An agricultural industry Internet platform integrating agricultural materials
By detecting the confidence of text boxes in agricultural input packaging images and performing appropriate pre-processing and equalization processing, combined with the super-resolution reconstruction of the SRCNN model, the problem of poor image quality of agricultural input packaging images is solved, and the accuracy of text recognition is significantly improved and reliability.
Patent Information
- Application Number
- CN202410292620.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-03-14
- Publication Date
- 2024-11-05
- Estimated Expiration
- 2044-03-14
AI Technical Summary
In the prior art, the poor quality of agricultural materials packaging images leads to the problem of low text recognition accuracy.
The image acquisition module is used to obtain agricultural supplies packaging images and backup images, detect text frame confidence through the object detection model, perform pre-processing and histogram equalization processing, and super-resolution reconstruction is used to improve image quality and improve text recognition. accuracy.
It improves the accuracy and reliability of text recognition, saves computing resources, avoids unnecessary preprocessing and super-resolution reconstruction, and optimizes the quality of agricultural input packaging images.
Smart Images

Figure CN118097684B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of agricultural product information collection, and in particular to an agricultural industry Internet platform integrating agricultural resources. Background Art
[0002] In modern agricultural production, the application of agricultural materials (agricultural production materials) plays a vital role in improving crop yield and quality. Agricultural materials include fertilizers, pesticides, seeds, soil conditioners, etc. The effective application of these products can not only promote crop growth, but also effectively prevent and control pests and diseases and improve agricultural production efficiency. However, problems in the agricultural materials market, such as counterfeit and inferior products and information opacity, have seriously affected the safety and efficiency of agricultural production. In addition, the correct use and management of agricultural materials is also the key to improving agricultural production efficiency and ensuring crop quality.
[0003] With the development of information technology, the establishment of the agricultural industry Internet platform provides a new way to solve the above problems. Through the agricultural industry Internet platform, the transparency of agricultural input information can be achieved, the efficiency of the agricultural input supply chain can be improved, and technical support can be provided for the correct use and management of agricultural inputs. However, to achieve this goal, it is necessary to effectively collect, analyze and process agricultural input packaging images to ensure the accuracy and reliability of agricultural input information.
[0004] Chinese patent application publication number: CN107437187A discloses an Internet-based agricultural product information collection system, which consists of a two-dimensional code data generator, a main control computer, a cloud server and a mobile client; the two-dimensional code data generator is arranged in the main control computer, and images are collected by a video camera and transmitted to the main control computer, and a two-dimensional code is generated by the two-dimensional code data generator; the embedded main control computer is provided with a memory for storing two-dimensional code data information, a data line interface module, a network port and a display; the embedded main control computer is connected to the cloud server through the network port and the network cable, and the cloud server is built with a historical database, a real-time database and a data storage device, and the mobile client accesses the cloud server through the Internet.
[0005] It can be seen that the prior art has the following problems: due to the complex patterns on the agricultural supplies packaging and the influence of reflection, the quality of the collected agricultural supplies packaging images is poor, resulting in low accuracy in text recognition in the agricultural supplies packaging images. Summary of the invention
[0006] To this end, the present invention provides an integrated agricultural industry Internet platform for agricultural supplies, which is used to overcome the problem that the quality of agricultural supply packaging images collected in the prior art is poor, thereby resulting in low accuracy in text recognition in agricultural supply packaging images.
