An image information retrieval method and system based on local matching

Abstract

The present invention is applicable to the field of image retrieval technology, and provides an image information retrieval method and system based on local matching, including the following steps: receiving a target image, performing feature recognition on each target image to obtain feature information, performing an intersection process on the color sets of all target images to obtain an intersection color feature, and determining an additional color set corresponding to each feature information; based on the intersection color feature, performing a preliminary screening on the images in the image library to obtain preliminary retrieved images; based on the additional color set, performing a secondary screening on the preliminary retrieved images to obtain secondary retrieved images; performing stroke drawing on the colors in the intersection color feature and the additional color set in the secondary retrieved images to obtain a retrieval contour for each color; calculating the similarity of the secondary retrieved images based on the contour feature. In this way, calculating the similarity only based on the contours of effective colors not only reduces the amount of calculation, but also excludes the interference of other colors, and the retrieval result is more accurate.

Description

Technical Field

[0001] The present invention relates to the technical field of image retrieval, and in particular, to an image information retrieval method and system based on local matching. Background Art

[0002] When users use image retrieval, they have the following requirements: given a target image, retrieve pictures that contain the target image, and the local part of the picture matches the target image. Most of the existing image retrieval technologies extract global features in the unit of the entire image, obtain feature vectors that can represent image information, and then perform similarity retrieval. This kind of global retrieval easily leads to low similarity between the pictures containing the target image and the target image itself. Especially when the picture itself is complex and the proportion of the target image is small, the retrieval accuracy is not high; and the image detection method based on global feature similarity matching requires a large amount of computation and the retrieval is not efficient. Therefore, it is necessary to provide an image information retrieval method and system based on local matching, aiming to solve or alleviate the above problems. Summary of the Invention

[0003] Aiming at the deficiencies of the existing technology, the purpose of the present invention is to provide an image information retrieval method and system based on local matching to solve or alleviate the problems existing in the above background art.

[0004] The present invention is implemented as follows. An image information retrieval method based on local matching, the method includes the following steps:

[0005] Receive a plurality of target images at different angles, perform feature recognition on each target image to obtain feature information, the feature information includes a color set, each color in the color set corresponds to a contour feature, and each color corresponds to the number of color blocks;

[0006] Perform intersection processing on the color sets of all target images to obtain intersection color features, and determine the corresponding additional color sets for each feature information based on the intersection color features;

[0007] Based on the intersection color features, preliminarily screen the images in the image library to obtain preliminarily retrieved images;

[0008] Based on the additional color sets, perform secondary screening on the preliminarily retrieved images to obtain secondarily retrieved images, and classify the secondarily retrieved images according to the additional color sets, and each category corresponds to feature information;

[0009] Stroke the colors in the intersection color features and the additional color sets in the secondarily retrieved images to obtain the retrieval contours of each color, and determine the number of color blocks corresponding to each color;

[0010] Calculate the similarity of the secondary retrieved images based on the number of color patches and contour features, and output the final retrieved images according to the calculation results.

[0011] As a further solution of the present invention: The step of performing feature recognition on each target image to obtain feature information specifically includes:

[0012] Identify the colors included in the target image to obtain a color set;

[0013] Perform edge tracing on each color edge to obtain the contour features of each color;

[0014] Determine the number of color patches according to the number of contours in the contour features of each color, and obtain feature information according to the color set, contour features, and the number of color patches.

[0015] As a further solution of the present invention: The step of performing secondary screening on the preliminary retrieved images based on the additional color set to obtain secondary retrieved images and classifying the secondary retrieved images according to the additional color set specifically includes:

[0016] Retrieve the additional color set corresponding to each feature information, where the additional color set contains zero, one, or more additional colors;

[0017] Arrange the additional color sets in descending order according to the number of types of additional colors;

[0018] Perform secondary screening on the preliminary retrieved images according to the descending order to obtain secondary retrieved images, and each secondary retrieved image corresponds to an additional color set;

[0019] Classify the secondary retrieved images according to the additional color sets, and each category corresponds to an additional color set.

