45results about How to "Improve binarization effect" patented technology

The invention provides a vehicle license plate recognition method based on a video. According to the vehicle license plate recognition method based on the video, moving vehicles are detected and separated out with the vehicle video which is obtained through actual photographing by means of a camera serving as input, the accurate position of a vehicle license plate area is determined by conducting vertical edge extraction on a target vehicle image obtained after pre-processing, a vehicle license plate image is separated out, color correction, binaryzation and inclination correction are conducted on a vehicle license plate image, each character in the positioned vehicle license plate area is separated to serve as an independent character, feature extraction is conducted one each character, obtained feature vectors are classified through a classifier which is well trained in advance, a classification result serves as a preliminary recognition result, secondary recognition is conducted on the stained vehicle license plate characters according to a template matching algorithm imitating the visual characteristics of human eyes, and then a final vehicle license plate recognition result is obtained. The vehicle license plate recognition method based on the video has the advantages that hardware cost is reduced, the management efficiency of an intelligent transportation system is improved, the anti-jamming performance and the robustness are high, the recognition efficiency is high, and the recognition speed is high.

The invention discloses a character recognition method for underwater video images. The method comprises the steps that the video images are subjected to preprocessing according to the morphological image processing principle, the contrast ratio is enhanced, and noise is filtered out; video characters are subjected to region segmentation according to preprocessing results, and text region location is carried out through a method combining edge detection and communication elements; according to video character region segmentation results, character segmentation is carried out on characters through a binary method combining a global threshold method and a local threshold method, and segmented characters are subjected to normalization to enable the segmented characters to be consistent with characters in a template base in size; the template base is designed according to the characteristics of video characters, the segmented characters and the characters in the template base are matched, and the characters can be recognized and stored in text.

The invention discloses an RMB sequence number identification method which comprises the following steps of: S1, performing pretreatment on a paper money image, including improvement of serious exposure, extraction of paper money image and registration of paper money image; S2, positioning the sequence number by a two-step method, namely the first step of approximate positioning using priori knowledge and the second step of accurate positioning of the sequence number; and performing character segmentation of the sequence number by use of a vertical projection method; and S3, performing multiple extraction for the characteristic value according to the characteristics of the confusable character by a new 13-point characteristic extraction method, and performing identification by use of a support vector machine according to the position-type relationship of the character to obtain relatively high identification accuracy. The method disclosed by the invention can improve the robustness of a sequence number identification system on the input paper money images different in angle, illumination, background and resolution, and improves the positioning and identification speed and the identification accuracy.

The invention provides an image character separation method, which is used to convert the image character detection results from a character area image to an OCR software recognizable binary image, and comprises following steps: (I) carrying out color component selection for the character area image; the characters are clearest among the selected color components; (II) carrying out binaryzation for the character area image on the base of the color component selected in step (I); (III) carrying out de-noising for the binary image derived in step (II). The invention can self-adaptive select the color component appropriate for binaryzation and thus achieving a better binaryzation effect; at the same time de-noising through a color-based clustering method can help acquiring a clearer binary character image with low noise and a better image character recognizing result.

The invention provides an identification method of a crown word number (serial number) of a bank note. The method comprises the following steps that: a crown word number (serial number) zone is localized on an RMB image under visible lights; a crown word number (serial number) identification method based on wavelet transformation local die minimal value and an improved crown word number (serial number) identification method supporting a vector machine are respectively used to independently identify a crown word number (serial number); and a weight-based discrimination method is used to carry out combination and decision on the two identification methods, so that a final identification result is obtained. According to the method, an RMB crown word number (serial number) can be identified rapidly and stably, so that it is beneficial to carry out monitoring and management on an RMB flow direction.

The invention discloses a template matching and optical flow method-based detection method for a power transmission line foreign matter of an unmanned aerial vehicle image, and belongs to the technical field of picture processing. The method comprises the steps of power transmission line video image collection, linear edge contrast enhancement, mean filtering, edge straight line feature extraction, power transmission line region setting, template matching, optical flow calculation, optical flow grayscale image binarization, feature point detection tracking and foreign matter identification, and the like. According to a feature that a power transmission line is approximate to a horizontal straight line, the linear edge contrast enhancement method and the edge straight line feature extraction method are proposed, so that an edge straight line of the power transmission line can be accurately extracted, and the adaptability and the edge straight line feature extraction are strong. According to a principle of large moving displacement of corresponding pixel points of an object relatively close to a camera in the image, a foreign matter region determination method is proposed by integrating template matching, optical flow calculation and image binarization, so that the accuracy is high and the robustness is strong. A foreign matter is tracked and identified through feature point matching of the foreign matter, so that the accuracy is high and the adaptability is strong.