[0007] To achieve the above objectives, the present invention provides an agricultural industry Internet platform integrating agricultural resources, including:
[0008] An image acquisition module, which is used to acquire images of spare agricultural material packaging while acquiring images of agricultural material packaging;
[0009] An image acquisition unit, connected to the image acquisition module, for acquiring the agricultural material packaging image and the spare agricultural material packaging image;
[0010] An image analysis unit connected to the image acquisition unit is used to determine the eligibility of the text box detected in the agricultural supplies packaging image according to the confidence of the target detection model; determine the preprocessing of the agricultural supplies packaging image according to the first relative difference between the confidence and the preset confidence; determine the marking of the pixel points in the agricultural supplies packaging image according to the grayscale value of the pixel points; determine the histogram equalization processing of the marked pixel points in the text box according to the ratio of the number of marked pixel points in the qualified text box in the agricultural supplies packaging image to the number of pixel points in the text box;
[0011] An image preprocessing unit connected to the image analysis unit is used to preprocess the agricultural supplies packaging image by using a sharpening filter, and determine whether to use the SRCNN model to perform super-resolution reconstruction on the preprocessed agricultural supplies packaging image according to a second relative difference between the grayscale variance of the preprocessed agricultural supplies packaging image and the grayscale variance of the preprocessed agricultural supplies packaging image;
[0012] A character recognition unit connected to the image analysis unit, for recognizing the characters in the text box according to a character recognition model, and determining the eligibility of the recognition result according to the average confidence of all the characters recognized in the text box;
[0013] an adjustment unit, which is connected to the image acquisition unit and the text recognition unit respectively, and is used to determine the adjustment of the grayscale value of the pixel point after the histogram equalization processing according to the grayscale value difference between the pixel point after the histogram equalization processing and the pixel point before the histogram equalization processing;
[0014] The agricultural material information storage unit is connected to the text recognition unit and is used to store the text information on the agricultural material packaging image.
[0015] Further, the image analysis unit determines that the text box is qualified based on the comparison result that the confidence level is greater than a preset confidence level.
[0016] Furthermore, the image analysis unit determines to pre-process the agricultural supplies packaging image based on a comparison result that the first relative difference is less than or equal to a first preset relative difference.
[0017] Furthermore, the image preprocessing unit determines to use the SRCNN model to perform super-resolution reconstruction on the preprocessed agricultural supplies packaging image based on the comparison result that the second relative difference is less than or equal to the second preset relative difference.
[0018] Furthermore, the image analysis unit determines to mark the pixel point based on a comparison result that the grayscale value of the pixel point in the agricultural supplies packaging image is greater than a preset grayscale value.
[0019] Furthermore, the image analysis unit determines to perform histogram equalization processing on the marked pixels in the text box based on the comparison result that the ratio is greater than a preset ratio.
[0020] Furthermore, the text recognition unit determines that the recognition result is qualified based on the comparison result that the average confidence is greater than a preset average confidence.
[0021] Furthermore, the adjustment unit determines to adjust the grayscale value of the pixel point processed by the histogram equalization based on the comparison result that the grayscale value difference is less than a preset grayscale value difference.
[0022] Furthermore, the adjustment unit adjusts the grayscale value of the pixel after the histogram equalization process according to the following adjustment coefficient, setting:
[0023]
[0024] Wherein, H represents the adjustment coefficient, ΔC represents the gray value difference, and C1 represents the gray value of the pixel before the histogram equalization processing.
[0025] Furthermore, the adjustment unit replaces the pixel points processed by the histogram equalization with the pixel points in the spare agricultural supplies packaging image.
[0026] Compared with the prior art, the beneficial effect of the present invention lies in that the present invention determines whether the text box is qualified by comparing the confidence of the text box detected by the target detection model in the agricultural supplies packaging image with the preset confidence, so as to pre-process the text box with low confidence, reduce the situation of text misrecognition, and improve the accuracy of text recognition.
[0027] Furthermore, the present invention determines whether to preprocess the agricultural input packaging image based on the comparison result of the first relative difference and the first preset relative difference, so as to more accurately decide which agricultural input packaging images need to be preprocessed to improve the accuracy of text recognition, and avoid preprocessing agricultural input packaging images that are difficult to achieve satisfactory recognition effects even after processing, thereby avoiding waste of computing resources.
[0028] Furthermore, the present invention determines whether to use the SRCNN model to perform super-resolution reconstruction on the preprocessed agricultural packaging image based on the comparison result of the second relative difference and the second preset relative difference, so as to judge whether the quality of the agricultural packaging image is substantially improved, and thus decides whether it is necessary to further use the SRCNN model for super-resolution reconstruction, thereby optimizing the quality of the final agricultural packaging image and saving computing resources.