[0020] As a further solution of the present invention: The step of performing secondary screening on the preliminary retrieved images according to the descending order to obtain secondary retrieved images specifically includes:

[0021] Determine the additional color set ranked first according to the descending order;

[0022] Perform secondary screening on the preliminary retrieved images according to the additional colors in the additional color set to obtain secondary retrieved images corresponding to the additional color set;

[0023] Remove the obtained secondary retrieved images from the preliminary retrieved images, and remove the additional color set from the descending order;

[0024] Repeat the above three steps until the secondary retrieved images are empty or the additional color sets in the descending order are empty.

[0025] As a further solution of the present invention: The step of calculating the similarity of the secondary retrieval image based on the number of color blocks and contour features specifically includes:

[0026] Determine the feature information corresponding to the secondary retrieval image, and screen the secondary retrieval image based on the number of color blocks of the feature information;

[0027] Calculate the contour similarity between the retrieval contour of each color in the screened secondary retrieval image and the contour feature of each color in the feature information;

[0028] Calculate the average value of the contour similarities of all colors to obtain the calculation result.

[0029] As a further solution of the present invention: The step of screening the secondary retrieval image based on the number of color blocks of the feature information specifically includes:

[0030] Compare the number of color blocks of each color in the feature information with the number of color blocks of the corresponding color in the secondary retrieval image;

[0031] When the number of color blocks of any color in the feature information > K × the number of color blocks of the corresponding color in the secondary retrieval image, where K is a fixed value greater than 1, the corresponding secondary retrieval image is deleted.

[0032] Another object of the present invention is to provide an image information retrieval system based on local matching. The system includes:

[0033] A target image upload module, configured to receive a plurality of target images from different angles, perform feature recognition on each target image to obtain feature information, where the feature information includes a color set, each color in the color set corresponds to a contour feature, and each color corresponds to a number of color blocks;

[0034] An intersection color feature module, configured to perform intersection processing on the color sets of all target images to obtain intersection color features, and determine an additional color set corresponding to each feature information based on the intersection color features;

[0035] An image preliminary retrieval module, configured to preliminarily screen the images in the image library based on the intersection color features to obtain preliminary retrieval images;

[0036] An image secondary retrieval module, configured to perform secondary screening on the preliminary retrieval images based on the additional color set to obtain secondary retrieval images, classify the secondary retrieval images according to the additional color set, and each category corresponds to feature information;

[0037] A secondary retrieval processing module, configured to stroke the colors in the intersection color features and the additional color set in the secondary retrieval images to obtain the retrieval contour of each color, and determine the number of color blocks corresponding to each color;

[0038] An image final retrieval module, which is used to calculate the similarity of the secondary retrieval images based on the number of color blocks and contour features, and output the final retrieval images according to the calculation results.

[0039] As a further solution of the present invention: the target image uploading module includes:

[0040] An image color recognition unit, which is used to recognize the colors contained in the target image to obtain a color set;

[0041] A contour feature determination unit, which is used to stroke the edges of each color to obtain the contour features of each color;

[0042] A feature information determination unit, which is used to determine the number of color blocks according to the number of contours in the contour features of each color, and obtain feature information according to the color set, contour features, and the number of color blocks.

[0043] As a further solution of the present invention: the image secondary retrieval module includes:

[0044] An additional color retrieval unit, which is used to retrieve the additional color set corresponding to each feature information, and the additional color set contains zero, one or more additional colors;

[0045] A set descending order arrangement unit, which is used to arrange the additional color sets in descending order according to the number of types of additional colors;

[0046] A secondary screening image unit, which is used to perform secondary screening on the preliminary retrieval images according to the descending order to obtain secondary retrieval images, and each secondary retrieval image corresponds to an additional color set;

[0047] A retrieval image classification unit, which is used to classify the secondary retrieval images according to the additional color sets, and each category corresponds to an additional color set.

[0048] As a further solution of the present invention: the image final retrieval module includes:

[0049] A color block number screening unit, which is used to determine the feature information corresponding to the secondary retrieval images, and screen the secondary retrieval images based on the number of color blocks of the feature information;

[0050] A contour similarity calculation unit, which is used to calculate the contour similarity between the retrieval contour of each color in the screened secondary retrieval images and the contour features of each color in the feature information;

[0051] A calculation result determination unit, which is used to calculate the average value of the contour similarities of all colors to obtain the calculation result.

