory coupled to the processor and storing executable program code;

[0060] The processor calls the executable program code stored in the memory to execute part or all of the steps of the image feature recognition and matching method disclosed in the first aspect of the embodiment of the present invention.

[0061] The fourth aspect of an embodiment of the present invention discloses a computer-readable storage medium, characterized in that the computer-readable storage medium stores computer instructions, and when the computer instructions are called, they are used to execute part or all of the steps of the image feature recognition and matching method disclosed in the first aspect of the embodiment of the present invention.

[0062] Compared with the prior art, the embodiments of the present invention have the following beneficial effects:

[0063] In the embodiment of the present invention, first image information is obtained; the first image information is identified and processed to obtain image feature identification type information; the first image information and the image feature identification type information are fused to obtain first trajectory information; the first trajectory information is calculated and processed to obtain first target trajectory information. It can be seen that the present application is conducive to improving the recognition accuracy of the sliding verification code so as to smoothly obtain the data source through the crawler technology. BRIEF DESCRIPTION OF THE DRAWINGS

[0064] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0065] Figure 1 A schematic diagram of a flow chart of an image feature recognition and matching method disclosed in an embodiment of the present invention;

[0066] Figure 2 A schematic diagram of the structure of an image feature recognition and matching device disclosed in an embodiment of the present invention;

[0067] Figure 3 The present invention is a schematic diagram of the structure of another image feature recognition and matching device disclosed in an embodiment of the present invention. DETAILED DESCRIPTION

[0068] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0069] The terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish different objects, rather than to describe a specific order. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, device, product or equipment that includes a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units that are not listed, or may optionally include other steps or units that are inherent to these processes, methods, products or equipment.

[0070] Reference to "embodiments" herein means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present invention. The appearance of the phrase in various places in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0071] The present invention discloses an image feature recognition and matching method and device, which is beneficial to improving the recognition accuracy of sliding verification codes so as to smoothly obtain data sources through crawler technology. Detailed descriptions are given below.

[0072] Embodiment 1

[0073] See also Figure 1 , Figure 1 : is a flow chart of an image feature recognition and matching method disclosed in an embodiment of the present invention. Figure 1 The described image feature recognition and matching method is applied to an image feature recognition and matching device, such as a local server or cloud server for image feature recognition and matching optimization management, and the embodiments of the present invention do not limit this. Figure 1 As shown, the image feature recognition and matching method may include the following operations:

[0074] S1, obtaining first image information;

[0075] S2, performing recognition processing on the first image information to obtain image feature recognition type information;

[0076] S3, fusing the first image information and the image feature recognition type information to obtain first trajectory information;

[0077] S4: Calculate and process the first trajectory information to obtain first target trajectory information.

[0078] It should be noted that the first image information is a sliding verification code image in a web page corresponding to the sliding verification code to be identified.

[0079] It can be seen that implementing the image feature recognition and matching method described in the embodiment of the present invention is conducive to improving the recognition accuracy of the sliding verification code, so as to smoothly obtain the data source through the crawler technology.

[0080] In an optional embodiment, the performing recognition processing on the first image information to obtain image feature recognition type information includes:

[0081] S21, preprocessing the first image information to obtain second image information;

[0082] It should be noted that the above preprocessing of the first image information includes converting the image into a grayscale image, binarization processing, and noise removal, etc. The conversion into a grayscale image may use a linear grayscale transformation method, the binarization processing may use an OTSU algorithm, etc., and the noise removal may use a Gaussian filtering method.

[0083] S22, performing feature analysis and recognition processing on the second image information to obtain image feature information;

[0084] It should be noted that the feature analysis and recognition processing of the second image information may use a boundary feature method, Canny edge detection, or CNN.

[0085] S23, performing fusion processing on the image feature information to obtain image feature recognition type information.

[0086] It should be noted that the fusion processing of the image feature information can be carried out using a pattern recognition algorithm. After processing by the pattern recognition algorithm, the sliding elements or puzzle pieces contained in the picture will be obtained. When the obtained sliding elements or puzzle pieces contain sliding bar information, the image feature recognition type information is determined to be a sliding bar verification code; when the obtained sliding elements or puzzle pieces contain puzzle information, the image feature recognition type information is determined to be a puzzle verification code; when the obtained sliding elements or puzzle pieces contain specific line segment information, the image feature recognition type information is determined to be a VAPTCHA verification code.