[0029] Furthermore, the present invention determines whether to mark the pixel points based on the comparison result of the grayscale value of the pixel points in the agricultural supplies packaging image with the preset grayscale value to mark the high-brightness area in the agricultural supplies packaging image, that is, the area where reflection occurs during the process of collecting the agricultural supplies packaging image, so as to process the high-brightness area in the subsequent processing process to reduce interference and improve the accuracy of subsequent recognition.
[0030] Furthermore, the present invention determines whether to perform histogram equalization processing on the marked pixels in the text box based on a comparison result of the ratio of the number of marked pixels in a qualified text box in the agricultural supplies packaging image to the number of pixels in the text box and a preset ratio, so as to significantly improve the readability of the text and make text recognition more accurate, while avoiding histogram equalization processing on the part outside the text box to save computing resources.
[0031] Furthermore, the present invention determines whether the recognition result is qualified based on the comparison result of the average confidence level of all the text in the text box being recognized with a preset average confidence level to comprehensively judge the effect of text recognition, thereby reducing the possibility of erroneous recognition and helping to improve the reliability and accuracy of the final recognition result.
[0032] Furthermore, the present invention determines whether to adjust the grayscale value of the pixel point after the histogram equalization processing according to the comparison result of the grayscale value difference and the preset grayscale value difference, and can adjust the part with poor processing effect on the high-brightness area again to improve the quality of the agricultural supplies packaging image and improve the accuracy of text recognition.
[0033] Furthermore, the present invention accurately adjusts the grayscale values of pixels after histogram equalization processing through a formula to avoid excessive adjustment leading to image quality degradation. At the same time, the parts that cannot continue to be adjusted by grayscale value are directly replaced, thereby improving processing efficiency and saving computing resources. BRIEF DESCRIPTION OF THE DRAWINGS
[0034] Figure 1 This is a schematic diagram of the structure of the agricultural industry Internet platform integrating agricultural resources according to an embodiment of the present invention;
[0035] Figure 2 It is a workflow diagram of the agricultural industry Internet platform integrating agricultural resources under the qualified condition of the text box according to the embodiment of the present invention;
[0036] Figure 3 This is a workflow diagram of the agricultural industry Internet platform with integrated agricultural resources according to an embodiment of the present invention under the condition that the text box is unqualified. DETAILED DESCRIPTION
[0037] In order to make the objects and advantages of the present invention more clearly understood, the present invention is further described below in conjunction with 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.
[0038] The preferred embodiments of the present invention are described below with reference to the accompanying drawings. It should be understood by those skilled in the art that these embodiments are only used to explain the technical principles of the present invention and are not intended to limit the protection scope of the present invention.
[0039] In addition, it should be noted that in the description of the present invention, unless otherwise clearly specified and limited, the terms "installed", "connected", and "connected" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two components. For those skilled in the art, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.
[0040] See also Figure 1-3 As shown, Figure 1 This is a schematic diagram of the structure of the agricultural industry Internet platform integrating agricultural resources according to an embodiment of the present invention; Figure 2 It is a workflow diagram of the agricultural industry Internet platform integrating agricultural resources under the qualified condition of the text box according to the embodiment of the present invention; Figure 3 This is a workflow diagram of the agricultural industry Internet platform with integrated agricultural resources according to an embodiment of the present invention under the condition that the text box is unqualified.