[0052] Compared with the prior art, the beneficial effects of the present invention are:

[0053] By performing feature recognition on each target image, the present invention obtains feature information, then performs an intersection process on the color sets of all target images to obtain intersection color features, and determines an additional color set corresponding to each feature information based on the intersection color features; then, based on the intersection color features, the images in the image library are preliminarily screened to obtain preliminarily retrieved images, and the preliminarily retrieved images are secondarily screened based on the additional color sets to obtain secondarily retrieved images. Thus, the present invention only uses color features for retrieval, which is efficient and fast. Then, the colors in the intersection color features and the additional color sets in the secondarily retrieved images are stroked to obtain the retrieval contours of each color, and the number of color blocks corresponding to each color is determined. Based on the number of color blocks and the contour features, the similarity of the secondarily retrieved images is calculated, and the final retrieved images are output according to the calculation results. In this way, the similarity is calculated only based on the contours of the colors in the intersection color features and the additional color sets, which not only reduces the calculation amount, but also excludes the interference of other colors, and introduces the number of color blocks, making the retrieval results more accurate. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] Figure 1 It is a flowchart of an image information retrieval method based on local matching.

[0055] Figure 2 It is a flowchart of performing feature recognition on a target image in an image information retrieval method based on local matching.

[0056] Figure 3 It is a flowchart of obtaining secondarily retrieved images in an image information retrieval method based on local matching.

[0057] Figure 4 It is a flowchart of secondarily screening the preliminarily retrieved images according to the descending order in an image information retrieval method based on local matching.

[0058] Figure 5 It is a flowchart of calculating similarity in an image information retrieval method based on local matching.

[0059] Figure 6 It is a flowchart of screening the secondarily retrieved images in an image information retrieval method based on local matching.

[0060] Figure 7 It is a schematic structural diagram of an image information retrieval system based on local matching. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0061] To make the objectives, technical solutions and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0062] The specific implementation of the present invention will be described in detail below with reference to specific embodiments.

[0063] As Figure 1 shown, an image information retrieval method based on local matching is provided in an embodiment of the present invention. The method includes the following steps:

[0064] S100, receive a plurality of target images at different angles, perform feature recognition on each target image to obtain feature information, where the feature information includes a color set, each color in the color set corresponds to a contour feature, and each color corresponds to the number of color blocks;

[0065] S200, perform an intersection process on the color sets of all target images to obtain an intersection color feature, and determine an additional color set corresponding to each feature information based on the intersection color feature;

[0066] S300, perform a preliminary screening on the images in the image library based on the intersection color feature to obtain preliminary retrieved images;

[0067] S400, perform a secondary screening on the preliminary retrieved images based on the additional color set to obtain secondary retrieved images, and classify the secondary retrieved images according to the additional color set, and each category corresponds to feature information;

[0068] S500, perform edge tracing on the colors in the intersection color feature and the additional color set in the secondary retrieved images to obtain a retrieval contour for each color, and determine the number of color blocks corresponding to each color;

[0069] S600, calculate the similarity of the secondary retrieved images based on the number of color blocks and the contour feature, and output the final retrieved images according to the calculation results.

[0070] In the embodiment of the present invention, first, the user needs to upload several target images from different angles. Of course, the user can also only upload one target image. Then, the embodiment of the present invention will perform feature recognition on each target image. Here, the feature recognition is simply to recognize the colors contained in the image to obtain feature information. The feature information includes a color set. Each color in the color set corresponds to a contour feature, and each color corresponds to the number of color blocks. For example, a certain color set is: red, blue, and yellow, where the number of color blocks corresponding to red is four. Each target image corresponds to one piece of feature information. Then, it is necessary to perform an intersection process on the color sets of all target images to obtain an intersection color feature, and determine an additional color set corresponding to each piece of feature information based on the intersection color feature. The additional color set is the set obtained by removing the intersection color feature from the color set. For example, if the color sets of all target images contain blue and yellow, then the intersection color feature is blue and yellow, and the additional color set of the color set of red, blue, and yellow is red. Then, based on the intersection color feature, the images in the image library are preliminarily screened to obtain preliminary retrieved images, and all the colors in the intersection color feature are included in the preliminary retrieved images; immediately afterwards, the preliminary retrieved images are secondarily screened based on the additional color set to obtain secondarily retrieved images, and all the colors in a certain additional color set are included in the secondarily retrieved images. So far, only color features are used for retrieval, which is efficient and fast. Then, the embodiment of the present invention will stroke the colors in the intersection color feature and the additional color set in the secondarily retrieved images to obtain the retrieval contour of each color, and determine the number of color blocks corresponding to each color. Based on the number of color blocks and the contour feature, the similarity of the secondarily retrieved images is calculated, and the final retrieved images are output according to the calculation results. In this way, the similarity is calculated only based on the contours of the colors in the intersection color feature and the additional color set, which not only reduces the amount of calculation, but also excludes the interference of other colors, and introduces the number of color blocks, making the retrieval result more accurate.