The invention relates to a method for binarizing document images, comprising the following steps of: learning: selecting a predetermined number of document images from at least one document image to be binarized in same types through a predetermined first binaryzation algorithm as training samples for learning so as to obtain the attribute of the binary images which correspond to the training samples, wherein the obtained attribute istaken as the public reference attribute of binary images which correspond to the at least one document image to be binarized; and optimal binaryzation processing: carrying out the optimal binaryzation processing on each of the at least one document image to be binarized according to the public reference attribute through a predetermined second binaryzation algorithm so that the attribute of an obtained final binary image of each document image to be binarized is consistent with the public reference attribute. The invention also provides a device capable of executing the method and binarizing the document images and a document image processor assembled with the same. The method and the device can obtain more optimal binaryzation effects and enhance the stability of binaryzation quality.

The invention relates to the technical field of image recognition, and provides a digital instrument reading recognition method, which comprises the following steps of: obtaining a first image of a digital display area of a digital instrument; preprocessing the first image to obtain a grayed second image; carrying out edge detection on the second image, carrying out binary classification through amaximum between-class difference method, and removing redundant edge pixel points; performing binarization processing on the second image to obtain a third image; obtaining a non-background area of the third image, performing binary classification through a maximum inter-class difference method, and removing redundant non-background pixel points; performing tilt correction on the third image; andcarrying out character segmentation and numerical recognition on the corrected third image. According to the method, the influence of illumination on the recognition result can be reduced, characterbonding can be effectively avoided, over-segmented characters are subjected to merging processing, miscellaneous points are removed, the character segmentation precision is ensured, a large number ofsamples do not need to be collected for training, and the method is suitable for reading recognition of various digital instruments.

The invention discloses a target detection and identification method based on structural similarity measurement, and the method comprises the steps: predefining a template image to form a data set, reading a to-be-detected image, carrying out gray transformation, balancing the brightness of the to-be-detected image, obtaining a preprocessed to-be-detected image, solving an optimal segmentation threshold value of the preprocessed to-be-detected image, performing binarization processing on the to-be-detected image preprocessed in the step 2 by taking the optimal segmentation threshold as a global threshold to obtain a binary image of the preprocessed to-be-detected image; 3, performing morphological opening operation on the binary image in the step 3 to eliminate noisy points in the binary image, filling holes in the binary image to obtain an expansion result image, and marking candidate connected regions in the expansion result image as candidate targets and extracting all candidate targets; and matching the target template image in the step 1 with the candidate targets in the step 5 to obtain the structural similarity measurement of the candidate target, and taking the candidate target corresponding to the maximum structural similarity measurement as a final target.

The invention discloses a thermometer verification reading automatic identification system and method. The system comprises a camera (1), a computer (2), a stepping motor (3) and a temperature controlgroove (4). The computer (2) is connected with the camera (1) and the stepping motor (3) through wires, and an image acquisition card and a motion control card are installed in the computer (2). Thestepping motor (3) is fixed at the top of the temperature control groove (4); A constant-temperature liquid environment is arranged in the temperature control groove (4) and used for providing the temperature of a verification point, a thermometer reading window (5) is formed in the surface of the temperature control groove (4), a hole is formed in the top end of the temperature control groove (4), and a thermometer clamp capable of being taken out independently is arranged inside the temperature control groove. The method comprises the steps of image display and acquisition, image format conversion, image grayscale conversion, image segmentation, image binaryzation, image morphological operation, image feature extraction and reading recognition. Human eyes are replaced with computer recognition software, automation of thermometer reading is achieved, and therefore the manual labor amount is reduced, and the reading precision is improved.

The invention discloses a degraded document image binarization method and system based on deep learning. The system network comprises a first operation module, a second operation module and a binary classifier, the first operation module is used for degrading degraded document images into feature maps with different proportions and gradually reduced resolutions by using a shallow layer network, so that the change of text pixels on different feature levels is predicted. The detailed information of the previous layer is combined in a coarse-to-fine manner by using deconvolution to sequentially reconstruct foreground images. And the second operation module is cascaded with a deep network for secondary training after the shallow network structure, and is connected with a binary classifier at the end of the network structure for distinguishing background noise from foreground characters, so that the final binarization effect is optimized, and the precision and accuracy of binarization of the degraded document are greatly improved.