[0041] The embodiment of the present invention discloses an agricultural industry Internet platform integrating agricultural materials, including:
[0042] An image acquisition module, which is used to acquire images of spare agricultural material packaging while acquiring images of agricultural material packaging;
[0043] An image acquisition unit, connected to the image acquisition module, for acquiring the agricultural material packaging image and the spare agricultural material packaging image;
[0044] An image analysis unit connected to the image acquisition unit, for determining the eligibility of the text box according to the confidence X of the text box detected by the target detection model in the agricultural supplies packaging image;
[0045] Determine the preprocessing of the agricultural supplies packaging image according to a first relative difference △X between the confidence X and a preset confidence X0;
[0046] Determine the mark of the pixel point according to the gray value L of the pixel point in the agricultural supplies packaging image;
[0047] Determine the histogram equalization processing of the marked pixels in the text box according to the ratio M of the number of marked pixels in the qualified text box in the agricultural supplies packaging image to the number of pixels in the text box;
[0048] An image preprocessing unit connected to the image analysis unit is used to preprocess the agricultural supplies packaging image by using a sharpening filter, and determine whether to use the SRCNN model to perform super-resolution reconstruction on the preprocessed agricultural supplies packaging image according to a second relative difference ΔF between a grayscale variance F2 of the preprocessed agricultural supplies packaging image and a grayscale variance F1 of the preprocessed agricultural supplies packaging image;
[0049] A text recognition unit connected to the image analysis unit, for recognizing the text in the text box according to a text recognition model, and determining the eligibility of the recognition result according to an average confidence level P of all the text in the text box;
[0050] An adjustment unit, which is connected to the image acquisition unit and the text recognition unit respectively, and is used to determine the adjustment of the grayscale value of the pixel point after the histogram equalization processing according to the grayscale value difference △C between the pixel point after the histogram equalization processing and the pixel point before the histogram equalization processing;
[0051] The agricultural material information storage unit is connected to the text recognition unit and is used to store the text information on the agricultural material packaging image.
[0052] In the embodiment of the present invention, the image acquisition module is preferably a mobile phone camera, the target detection model is preferably YOLOV7, the text recognition model is preferably Google Vision OCR, and the agricultural supplies packaging images include but are not limited to fertilizer packaging, pesticide packaging, seed packaging, and soil conditioner packaging.
[0053] Specifically, the image analysis unit determines whether the text box is qualified by comparing the confidence X of the text box detected by the target detection model in the agricultural supplies packaging image with the preset confidence X0;
[0054] If X≤X0, the image analysis unit determines that the text box is unqualified;
[0055] If X>X0, the image analysis unit determines that the text box is qualified;
[0056] In the embodiment of the present invention, the preset reliability X0 is set to 0.9, and those skilled in the art may adjust the preset reliability X0 according to specific circumstances.
[0057] Specifically, the present invention determines whether the text box is qualified by comparing the confidence of the text box detected by the target detection model in the agricultural supplies packaging image with the preset confidence, so as to pre-process the text box with low confidence, reduce the situation of text misrecognition, and improve the accuracy of text recognition.
[0058] Specifically, when the text box is unqualified, the image analysis unit calculates a first relative difference △X between the confidence X and a preset confidence X0, and determines whether to pre-process the agricultural supplies packaging image according to a comparison result of the first relative difference △X and a first preset relative difference △X0, and sets △X=(X0-X) / X0;
[0059] If △X≤△X0, the image analysis unit determines to pre-process the agricultural supplies packaging image;
[0060] If △X>△X0, the image analysis unit determines not to pre-process the agricultural supplies packaging image and reports an error on the agricultural supplies packaging image;
[0061] The image preprocessing unit preprocesses the agricultural supplies packaging image by using a sharpening filter under the agricultural supplies packaging image preprocessing condition determined by the image analysis unit.
[0062] In the embodiment of the present invention, the first preset relative difference △X0 is set to 0.33. The first preset relative difference △X0 is obtained when the confidence X is 0.6. Those skilled in the art can adjust the first preset relative difference △X0 according to specific circumstances.
[0063] Specifically, the present invention determines whether to preprocess the agricultural input packaging image based on the comparison result of the first relative difference and the first preset relative difference, so as to more accurately decide which agricultural input packaging images need to be preprocessed to improve the accuracy of text recognition, and avoid preprocessing agricultural input packaging images that are difficult to achieve satisfactory recognition effects even after processing, thereby avoiding waste of computing resources.