[0087] It can be seen that implementing the image feature recognition and matching method described in the embodiment of the present invention is conducive to improving the recognition accuracy of the sliding verification code, so as to smoothly obtain the data source through the crawler technology.

[0088] In another optional embodiment, the fusing the first image information and the image feature recognition type information to obtain the first trajectory information includes:

[0089] S31, performing an extraction operation on the image feature recognition type information to obtain first extraction result information;

[0090] When the first extraction result information is first image feature recognition type information, executing S32;

[0091] When the first extraction result information is second image feature recognition type information, executing S33;

[0092] When the first extraction result information is third image feature recognition type information, executing S34;

[0093] It should be noted that the first image feature recognition type information is a sliding bar verification code, the second image feature recognition type information is a puzzle verification code, and the third image feature recognition type information is a VAPTCHA verification code.

[0094] S32, performing calculation processing on the first image information to obtain second trajectory information, and determining the second trajectory information as the first trajectory information;

[0095] It should be noted that the first image information is calculated and processed to obtain the second trajectory information, and the web page corresponding to the first image information can be used to parse its elements to obtain the starting position and the end position of the sliding bar verification code, such as using Python's BeautifulSoup for parsing. After the starting position and the end position are obtained, the straight line connecting the starting position and the end position is the second trajectory information.

[0096] S33, performing calculation processing on the first image information to obtain third trajectory information, and determining the third trajectory information as the first trajectory information;

[0097] S34: perform calculation processing on the first image information to obtain fourth trajectory information, and determine the fourth trajectory information as the first trajectory information.

[0098] It can be seen that implementing the image feature recognition and matching method described in the embodiment of the present invention is conducive to improving the recognition accuracy of the sliding verification code, so as to smoothly obtain the data source through the crawler technology.

[0099] In an optional embodiment, the performing calculation processing on the first image information to obtain the third trajectory information includes:

[0100] S331, obtaining a first calculation model and first starting point information;

[0101] It should be noted that the first calculation model is Faster R-CNN, and the first starting point information is the position of the sliding block in the first image information.

[0102] S332, performing model training processing on the first computing model to obtain a second computing model;

[0103] It should be noted that the first computing model is subjected to model training processing to obtain the second computing model, which is obtained by training Faster R-CNN through ImageNet. The trained Faster R-CNN obtained is the second computing model, including:

[0104] 1) Use the model parameters pre-trained by ImageNet to initialize the RPN network in Faster R-CNN to obtain the candidate box;

[0105] 2) Initialize the Fast R-CNN part of Faster R-CNN using the model parameters pre-trained by ImageNet, and train it using the candidate boxes and training set generated in 1) to obtain the parameters of the trained Fast R-CNN;

[0106] 3) Use the parameters of Fast R-CNN to initialize the RPN part for fine-tuning, keep the parameters of the feature extraction part unchanged, and only fine-tune the RPN part to obtain the RPN parameters after training;

[0107] 4) Initialize Fast R-CNN with the RPN parameters in step 3), and also keep the feature extraction parameters unchanged and only fine-tune the unique parameters of Fast R-CNN. At this point, the entire Faster R-CNN detection network is trained, and the trained Faster R-CNN model is obtained, which is the second calculation model.

[0108] S333, using a second calculation model, performing calculation processing on the first image information to obtain first endpoint information;

[0109] It should be noted that the first image information is calculated and processed using the second calculation model to obtain the first endpoint information, which is to use the first image information as input and use the Faster R-CNN model to calculate to obtain the first endpoint information, and the first endpoint information is the gap position of the sliding block in the first image information.

[0110] S334: merge the first starting point information and the first end point information to obtain third trajectory information.

[0111] It should be noted that the first starting point information and the first end point information are fused to obtain the third trajectory information, which is obtained by connecting the first starting point information and the first end point information, and the straight line segment trajectory obtained is the third trajectory information;

[0112] It should be noted that the first starting point information and the first end point information are fused to obtain the third trajectory information, which can also be obtained using the sixth calculation model;

[0113] Wherein, the sixth calculation model is:

[0114]

[0115] Wherein, (A, B) is the third trajectory information, (A1, B1) is the first starting point information, (A2, B2) is the first end point information, is the second position offset, and the second position offset will randomly fluctuate within a preset second offset range. The value range of A is between A1 and A2, and the value range of B is between B1 and B2.