Embodiments of the invention provide a QR (Quick Response) two-dimensional code binarization method and system. The method comprises the steps of obtaining a grayscale image of a QR two-dimensional code, and obtaining a grayscale range of the grayscale image; dividing the grayscale image into a plurality of grayscale fuzzy regions according to the grayscale range; obtaining grayscale thresholds corresponding to the grayscale fuzzy regions respectively; and performing region binarization on the grayscale fuzzy regions according to the thresholds corresponding to the grayscale fuzzy regions. According to the method, a plurality of thresholds are selected for a grayscale entropy of the QR two-dimensional code image with wide grayscale, and the QR two-dimensional code image is binarized; and the binarization problem of the QR two-dimensional code under the condition of uneven illumination, illumination overexposure or the like is solved, a grayscale maximum entropy algorithm is improved, and a multi-threshold method is proposed for performing binarization, so that the binarization effect is improved.

The invention discloses an image binaryzation system based on an adaptive window and a smoothness threshold value method. The system comprises an image graying module, a smoothness threshold calculating module, an image binaryzation module and a special area searching module. The image graying module performs graying of an input image. The smoothness threshold calculating module calculates the smoothness threshold value of the image to generate a smoothness threshold value table. The image binaryzation module performs image binaryzation according to the smoothness threshold value table. The special area searching module searches a special area in the image according to a binary image. The system adopts the smoothness threshold value method to perform image binaryzation, has a good adaptability for images in different conditions, can obtain good binary effects for situations of special color characters, uneven illumination, noise influences, image blurring and low contrast of the image. The system ensures the binary effect of a common image and also can deal with situations of low resolution and complex illumination scenes well.

The invention discloses a Manchu historical document image binaryzation method based on a U-shaped convolutional neural network, which belongs to the technical field of image binaryzation, and comprises the steps of S1, dividing a Manchu historical document image into a plurality of sub-block images by adopting a block adaptive homomorphic filtering method, and performing homomorphic filtering onthe sub-block images to solve the problem of uneven image illumination; S2, binarizing the Manchu history document subjected to homomorphic filtering by using a U-shaped convolutional neural network.The method can enable the brightness of the image to be moderate, maintains the local details of the image, is higher in pixel recognition precision, can effectively remove a background part, maintains the text content, and is higher in flexibility and adaptability.

The invention provides a license plate image processing method and device. The method comprises the steps of obtaining a license plate image and a boundary formed due to brightness difference on the license plate image, wherein the license plate image is divided into upper and lower parts through the boundary; binarizing the upper and lower part license plate images; discretizing the boundary, thereby obtaining discrete points on the boundary; and binarizing pixels according to pixel values of the pixels in a preset neighbourhood range of a column in which the discrete points are located. According to the method, the parts in the license plate image with different brightness are binarized through adoption of different thresholds; compared a the mode of carrying out general binarization through adoption of a uniform threshold, the mode has the advantages that the obtained binarized image is beneficial for follow-up character extraction and identification, the binarization effect of vehicle light and shade plates is improved, and the identification accuracy of a vehicle tail plate is improved.

An embodiment of the invention discloses an image binarization method and device. The method comprises the following steps: segmenting a grayscale image to be processed into a plurality of image blocks, and calculating binarization threshold values of the plurality of image blocks respectively; carrying out interpolation calculation on the binarization threshold values of the plurality of image blocks respectively through a preset interpolation formula to obtain a binarization threshold value of each row or a binarization threshold value of each line of the plurality of image blocks; and carrying out binarization on the image blocks according to the binarization threshold value of each row or the binarization threshold value of each line of the plurality of image blocks to obtain corresponding binary images. According to the technical scheme, the binarization threshold value is set for each line or each row according to the preset interpolation formula, and pixel points are subjected to binarization according to the binarization threshold value of the line or row, so that a better binarization effect can be ensured; and meanwhile, the binarization threshold values cannot be calculated pixel point by pixel point, thereby reducing calculation amount, saving calculation time and improving binarization efficiency.