[0071] As Figure 2 shown, as a preferred embodiment of the present invention, the step of performing feature recognition on each target image to obtain feature information specifically includes:

[0072] S101, recognizing the colors contained in the target image to obtain a color set;

[0073] S102, performing a stroke process on the edge of each color to obtain the contour feature of each color;

[0074] S103, determining the number of color blocks according to the number of contours in the contour feature of each color, and obtaining feature information according to the color set, the contour feature, and the number of color blocks.

[0075] In the embodiments of the present invention, in order to determine the feature information, first, the colors included in the target image are identified to obtain a color set, and then edge tracing is performed on the edges of each color. For example, edge picking is performed to obtain the contour features of each color. At the same time, the number of color blocks is determined according to the number of contours in the contour features of each color, so that the feature information of the target image is obtained.

[0076] As Figure 3 shown, as a preferred embodiment of the present invention, the step of performing a secondary screening on the preliminary retrieved image based on the additional color set to obtain a secondary retrieved image and classifying the secondary retrieved image according to the additional color set specifically includes:

[0077] S401, retrieve the additional color set corresponding to each feature information, where the additional color set contains zero, one or more additional colors;

[0078] S402, sort the additional color sets in descending order according to the number of types of additional colors;

[0079] S403, perform a secondary screening on the preliminary retrieved image according to the descending order to obtain a secondary retrieved image, and each secondary retrieved image corresponds to an additional color set;

[0080] S404, classify the secondary retrieved images according to the additional color sets, and each category corresponds to an additional color set.

[0081] In the embodiments of the present invention, before performing the secondary retrieval, it is necessary to determine the additional color set corresponding to each feature information. It is easy to understand that the additional color set may contain zero, one or more additional colors. Then, the additional color sets are sorted in descending order according to the number of types of additional colors. Next, the preliminary retrieved image is subjected to secondary screening according to the descending order to obtain a secondary retrieved image, and each secondary retrieved image corresponds to an additional color set, which facilitates subsequent classification, that is, classifying the secondary retrieved images according to the additional color sets, and each category corresponds to an additional color set, so that the feature information corresponding to each secondary retrieved image can be determined.

[0082] As Figure 4 shown, as a preferred embodiment of the present invention, the step of performing a secondary screening on the preliminary retrieved image according to the descending order to obtain a secondary retrieved image specifically includes:

[0083] S4031, determine the additional color set ranked first according to the descending order;

[0084] S4032, perform a secondary screening on the preliminary retrieved image according to the additional colors in the additional color set to obtain a secondary retrieved image corresponding to the additional color set;

[0085] S4033, removing the obtained secondary search image from the preliminary search image, and removing the extra color set from the descending order;

[0086] S4034, repeat steps S4031-S4033 until the secondary retrieval image is empty or the additional color set in descending order is empty.

[0087] In the embodiment of the present invention, when the secondary search is formally performed, the first additional color set is determined according to the descending order, and the primary search image is secondary screened according to the additional colors in the first additional color set to obtain a secondary search image corresponding to the additional color set; then the obtained secondary search image is removed from the primary search image, and the additional color set is removed from the descending order. Then the above search steps are repeated and looped until the secondary search image is empty or the additional color set in the descending order is empty, and the step is stopped.

[0088] like Figure 5 As shown in FIG. 1 , as a preferred embodiment of the present invention, the step of calculating the similarity of the secondary retrieval image based on the number of color blocks and the contour features specifically includes:

[0089] S601, determining feature information corresponding to the secondary search image, and screening the secondary search image based on the number of color blocks of the feature information;

[0090] S602, calculating the contour similarity between the search contour of each color in the filtered secondary search image and the contour feature of each color in the feature information;

[0091] S603, calculating the average value of the contour similarities of all colors to obtain a calculation result.