[0064] Specifically, the image preprocessing unit calculates a second relative difference △F between the grayscale variance F2 of the preprocessed agricultural material packaging image and the grayscale variance F1 of the preprocessed agricultural material packaging image under the condition that the preprocessing of the agricultural material packaging image is completed, and determines whether to use the SRCNN model to perform super-resolution reconstruction on the preprocessed agricultural material packaging image according to the comparison result of the second relative difference △F and the second preset relative difference △F0, and sets △F=(F2-F1) / F1;
[0065] If △F≤△F0, the image preprocessing unit determines to use the SRCNN model to perform super-resolution reconstruction on the preprocessed agricultural supplies packaging image;
[0066] If △F>△F0, the image preprocessing unit determines not to use the SRCNN model to perform super-resolution reconstruction on the preprocessed agricultural supplies packaging image;
[0067] In the embodiment of the present invention, the second preset relative difference ΔF0 has a value range of 0-0.3, and the second preset relative difference ΔF0 preferably has a value of 0.1. Those skilled in the art may adjust the second preset relative difference ΔF0 according to specific circumstances.
[0068] Specifically, the present invention determines whether to use the SRCNN model to perform super-resolution reconstruction on the preprocessed agricultural packaging image according to the comparison result of the second relative difference and the second preset relative difference, so as to judge whether the quality of the agricultural packaging image is substantially improved, and thus decides whether it is necessary to further use the SRCNN model for super-resolution reconstruction, thereby optimizing the quality of the final agricultural packaging image and saving computing resources.
[0069] Specifically, the image analysis unit determines whether to mark the pixel point according to the comparison result between the gray value L of the pixel point in the agricultural supplies packaging image and the preset gray value L0, when determining that the text box is qualified or the agricultural supplies packaging image is preprocessed or the preprocessed agricultural supplies packaging image is super-resolution reconstructed using the SRCNN model;
[0070] If L≤L0, the image analysis unit determines not to mark the pixel point;
[0071] If L>L0, the image analysis unit determines to mark the pixel point;
[0072] In the embodiment of the present invention, the preset grayscale value L0 ranges from 0.9 to 1 of the maximum grayscale value of the image, and the preset grayscale value L0 is preferably 0.98. Those skilled in the art may adjust the preset grayscale value L0 according to specific circumstances.
[0073] Specifically, the present invention determines whether to mark the pixel points according to the comparison result of the grayscale value of the pixel points in the agricultural supplies packaging image with the preset grayscale value to mark the high-brightness area in the agricultural supplies packaging image, that is, the area where reflection occurs during the process of collecting the agricultural supplies packaging image, so as to process the high-brightness area in the subsequent processing process to reduce interference and improve the accuracy of subsequent recognition.
[0074] Specifically, the image analysis unit determines whether to perform histogram equalization processing on the marked pixels in the text box according to a comparison result of a ratio M of the number of marked pixels in the qualified text box in the agricultural supplies packaging image to the number of pixels in the text box and a preset ratio M0;
[0075] If M≤M0, the image analysis unit determines not to perform histogram equalization processing on the marked pixel points in the text box;
[0076] If M>M0, the image analysis unit determines to perform histogram equalization processing on the marked pixels in the text box;
[0077] In the embodiment of the present invention, 0.1≤M0≤0.5, and the preset ratio M0 is preferably 0.3. Those skilled in the art may adjust the preset ratio M0 according to specific circumstances.
[0078] Specifically, the present invention determines whether to perform histogram equalization processing on the marked pixels in the text box according to the comparison result of the ratio of the number of marked pixels in the qualified text box in the agricultural supplies packaging image to the number of pixels in the text box and a preset ratio, so as to significantly improve the readability of the text and make text recognition more accurate, while avoiding histogram equalization processing on the part outside the text box to save computing resources.