[0116] It can be seen that implementing the image feature recognition and matching method described in the embodiment of the present invention is conducive to improving the recognition accuracy of the sliding verification code, so as to smoothly obtain the data source through the crawler technology.

[0117] In an optional embodiment, the performing calculation processing on the first image information to obtain fourth trajectory information includes:

[0118] S341, performing calculation processing on the first image information to obtain second starting point information, inflection point information and second end point information; the inflection point information includes one or two second inflection point information;

[0119] When the inflection point information includes one piece of the second inflection point information, executing S342;

[0120] When the inflection point information includes two pieces of the second inflection point information, executing S343;

[0121] It should be noted that when the first image information is the type of VAPTCHA verification code, the dragging trajectory of the sliding verification code is composed of a line segment with a certain width and related icons, among which the starting point of the line segment has a more obvious feature, which is a semicircular shape with a certain curvature, and the end point of the line segment is an arrow-shaped pattern, and there are several (one or two) inflection points in the middle.

[0122] It should be noted that the calculation and processing of the first image information to obtain the second starting point information, the inflection point information and the second end point information is obtained by Faster R-CNN. When the inflection point information is calculated by Faster R-CNN, more than two second inflection point information may be obtained. At this time, the confidence output by Faster R-CNN is used to determine the two second inflection point information with the highest confidence as the inflection point information in the first image information.

[0123] S342, calculating and processing the second starting point information, the inflection point information and the second end point information to obtain fourth trajectory information;

[0124] It should be noted that the straight line segment from the second starting point information to the second inflection point information plus the straight line segment from the second inflection point information to the second end point information in sequence is determined as the fourth trajectory information, and the moving direction of the trajectory is from the second starting point information to the second inflection point information to the second end point information.

[0125] S343, performing calculation processing on the first image information, the second starting point information and the inflection point information to obtain third inflection point information and fourth inflection point information;

[0126] It should be noted that the first image information, the second starting point information and the inflection point information are calculated and processed to obtain the third inflection point information and the fourth inflection point information by LeNet-5. The third inflection point information is directly connected to the second starting point information, and the fourth inflection point information is directly connected to the second end point information.

[0127] It should be noted that the fourth trajectory information in the first image information is a line segment with a certain width, that is, there is a more obvious feature between the two connected points. Therefore, it is only necessary to cut out the fragments connected to the two second inflection point information and the second starting point information to determine the order of the second inflection points. At this time, LeNet-5 is used to implement the binary classification task to complete the order judgment, where the LeNet-5 network structure consists of 3 convolutional layers, 2 maximum pooling layers and 2 fully connected layers. All convolution kernel sizes are 5*5, the number of channels is 16, 32 and 120, the number of channels of the first fully connected layer is 84, and because it is a binary classification task, the number of channels of the last fully connected layer is 2.

[0128] S344: Fusing the second starting point information, the third inflection point information, the fourth inflection point information, and the second end point information to obtain the fourth trajectory information.

[0129] It should be noted that the second starting point information, the third inflection point information, the fourth inflection point information and the second end point information are fused to obtain the fourth trajectory information, which is a line segment connected in sequence according to the second starting point information, the third inflection point information, the fourth inflection point information and the second end point information, that is, the fourth trajectory information.

[0130] It should be noted that the fourth trajectory information can be obtained by fusing the second starting point information, the third inflection point information, the fourth inflection point information and the second end point information, which can also be obtained by using the seventh calculation model;

[0131] Among them, the seventh calculation model is:

[0132]

[0133] Wherein, f(t) is a point on the fourth trajectory information, XA, XB, XC and XD are the second starting point information, the third inflection point information, the fourth inflection point information and the second end point information respectively, and They are respectively the third position offset, the fourth position offset and the fifth position offset, wherein the third position offset will randomly fluctuate within a preset third offset range, the fourth position offset will randomly fluctuate within a preset fourth offset range, and the fifth position offset will randomly fluctuate within a preset fifth offset range.

[0134] It should be noted that the value of t is a preset trajectory step. For example, if the step is 0.01, the value of t is 0, 0.01, 0.02, 0.03, etc., up to 3. When t is taken from 1 to 3, the point f(t) on the fourth trajectory information obtained is the complete fourth trajectory information.

[0135] It can be seen that implementing the image feature recognition and matching method described in the embodiment of the present invention is conducive to improving the recognition accuracy of the sliding verification code, so as to smoothly obtain the data source through the crawler technology.