The invention relates to a license plate image binaryzation method based on a Laplacian extension operator, which is characterized by including the following steps: carrying out convolution operationon a license plate gray image by using a Laplacian extension operator, and acquiring positive and negative edge distribution features; acquiring a threshold T1 of positive edge distribution feature and a threshold T2 of negative edge distribution feature through an OTSU method, and generating a ternary image; and distinguishing between the characters and background of the license plate according to the pixel values of the points of the ternary image, and obtaining a binary image. The defects of the application of the traditional method based on gray binaryzation in special scenes are overcome.For a license plate image with shadow, local overexposure or the like, the integrity of the strokes can be fully retained, and the binaryzation effect of the license plate image can be improved.

The invention provides an image preprocessing method and device, electronic equipment and a readable storage medium. The method comprises the following steps: converting an original image into a grayscale image; performing brightness equalization and contrast improvement processing on the grayscale image to obtain a first image; performing enhancement processing on the first image to obtain a second image; and determining a connected domain corresponding to the character strokes in the second image, and obtaining a character region according to the connected domain. According to the method, an effective character segmentation result is obtained under the working conditions that illumination is not uniform, the background gray level is similar to the character gray level, and the background area is large.

The invention discloses a method for quickly positioning a damaged QR code according to image redundancy information. The method comprises the following steps: performing gray processing on a picturecontaining the QR code; carrying out definition evaluation on the grayscale picture; performing filtering and edge detection on the grayscale picture; determining three positioning patterns of the QRcode according to the width ratio of the stripes; performing adaptive binarization on the QR code region in combination with the size of the positioning pattern; and searching four edges of the QR code through affine transformation, and finally performing positioning correction by using homography transformation. According to the method for quickly positioning the damaged QR code, redundant information in the image is fully utilized, the image traversal frequency is reduced as much as possible, and the algorithm speed is increased; self-adaptive threshold edge detection and self-adaptive threshold binaryzation are adopted, so that the edge detection and binaryzation effects are improved; unclear images are filtered, and excessive waste of computing resources can be avoided.

The method is used for extracting an instant floating threshold of a binarization segmentation threshold level. It uses background sampling approach to select a segmentation threshold and comprises: along the row direction, extracting the level at rear edge of gate as the background level of current row; the said background level is taken as the segmentation threshold to segment the video of the said row.

The invention discloses an endoscope resolution recognition method, relates to the technical field of endoscope equipment, and provides an endoscope resolution image segmentation algorithm based on priori knowledge in order to avoid excessive segmentation or wrong segmentation under the condition that all levels of rhombus patterns in a resolution image are different in definition. According to the method, prior knowledge in a resolution plate standard is utilized, an original grey-scale map is subjected to corrosion operation firstly, connectivity between rhombic patterns is eliminated, and then cyclic binarization and contour detection operation is performed on a processed image; when the number of contours meeting the condition reaches a set prior value, an image with a good binarization effect is obtained, and meanwhile, region-of-interest extraction is carried out to obtain an independent rhombus pattern image. According to the algorithm, in the resolution image of the endoscope,relatively accurate extraction of the rhombus pattern is realized, and the accuracy of a detection result is improved.

The invention relates to a method for binarizing document images, comprising the following steps of: learning: selecting a predetermined number of document images from at least one document image to be binarized in same types through a predetermined first binaryzation algorithm as training samples for learning so as to obtain the attribute of the binary images which correspond to the training samples, wherein the obtained attribute istaken as the public reference attribute of binary images which correspond to the at least one document image to be binarized; and optimal binaryzation processing: carrying out the optimal binaryzation processing on each of the at least one document image to be binarized according to the public reference attribute through a predetermined second binaryzation algorithm so that the attribute of an obtained final binary image of each document image to be binarized is consistent with the public reference attribute. The invention also provides a device capable of executing the method and binarizing the document images and a document image processor assembled with the same. The method and the device can obtain more optimal binaryzation effects and enhance the stability of binaryzation quality.

The invention belongs to the technical field of insect identification. The invention discloses an insect identification method based on three-dimensional simulation, which is provided with a processorbox body, a camera photographing module is fixed in the middle of the processor box body through screws, a sound collecting module is installed beside the camera photographing module through screws,and a glass cover is installed on the surface of the processor box body through a rotating shaft. The camera shooting module is connected with a picture data memory, the picture data memory is connected with a picture binary processor, and the picture binary processor is connected with a picture feature sampler. The sound acquisition module is connected with the sound data memory, and the sound data memory is connected with the sound characteristic sampler. The picture feature sampler is coupled to a central processing unit which transfers data to a data comparison analyzer together with the sound feature sampler. By collecting and comparing the appearance pictures and sound information, we can accurately analyze the insect species, so that ordinary people can quickly grasp the important information of the insect.