[0079] Specifically, the text recognition unit recognizes the text in the text box under the condition that the image analysis unit completes the histogram equalization processing on the marked pixels in the text box or the image analysis unit determines not to perform the histogram equalization processing on the marked pixels in the text box, and determines whether the recognition result is qualified according to the comparison result of the average confidence P of all the characters recognized in the text box and the preset average confidence P0;
[0080] If P≤P0, the character recognition unit determines that the recognition result is unqualified;
[0081] If P>P0, the character recognition unit determines that the recognition result is qualified;
[0082] In the embodiment of the present invention, the preset average confidence level P0 is set to 0.9, and those skilled in the art may adjust the preset average confidence level P0 according to specific circumstances.
[0083] Specifically, the present invention determines whether the recognition result is qualified based on the comparison result of the average confidence level of all the text in the text box being recognized with the preset average confidence level to comprehensively judge the effect of text recognition, reduces the possibility of erroneous recognition, and helps to improve the reliability and accuracy of the final recognition result.
[0084] Specifically, the adjustment unit calculates a grayscale value difference △C between a pixel after the histogram equalization process and a pixel before the histogram equalization process under the condition that the text recognition unit determines that the recognition result is unqualified, and determines whether to adjust the grayscale value of the pixel after the histogram equalization process according to a comparison result of the grayscale value difference △C and a preset grayscale value difference △C0, and sets △C=C1-C2, wherein C2 represents the grayscale value of the pixel after the histogram equalization process, and C1 represents the grayscale value of the pixel before the histogram equalization process;
[0085] If △C≤△C0, the adjustment unit determines to adjust the grayscale value of the pixel point processed by the histogram equalization;
[0086] If △C>△C0, the adjustment unit determines not to adjust the grayscale value of the pixel point processed by the histogram equalization;
[0087] In the embodiment of the present invention, the preset gray value difference △C0 is 10% of the gray value of the pixel before the histogram equalization process. Those skilled in the art may adjust the preset gray value difference △C0 according to specific circumstances.
[0088] Specifically, the present invention determines whether to adjust the grayscale value of the pixel point after the histogram equalization processing according to the comparison result of the grayscale value difference and the preset grayscale value difference, and can adjust the part with poor processing effect on the high-brightness area again to improve the quality of the agricultural supplies packaging image and improve the accuracy of text recognition.
[0089] Specifically, the adjustment unit adjusts the grayscale value of the pixel after the histogram equalization process according to the following adjustment coefficient H, setting:
[0090]
[0091] The adjusted grayscale value of the pixel point processed by the histogram equalization is set to Cc2=C2×H.
[0092] Specifically, the adjustment unit replaces the pixel points processed by the histogram equalization with the pixel points in the spare agricultural supplies packaging image under the condition that it is determined that the grayscale value of the pixel points processed by the histogram equalization is not adjusted.
[0093] In the embodiment of the present invention, the spare agricultural material packaging image is an agricultural material packaging image whose ISO is 50% of the agricultural material packaging image and is acquired at the same time as the agricultural material packaging image.
[0094] Specifically, the present invention accurately adjusts the grayscale values of pixels after histogram equalization processing through a formula to avoid excessive adjustment that may lead to a decrease in image quality. At the same time, the parts that cannot continue to be adjusted by grayscale value are directly replaced, thereby improving processing efficiency and saving computing resources.
[0095] So far, the technical solutions of the present invention have been described in conjunction with the preferred embodiments shown in the accompanying drawings. However, it is easy for those skilled in the art to understand that the protection scope of the present invention is obviously not limited to these specific embodiments. Without departing from the principle of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will fall within the protection scope of the present invention.
[0096] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.