[0136] In an optional embodiment, the calculating and processing the first trajectory information to obtain the first target trajectory information includes:

[0137] S41, performing an extraction operation on preset target trajectory pattern information to obtain second extraction result information;

[0138] When the second extraction result information is the first target trajectory pattern information, executing S42;

[0139] When the second extraction result information is second target trajectory pattern information, executing S43;

[0140] S42, performing calculation processing on the first trajectory information to obtain second target trajectory information, and determining the second target trajectory information as the first target trajectory information;

[0141] S43: Calculate and process the first trajectory information to obtain third target trajectory information, and determine the third target trajectory information as the first target trajectory information.

[0142] It can be seen that implementing the image feature recognition and matching method described in the embodiment of the present invention is conducive to improving the recognition accuracy of the sliding verification code, so as to smoothly obtain the data source through the crawler technology.

[0143] In an optional embodiment, the calculating and processing the first trajectory information to obtain the second target trajectory information includes:

[0144] S421, presetting first speed information, first time information, and second time information;

[0145] It should be noted that the second time information is the total time for the simulated mouse to move from the starting point of the first trajectory information to the end point along the trajectory direction during the sliding verification code recognition process. The first time information is a certain time in the preset mouse movement process, and the first speed information is the initial speed of the mouse when it starts moving from the starting point.

[0146] S422, using a third calculation model, calculating and processing the first speed information, the first time information, and the second time information to obtain first displacement information;

[0147] Wherein, the third calculation model is:

[0148]

[0149] Wherein, MB(s1) is the first displacement information, v1 is the first velocity information, u0 is the preset initial acceleration information, w1 is the preset first position offset, S′ and S″ are the first time information and the second time information, σ1, σ2, σ3, σ4 and σ5 are the first weight parameter, the second weight parameter, the third weight parameter, the fourth weight parameter and the fifth weight parameter;

[0150] It should be noted that the first weight parameter, the second weight parameter, the third weight parameter, the fourth weight parameter and the fifth weight parameter may be set by the user or may be obtained based on historical data, and the present invention does not make any specific limitation thereto.

[0151] S423: Calculate and process the first trajectory information, the second time information, and the first displacement information to obtain second target trajectory information.

[0152] It should be noted that the calculation and processing of the first trajectory information, the second time information and the first displacement information to obtain the second target trajectory information is to superimpose the first displacement information on all trajectory positions of the first trajectory information in the process of simulating the movement of the mouse from the starting point to the end point of the first trajectory information. For example, at time Z1, the mouse moves to the F1 trajectory position of the first trajectory information. At this time, the F1 trajectory position of the first trajectory information is added to the first displacement information MB(Z1) at time Z1, and the trajectory position F1+MB(Z1) of the second target trajectory information corresponding to the F1 trajectory position in the first trajectory information is obtained. The second target trajectory information is obtained by performing the above-mentioned superposition operation on all trajectory positions of the first trajectory information within the time of trajectory movement, that is, within the second time information.

[0153] It should be noted that after obtaining the second target trajectory information, the simulated mouse is moved from the starting point of the second target trajectory information to the end point of the second target trajectory information within the second time information, thereby completing the sliding verification code recognition.

[0154] It can be seen that implementing the image feature recognition and matching method described in the embodiment of the present invention is conducive to improving the recognition accuracy of the sliding verification code, so as to smoothly obtain the data source through the crawler technology.

[0155] In an optional embodiment, the calculating and processing the first trajectory information to obtain the third target trajectory information includes:

[0156] S431, using a fourth calculation model, calculating and processing the first trajectory information to obtain first coefficient information;

[0157] Wherein, the fourth calculation model is:

[0158]

[0159] Wherein, FD is the first coefficient information, JL is the distance from the starting point to the end point in the first trajectory information, T1 is the preset first coefficient parameter, σ6 is the sixth weight parameter, Indicates rounding down;

[0160] It should be noted that the sixth weight parameter may be set by a user or obtained based on historical data, and the present invention does not make any specific limitation thereto.

[0161] S432, using a fifth calculation model, calculating and processing the first coefficient information to obtain second displacement information;

[0162] Wherein, the fifth calculation model is:

[0163]

[0164] Wherein, NB(T2) is the second displacement information, FD is the first coefficient information, JL is the distance from the starting point to the end point in the first trajectory information, σ7 is the seventh weight parameter, Indicates rounding down;

[0165] It should be noted that the seventh weight parameter may be set by a user or obtained based on historical data, and the present invention does not make any specific limitation thereto.