[0092] In an embodiment of the present invention, in order to further improve the retrieval efficiency and retrieval accuracy, the feature information corresponding to each secondary retrieval image is determined, the secondary retrieval image is screened based on the number of color blocks of the feature information, and then only the contour similarity between the retrieval contour of each color in the secondary retrieval image retained after screening and the contour feature of each color in the feature information is calculated. The contour similarity calculation can adopt methods such as Hausdorff distance and Fréchet distance, and a separate similarity calculation is performed for the contour of each color, eliminating interference from other factors as much as possible. For example, the retrieval contour includes three categories, namely, a red retrieval contour, a blue retrieval contour and a yellow retrieval contour, so that three contour similarities will be obtained. Finally, the average value of the contour similarities of the three colors is calculated to obtain a calculation result, and the final retrieval image is output according to the calculation result. For example, if the calculation result is higher than the set threshold, the corresponding secondary retrieval image will be regarded as the final retrieval image and thus output.

[0093] As shown in Figure 6 the figure, as a preferred embodiment of the present invention, the step of screening the secondary retrieval images based on the number of color blocks of the feature information specifically includes:

[0094] S6011, comparing the number of color blocks of each color in the feature information with the number of color blocks of the corresponding color in the secondary retrieval image;

[0095] S6012, when the number of color blocks of any color in the feature information > K × the number of color blocks of the corresponding color in the secondary retrieval image, where K is a fixed value greater than 1, the corresponding secondary retrieval image is deleted.

[0096] In the embodiment of the present invention, it is easy to understand that the finally retrieved images need to contain the target image, and the retrieved images contain more elements. Therefore, the number of color blocks of a certain color in the retrieved images should be greater than the number of color blocks of that color in the target image. Based on this, in the embodiment of the present invention, the number of color blocks of each color in the feature information will be compared with the number of color blocks of the corresponding color in the secondary retrieval image. When the number of color blocks of any color in the feature information > K × the number of color blocks of the corresponding color in the secondary retrieval image, it is determined that the corresponding secondary retrieval image does not meet the requirements and is deleted.

[0097] As shown in Figure 7 the figure, the embodiment of the present invention also provides an image information retrieval system based on local matching. The system includes:

[0098] A target image upload module 100, configured to receive a plurality of target images from different angles, perform feature recognition on each target image to obtain feature information, where the feature information includes a color set, each color in the color set corresponds to a contour feature, and each color corresponds to a number of color blocks;

[0099] An intersection color feature module 200, configured to perform an intersection process on the color sets of all target images to obtain intersection color features, and determine an additional color set corresponding to each feature information based on the intersection color features;

[0100] An image preliminary retrieval module 300, configured to perform preliminary screening on the images in the image library based on the intersection color features to obtain preliminary retrieval images;

[0101] An image secondary retrieval module 400, configured to perform secondary screening on the preliminary retrieval images based on the additional color set to obtain secondary retrieval images, and classify the secondary retrieval images according to the additional color set, and each category corresponds to feature information;

[0102] The secondary retrieval processing module 500 is configured to stroke the intersection color features in the secondary retrieval image and the colors in the additional color set to obtain the retrieval contour of each color, and determine the number of color blocks corresponding to each color;

[0103] The image final retrieval module 600 is configured to calculate the similarity of the secondary retrieval image based on the number of color blocks and the contour features, and output the final retrieval image according to the calculation result.

[0104] As a preferred embodiment of the present invention, the target image uploading module 100 includes:

[0105] An image color recognition unit, configured to recognize the colors included in the target image to obtain a color set;

[0106] A contour feature determination unit, configured to stroke the edge of each color to obtain the contour feature of each color;

[0107] A feature information determination unit, configured to determine the number of color blocks according to the number of contours in the contour feature of each color, and obtain feature information according to the color set, the contour feature, and the number of color blocks.