Claims
1. An agricultural industry Internet platform integrating agricultural resources, characterized in that: include: An image acquisition module, which is used to acquire images of spare agricultural material packaging while acquiring images of agricultural material packaging; An image acquisition unit, connected to the image acquisition module, for acquiring the agricultural material packaging image and the spare agricultural material packaging image; An image analysis unit connected to the image acquisition unit, for determining the eligibility of the text box according to the confidence of the text box detected by the target detection model in the agricultural supplies packaging image; Determine preprocessing of the agricultural supplies packaging image according to a first relative difference between the confidence and a preset confidence; Determine a mark for a pixel point according to a grayscale value of the pixel point in the agricultural supplies packaging image; Determine the histogram equalization processing of the marked pixels in the text box according to the ratio of the number of marked pixels in the qualified text box in the agricultural supplies packaging image to the number of pixels in the text box; An image preprocessing unit connected to the image analysis unit is used to preprocess the agricultural supplies packaging image by using a sharpening filter, and determine whether to use the SRCNN model to perform super-resolution reconstruction on the preprocessed agricultural supplies packaging image according to a second relative difference between the grayscale variance of the preprocessed agricultural supplies packaging image and the grayscale variance of the preprocessed agricultural supplies packaging image; A character recognition unit connected to the image analysis unit, for recognizing the characters in the text box according to a character recognition model, and determining the eligibility of the recognition result according to the average confidence of all the characters recognized in the text box; An adjustment unit, which is connected to the image acquisition unit and the text recognition unit respectively, and is used to determine the adjustment of the grayscale value of the pixel after the histogram equalization processing according to the grayscale value difference between the pixel after the histogram equalization processing and the pixel before the histogram equalization processing, specifically including: determining to adjust the grayscale value of the pixel after the histogram equalization processing based on the comparison result that the grayscale value difference is less than the preset grayscale value difference, and replacing the pixel after the histogram equalization processing with the pixel in the spare agricultural material packaging image based on the comparison result that the grayscale value difference is greater than or equal to the preset grayscale value difference; The grayscale value of the pixel point processed by the histogram equalization after adjustment is calculated according to the following formula: Cc2=C2×H, where Cc2 is the grayscale value of the pixel processed by histogram equalization after adjustment, C2 is the grayscale value of the pixel processed by histogram equalization, and H is the adjustment coefficient; The adjustment coefficient is calculated according to the following formula, Wherein, H represents the adjustment coefficient, △C represents the gray value difference between the pixel point after the histogram equalization processing and the pixel point before the histogram equalization processing, and C1 represents the gray value of the pixel point before the histogram equalization processing; The agricultural material information storage unit is connected to the text recognition unit and is used to store the text information on the agricultural material packaging image.
2. The agricultural industry Internet platform for integrated agricultural resources according to claim 1 is characterized in that: The image analysis unit determines that the text box is qualified based on the comparison result that the confidence level is greater than a preset confidence level.
3. The agricultural industry Internet platform integrating agricultural resources according to claim 2 is characterized in that: The image analysis unit determines to pre-process the agricultural supplies packaging image based on a comparison result that the first relative difference is less than or equal to a first preset relative difference.
4. The agricultural industry Internet platform integrating agricultural resources according to claim 3 is characterized in that: The image preprocessing unit determines to use the SRCNN model to perform super-resolution reconstruction on the preprocessed agricultural supplies packaging image based on the comparison result that the second relative difference is less than or equal to the second preset relative difference.
5. The agricultural industry Internet platform integrating agricultural resources according to claim 4 is characterized in that: The image analysis unit determines to mark the pixel point based on a comparison result that the grayscale value of the pixel point in the agricultural supplies packaging image is greater than a preset grayscale value.
6. The agricultural industry Internet platform integrating agricultural resources according to claim 5 is characterized in that: The image analysis unit determines to perform histogram equalization processing on the marked pixels in the text box based on the comparison result that the ratio is greater than a preset ratio.
7. The agricultural industry Internet platform integrating agricultural resources according to claim 6 is characterized in that: The text recognition unit determines that the recognition result is qualified based on the comparison result that the average confidence level is greater than a preset average confidence level.
8. The agricultural industry Internet platform integrating agricultural resources according to claim 7 is characterized in that: The adjustment unit determines to adjust the grayscale value of the pixel point processed by the histogram equalization based on the comparison result that the grayscale value difference is less than a preset grayscale value difference.