[0166] S433: Calculate and process the first trajectory information and the second displacement information to obtain third target trajectory information.

[0167] It should be noted that the calculation and processing of the first trajectory information and the second displacement information to obtain the third target trajectory information is to superimpose the second displacement information on all trajectory positions of the first trajectory information in the process of simulating the movement of the mouse from the starting point to the end point of the first trajectory information. For example, at time Z1, the mouse moves to the F1 trajectory position of the first trajectory information. At this time, the F1 trajectory position of the first trajectory information is added to the second displacement information NB(Z1) at time Z1 to obtain the trajectory position F1+NB(Z1) of the third target trajectory information corresponding to the F1 trajectory position in the first trajectory information. By calculating the time during which the trajectory moves, that is, Within the time, the above-mentioned superposition operation is performed on all the trajectory positions of the first trajectory information to obtain the third target trajectory information.

[0168] It should be noted that after obtaining the third target trajectory information, the simulated mouse Move from the starting point of the third target trajectory information to the end point of the third target trajectory information within the time, thereby completing the sliding verification code recognition.

[0169] It can be seen that implementing the image feature recognition and matching method described in the embodiment of the present invention is conducive to improving the recognition accuracy of the sliding verification code, so as to smoothly obtain the data source through the crawler technology.

[0170] It should be noted that the simulated mouse movement in the first embodiment may be set by the user, or may be performed using other existing tools or methods, and the present invention does not make any specific limitation thereto.

[0171] Embodiment 2

[0172] See also Figure 2 , Figure 2 : is a schematic diagram of the structure of an image feature recognition and matching device disclosed in an embodiment of the present invention. Figure 2 The described image feature recognition and matching device is applied to an image feature recognition and matching optimization system, such as a local server or cloud server for image feature recognition and matching, which is not limited in the embodiments of the present invention. Figure 2 As shown, the image feature recognition and matching device includes:

[0173] An acquisition module 201 is used to acquire first image information;

[0174] The recognition module 202 is used to perform recognition processing on the first image information to obtain image feature recognition type information;

[0175] A first calculation module 203 is used to perform a fusion process on the first image information and the image feature recognition type information to obtain first trajectory information;

[0176] The second calculation module 204 is used to calculate and process the first trajectory information to obtain first target trajectory information.

[0177] It can be seen that implementing the image feature recognition and matching device described in the embodiment of the present invention is conducive to improving the recognition accuracy of the sliding verification code, so as to smoothly obtain the data source through the crawler technology.

[0178] Embodiment 3

[0179] See also Figure 3 , Figure 3 : is a schematic diagram of the structure of an image feature recognition and matching device disclosed in an embodiment of the present invention. Figure 3 The described image feature recognition and matching device is applied to an image feature recognition and matching optimization system, such as a local server or cloud server for image feature recognition and matching, which is not limited in the embodiments of the present invention. Figure 3 As shown, the image feature recognition and matching device includes:

[0180] Processor 301;

[0181] A memory 302 coupled to the processor 301 and storing executable program codes;

[0182] The processor 301 calls the executable program code stored in the memory 302 to execute part or all of the steps of the image feature recognition and matching method described in the first embodiment.

[0183] It can be seen that implementing the image feature recognition and matching device described in the embodiment of the present invention is conducive to improving the recognition accuracy of the sliding verification code, so as to smoothly obtain the data source through the crawler technology.

[0184] Embodiment 4

[0185] An embodiment of the present invention discloses a computer-readable storage medium, characterized in that the computer-readable storage medium stores computer instructions, and when the computer instructions are called, they are used to execute part or all of the steps of the image feature recognition and matching method described in Example 1.

[0186] Embodiment 5

[0187] An embodiment of the present invention discloses a computer program product, which includes a non-transitory computer-readable storage medium storing a computer program, and the computer program is operable to enable a computer to execute some or all of the steps in the image feature recognition and matching method described in Embodiment 1.

[0188] The system embodiments described above are only illustrative, wherein the modules described as separate components may or may not be physically separated, and the components displayed as modules may or may not be physical modules, i.e., they may be located in one place, or they may be distributed on multiple network modules. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Those of ordinary skill in the art may understand and implement it without creative labor.