[0108] As a preferred embodiment of the present invention, the image secondary retrieval module 400 includes:

[0109] An additional color retrieval unit, configured to retrieve the additional color set corresponding to each feature information, where the additional color set includes zero, one, or more additional colors;

[0110] A set descending order arrangement unit, configured to arrange the additional color sets in descending order according to the number of types of additional colors;

[0111] A secondary screening image unit, configured to perform secondary screening on the preliminary retrieval image according to the descending order to obtain a secondary retrieval image, and each secondary retrieval image corresponds to an additional color set;

[0112] A retrieval image classification unit, configured to classify the secondary retrieval images according to the additional color sets, and each category corresponds to an additional color set.

[0113] As a preferred embodiment of the present invention, the image final retrieval module 600 includes:

[0114] A color block number screening unit, configured to determine the feature information corresponding to the secondary retrieval image, and screen the secondary retrieval image based on the number of color blocks of the feature information;

[0115] A contour similarity calculation unit, configured to calculate the contour similarity between the retrieval contour of each color in the screened secondary retrieval image and the contour feature of each color in the feature information;

[0116] A calculation result determination unit is configured to calculate an average value of the contour similarities of all colors to obtain a calculation result.

[0117] The above only describes the preferred embodiments of the present invention in detail, and is not intended to limit the present invention. Any modifications, equivalent replacements, and improvements made within the spirit and principles of the present invention shall be included within the protection scope of the present invention.

[0118] It should be understood that although the steps in the flowcharts of the embodiments of the present invention are shown in sequence according to the arrows, these steps do not necessarily have to be executed in the order indicated by the arrows. Unless there is a clear description in this article, the execution of these steps does not have a strict order limit, and these steps can be executed in other orders. Moreover, at least a part of the steps in each embodiment may include multiple sub-steps or multiple stages. These sub-steps or stages do not necessarily have to be executed at the same time, but can be executed at different times. The execution order of these sub-steps or stages does not necessarily have to be sequential, but can be executed alternately or alternately with at least a part of other steps or sub-steps or stages of other steps.

[0119] Those of ordinary skill in the art can understand that all or part of the processes of implementing the methods in the above embodiments can be completed by instructing relevant hardware through a computer program. The program can be stored in a non-volatile computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above methods. Among them, any reference to a memory, storage, database, or other medium used in the embodiments provided in the present application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDR SDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0120] Other embodiments of the present disclosure will be readily contemplated by those skilled in the art after considering the disclosure in the specification and the embodiments. This application is intended to cover any variations, uses, or adaptations of the present disclosure that follow the general principles of the present disclosure and include known common general knowledge or conventional technical means in the technical field not disclosed in the present disclosure. The specification and the embodiments are only regarded as exemplary, and the true scope and spirit of the present disclosure are pointed out by the claims.

Claims

1. A local matching-based image information retrieval method, characterized in that: The method comprises the following steps: Receive a number of target images at different angles, perform feature recognition on each target image, and obtain feature information, wherein the feature information includes a color set, each color in the color set corresponds to a contour feature, and each color corresponds to the number of color blocks; Perform intersection processing on the color sets of all target images to obtain intersection color features, and determine the additional color set corresponding to each feature information based on the intersection color features; Preliminarily screening the images in the image library based on the intersection color feature to obtain a preliminary retrieval image; Perform secondary screening on the preliminary retrieval image based on the additional color set to obtain a secondary retrieval image, and classify the secondary retrieval image according to the additional color set, and each category corresponds to feature information; The intersection color features in the secondary retrieval image and the colors in the additional color set are traced to obtain the retrieval outline of each color, and the number of color blocks corresponding to each color is determined; Calculate the similarity of the secondary retrieval image based on the number of color blocks and contour features, and output the final retrieval image according to the calculation results; Among them, the steps of performing secondary screening on the preliminary search image based on the additional color set to obtain the secondary search image, and classifying the secondary search image according to the additional color set specifically include: retrieving the additional color set corresponding to each feature information, wherein the additional color set contains zero, one or more additional colors; arranging the additional color set in descending order according to the number of types of additional colors; performing secondary screening on the preliminary search image in descending order to obtain the secondary search image, wherein each secondary search image corresponds to an additional color set; and classifying the secondary search image according to the additional color set, wherein each category corresponds to an additional color set.

2. The image information retrieval method based on local matching according to claim 1, characterized in that: The step of performing feature recognition on each target image to obtain feature information specifically includes: Identify the colors contained in the target image and obtain a color set; Stroke the edge of each color to obtain the contour features of each color; The number of color blocks is determined according to the number of contours in the contour feature of each color, and the feature information is obtained according to the color set, the contour feature and the number of color blocks.