The invention provides a license plate positioning and identification method. The method comprises the following steps of acquiring an original image snapshot by a camera; using a preset zoom coefficient to zoom the original image for many times so as to acquire a plurality of zoom images, and detecting a license plate area image in each zoom image; mapping the detected license plate area images to the original image, through filtering, merging and superposing images, according to a characteristic intensity of each license plate area image, ordering the license plate area images and acquiring all the license plate area images; carrying out downsampling on all the license plate area images for one time, carrying out graying, vertical edge detection and binaryzation on the downsampled images, and using vertical edge information to carry out license plate coarse positioning and acquiring all the license plate areas; correcting a gradient of the license plate areas; carrying out character segmentation on the license plate areas; and identifying segmented characters. Through the above mode, a license plate can be identified in a high-definition image.

The invention discloses a method and apparatus for checking a new-edition RMB banknote. The method comprises the steps of intercepting an area to be checked in a new-edition RMB banknote to be checked, and the area to be checked containing a second safety line feature area, conducting binary processing for the second safety line feature area, calculating the left and right boundaries according to a second safety line feature, matching a second safety line feature within a scope defined by left and right boundaries with a template, and determining the new-edition RMB banknote is authentic if the second safety line feature within a scope defined by left and right boundaries matches with the template. The method is low in algorithm complexity, and the time efficiency is high for the feature template matching algorithm. Selection of a binary threshold is set according to the proportion actually occupied by a black block, the binary effect is good, and the interference of noise is unlikely to occur. The method for checking black-and-white characteristics based on template matching ensures high accuracy.

Embodiments of the invention provide a QR (Quick Response) two-dimensional code binarization method and system. The method comprises the steps of obtaining a grayscale image of a QR two-dimensional code, and obtaining a grayscale range of the grayscale image; dividing the grayscale image into a plurality of grayscale fuzzy regions according to the grayscale range; obtaining grayscale thresholds corresponding to the grayscale fuzzy regions respectively; and performing region binarization on the grayscale fuzzy regions according to the thresholds corresponding to the grayscale fuzzy regions. According to the method, a plurality of thresholds are selected for a grayscale entropy of the QR two-dimensional code image with wide grayscale, and the QR two-dimensional code image is binarized; and the binarization problem of the QR two-dimensional code under the condition of uneven illumination, illumination overexposure or the like is solved, a grayscale maximum entropy algorithm is improved, and a multi-threshold method is proposed for performing binarization, so that the binarization effect is improved.

The invention provides an identification method of a crown word number (serial number) of a bank note. The method comprises the following steps that: a crown word number (serial number) zone is localized on an RMB image under visible lights; a crown word number (serial number) identification method based on wavelet transformation local die minimal value and an improved crown word number (serial number) identification method supporting a vector machine are respectively used to independently identify a crown word number (serial number); and a weight-based discrimination method is used to carry out combination and decision on the two identification methods, so that a final identification result is obtained. According to the method, an RMB crown word number (serial number) can be identified rapidly and stably, so that it is beneficial to carry out monitoring and management on an RMB flow direction.

The invention discloses a method for quickly locating a damaged QR code according to image redundancy information. The steps include: performing grayscale processing on a picture containing the QR code; performing clarity evaluation on the grayscale picture; and filtering the grayscale picture and edge detection; determine the three positioning patterns of the QR code according to the width ratio of the stripes; combine the size of the positioning pattern to adaptively binarize the QR code area; find the four edges of the QR code through affine transformation, and finally use the homography Transform for positioning correction. The method for quickly locating damaged QR codes makes full use of the redundant information in the image, reduces the number of image traversals as much as possible, and improves the algorithm speed; uses adaptive threshold edge detection and adaptive threshold binarization to improve edge detection and binarization effect; filtering unclear images can avoid excessive waste of computing resources.

The invention relates to a method for identifying a license plate in a high-definition image, which comprises steps of candidate license plate area detection, grayscale image binarization, obliquity correction, character segmentation, character identification and post-processing. Compared with the prior art, the method for identifying the license plate in the high-definition image has the advantages of simultaneous and quick positioning and identification of a plurality of license plates in a complex background with high definition (more than 5,000,000 pixels), good image binarization effect under the condition of non-uniform illumination, variability of license plate obliquity and size in a relatively larger range, and the like, and simultaneously greatly reduces the requirements on the installation of a video camera.