[0189] Through the specific description of the above embodiments, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on such an understanding, the above technical solution can be essentially or partly contributed to the prior art in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, and the storage medium includes a read-only memory (ROM), a random access memory (RAM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), a one-time programmable read-only memory (OTPROM), an electronically erasable rewritable read-only memory (EEPROM), a compact disc (CD-ROM) or other optical disc storage, magnetic disk storage, magnetic tape storage, or any other computer-readable medium that can be used to carry or store data.

[0190] Finally, it should be noted that the image feature recognition and matching method and device disclosed in the embodiment of the present invention discloses only the preferred embodiments of the present invention, which are only used to illustrate the technical solution of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments can still be modified, or some of the technical features can be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for image feature recognition and matching, characterized in that: The method comprises: S1, obtaining first image information; S2, performing recognition processing on the first image information to obtain image feature recognition type information; S3, fusing the first image information and the image feature recognition type information to obtain first trajectory information, including: S31, performing an extraction operation on the image feature recognition type information to obtain first extraction result information; When the first extraction result information is first image feature recognition type information, executing S32; When the first extraction result information is second image feature recognition type information, executing S33; When the first extraction result information is third image feature recognition type information, executing S34; S32, performing calculation processing on the first image information to obtain second trajectory information, and determining the second trajectory information as the first trajectory information; S33, performing calculation processing on the first image information to obtain third trajectory information, and determining the third trajectory information as the first trajectory information, including: S331, obtaining a first calculation model and first starting point information; S332, performing model training processing on the first computing model to obtain a second computing model; S333, using a second calculation model, performing calculation processing on the first image information to obtain first endpoint information; S334, fusing the first starting point information and the first end point information to obtain third trajectory information; S34, performing calculation processing on the first image information to obtain fourth trajectory information, and determining the fourth trajectory information as the first trajectory information, including: S341, performing calculation processing on the first image information to obtain second starting point information, inflection point information and second end point information; the inflection point information includes one or two second inflection point information; When the inflection point information includes one piece of the second inflection point information, executing S342; When the inflection point information includes two pieces of the second inflection point information, executing S343; S342, calculating and processing the second starting point information, the inflection point information and the second end point information to obtain fourth trajectory information; S343, performing calculation processing on the first image information, the second starting point information and the inflection point information to obtain third inflection point information and fourth inflection point information; S344, fusing the second starting point information, the third inflection point information, the fourth inflection point information and the second end point information to obtain the fourth trajectory information; S4, calculating and processing the first trajectory information to obtain first target trajectory information, including: S41, performing an extraction operation on preset target trajectory pattern information to obtain second extraction result information; When the second extraction result information is the first target trajectory pattern information, executing S42; When the second extraction result information is second target trajectory pattern information, executing S43; S42, performing calculation processing on the first trajectory information to obtain second target trajectory information, and determining the second target trajectory information as the first target trajectory information, including: S421, presetting first speed information, first time information, and second time information; S422, using a third calculation model, calculating and processing the first speed information, the first time information, and the second time information to obtain first displacement information; Wherein, the third calculation model is: Wherein, MB(s1) is the first displacement information, v1 is the first velocity information, u0 is the preset initial acceleration information, w1 is the preset first position offset, S′ and S″ are the first time information and the second time information, σ1, σ2, σ3, σ4 and σ5 are the first weight parameter, the second weight parameter, the third weight parameter, the fourth weight parameter and the fifth weight parameter; S423, calculating and processing the first trajectory information, the second time information, and the first displacement information to obtain second target trajectory information; S43, performing calculation processing on the first trajectory information to obtain third target trajectory information, and determining the third target trajectory information as the first target trajectory information, including: S431, using a fourth calculation model, calculating and processing the first trajectory information to obtain first coefficient information; Wherein, the fourth calculation model is: Wherein, FD is the first coefficient information, JL is the distance from the starting point to the end point in the first trajectory information, T1 is the preset first coefficient parameter, σ6 is the sixth weight parameter, Indicates rounding down; S432, using a fifth calculation model, calculating and processing the first coefficient information to obtain second displacement information; Wherein, the fifth calculation model is: Wherein, NB(T2) is the second displacement information, FD is the first coefficient information, JL is the distance from the starting point to the end point in the first trajectory information, σ7 is the seventh weight parameter, Indicates rounding down; S433: Calculate and process the first trajectory information and the second displacement information to obtain third target trajectory information.

Images