3. The image information retrieval method based on local matching according to claim 1, characterized in that: The step of performing secondary screening on the preliminary search images according to the descending order to obtain the secondary search images specifically includes: Determine the additional set of colors that are ranked first according to the descending order; Performing a secondary screening on the preliminary search image according to the additional colors in the additional color set to obtain a secondary search image corresponding to the additional color set; Eliminating the obtained secondary retrieval image from the preliminary retrieval image, and excluding the additional color set from the descending order; Repeat the above three steps until the secondary retrieval image is empty or the additional color set in descending order is empty.

4. The image information retrieval method based on local matching according to claim 1, characterized in that: The step of calculating the similarity of the secondary retrieval image based on the number of color blocks and the contour features specifically includes: Determining feature information corresponding to the secondary search image, and screening the secondary search image based on the number of color blocks of the feature information; Calculating the contour similarity between the retrieval contour of each color in the screened secondary retrieval image and the contour feature of each color in the feature information; The average value of the contour similarities of all colors is calculated to obtain the calculation result.

5. The image information retrieval method based on local matching according to claim 4 is characterized in that: The step of screening the secondary search image based on the number of color blocks of the feature information specifically includes: Comparing the number of color blocks of each color in the feature information with the number of color blocks of the corresponding color in the secondary retrieval image; When the number of color blocks of any color in the feature information is greater than K×the number of color blocks of the corresponding color in the secondary retrieval image, where K is a constant value greater than 1, the corresponding secondary retrieval image is deleted.

6. An image information retrieval system based on local matching, characterized in that: The system comprises: The target image uploading module is used to receive a number of target images at different angles, perform feature recognition on each target image, and obtain feature information, wherein the feature information includes a color set, each color in the color set corresponds to a contour feature, and each color corresponds to the number of color blocks; An intersection color feature module is used to perform intersection processing on the color sets of all target images to obtain intersection color features, and determine the additional color set corresponding to each feature information based on the intersection color features; An image preliminary retrieval module, used for performing preliminary screening of images in an image library based on the intersection color feature to obtain a preliminary retrieval image; The image secondary retrieval module is used to perform secondary screening on the primary retrieval image based on the additional color set to obtain a secondary retrieval image, and classify the secondary retrieval image according to the additional color set, and each category corresponds to feature information; A secondary retrieval processing module is used to stroke the intersection color features in the secondary retrieval image and the colors in the additional color set to obtain the retrieval outline of each color, and determine the number of color blocks corresponding to each color; The image final retrieval module is used to calculate the similarity of the secondary retrieval image based on the number of color blocks and contour features, and output the final retrieval image according to the calculation results; Among them, the image secondary retrieval module includes: an additional color calling unit, which is used to call the additional color set corresponding to each feature information, and the additional color set contains zero, one or more additional colors; a set descending order arrangement unit, which is used to arrange the additional color set in descending order according to the number of types of additional colors; a secondary screening image unit, which is used to perform secondary screening on the preliminary retrieval image according to the descending order to obtain a secondary retrieval image, and each secondary retrieval image corresponds to an additional color set; a retrieval image classification unit, which is used to classify the secondary retrieval image according to the additional color set, and each category corresponds to an additional color set.

7. The image information retrieval system based on local matching according to claim 6, characterized in that: The target image uploading module comprises: An image color recognition unit, used for recognizing the colors contained in the target image and obtaining a color set; A contour feature determination unit, used for performing a stroke process on the edge of each color to obtain a contour feature of each color; The feature information determination unit is used to determine the number of color blocks according to the number of contours in the contour feature of each color, and obtain feature information according to the color set, the contour feature and the number of color blocks.

8. The image information retrieval system based on local matching according to claim 6, characterized in that: The image final retrieval module comprises: A color block quantity screening unit, used to determine feature information corresponding to the secondary search image, and screen the secondary search image based on the number of color blocks of the feature information; A contour similarity calculation unit, used to calculate the contour similarity between the retrieval contour of each color in the screened secondary retrieval image and the contour feature of each color in the feature information; The calculation result determination unit is used to calculate the average value of the contour similarities of all colors to obtain the calculation result.

Images