63 results about "Otsu's method" patented technology

The invention discloses a method for detecting lane change of a vehicle based on a vehicle-mounted camera. The method comprises the following steps: firstly, initializing a read-in image, and converting the read-in image to a gray space; secondly, segmenting an sky region and a ground region of the image, and acquiring the image of the ground region; thirdly, carrying out edge detection by utilizing the sobel operator; fourthly, carrying out binaryzation by utilizing an Otsu's method; fifthly, restricting a fitting range, wherein hough transformation is restricted by the minimum fitting points, and extracting a lane line equation; sixthly, determining the type of a lane line; seventhly, classifying lane line treatment results; and eighthly, determining the current lane change situation ofthe vehicle. According to the invention, the current lane change situation of the vehicle is detected by adopting the mode of images, the existing vehicle-mounted camera of a driving school is utilized, has low cost, high practicability, diversity detection data, and has accurate and objective evaluations of lane change level for students, the equipment is simple to install, and has wide applications.

The invention discloses a method adopting image segmentation for automatically testing the LED indicator light of automobile instruments and provides a region-growing method based on the maximum between-cluster variance (Otsu's Method). The method comprises the following steps: automatically selecting the centroid in the light-emitting region of the LED light as the seed point to carry out the region-growing; realizing the region segmentation and extraction; and judging whether the brightness and color of the light are correct upon the seed point and the area of the region. The invention can effectively improve the testing accuracy rate and testing efficiency of the LED indicator light of the automobile instruments and provide effective means for the automatic production of the automobile instruments.

The invention discloses a rural resident point information extraction method based on a high-resolution remote-sensing image. The rural resident point information extraction method comprises the following steps: firstly, carrying out Canny edge detection on the high-resolution remote-sensing image, and obtaining a vector edge line through chain code tracking; then, carrying out Douglas-Peucker abbreviation on a vector edge to extract straight-line segments, and rejecting shorter straight-line segments and keeping longer straight-line segments according to length constraints; secondly, according to adjacency relation between the straight-line segments and included angle constraints, extracting straight angle points, calculating the density characteristics of the straight angle points within a certain space region range, and generating a straight angle point density characteristic image, wherein the characteristic image and the input high-resolution remote-sensing image have same spatial range and spatial resolution; and finally, carrying out binarization processing on the density characteristic image by utilizing Otsu, extracting rural resident point vector pattern spots through connected component analysis so as to extract rural resident point information. Precision and effects are improved, and robustness and universality are better.

The invention discloses a lane departure warning method based on Hough transform. The method comprises the steps of carrying out region of interest segmentation on an image; graying region of interestimages; carrying out edge detection on the grayed images through utilization of an improved unidirectional gradient operator, thereby obtaining edge images; carrying out binarization processing on the images through utilization of an otsu algorithm; obtaining a candidate straight line set through Hough transform, dividing the candidate straight line set into a left lane straight line set and a right lane straight line set according to slopes, screening the two straight line sets through combination of vanishing point constraint and a random sample consistency algorithm, thereby obtaining optimum left and right lane line parameters; establishing two Kalman filters to track left and right optimum lane lines; and estimating relative lateral offset and a change trend through utilization of the left and right lane line slopes, identifying lane departure and carrying out warning. According to the method, the lane departure problem in driving can be effectively identified, and the robustnessand timeliness of the a lane departure warning system are improved.

The invention relates to a PCB labeling printing quality detection method based on deep learning. The method comprises the following steps of carrying out a series of pretreatment such as Gaussian filtering, conversion into a gray scale image, edge detection, hole filling and optimal threshold binarization segmentation on an obtained PCB original image so as to effectively improve recognition detection efficiency; adopting an Otsu algorithm for carrying out optimal threshold segmentation on the image to obtain a binary image, so that the test precision of the method is improved; effectively collecting and expanding image sample data to increase representativeness of samples. The model overfitting phenomenon caused by few data is prevented. A model obtained through training is affected by irrelevant factors as little as possible. Robustness of the model is enhanced, training time of a deep learning network is shorter, and the convergence speed is higher. Target features of the image areextracted efficiently through the deep neural network, and the defects of easy overfitting, long training time, high parameter adjustment difficulty and the like caused by a traditional artificial neural network feature extraction method can be effectively avoided.

The invention discloses an impurity detection method based on image processing. The method comprises steps: a cotton sample image is obtained; the cotton sample image is enhanced to obtain cotton enhancement image; according to HSV, the cotton enhancement image is segmented by threshold, and the background color of the cotton enhancement image is removed to obtain the target cotton image; according to a Otsu algorithm, the cotton image is segmented by threshold, and the binary image of the target cotton image is obtained; an impurity area in cotton is calculated according to a target cotton image and a binary image of the target cotton image; a connected region of a binary graph of a target cotton image is determined, and an impurity position and an impurity number in the binary graph of the target cotton image are identified according to the connected region; an L-M algorithm and S-component histogram of HSV color space are used to train cotton and background threshold and stray areaadaptive BP network, and the dynamic threshold region of background is obtained.

The invention discloses a two-dimension code preprocessing method which comprises the following steps of luminance equalization: performing self-adaptation brightness equalization processing on the captured images to remove a too faint part and a too bright part in each image; binarization processing: performing binarization processing on the images subjected to the self-adaptation brightness equalization processing through an Otsu's method. The invention further discloses a two-dimension code preprocessing device which comprises a luminance equalization processing module for performing self-adaptation brightness equalization processing on the captured images to remove the too faint part and the too bright part in each image and a binarization processing module for performing binarization processing on the images subjected to the self-adaptation brightness equalization processing through the Otsu's method. By means of the two-dimension code preprocessing method and device, effect caused by uneven illumination in the bar code images is eliminated effectively, binarization can be performed on the bar code images with uneven illumination correctively, and the recognition rate of two-dimension codes is increased.

The embodiment of the invention discloses a strabismus detection method based on a cascade convolutional neural network. The method comprises the steps that the strabismus images photographed by the camera are collated and a strabismus image library of the images is established; the cascade convolutional neural network is trained by using the face database iBUG23, LPFW24, Helen25 and AFW, and thelearning parameters in the cascade convolutional neural network are determined; the eyes of the strabismus images in the strabismus image library are segmented by using the completely trained cascadeconvolutional neural network; the eye iris of the strabismus images is segmented by using the Otsu algorithm after completing eye segmentation of the strabismus images; and whether the person has thestrabismus is determined according to the relative position relation of the iris in the eyes. The high recognition and segmentation capacity of the cascade convolutional neural network is fully utilized, whether the person has the strabismus can be efficiently and accurately determined through combination of the image processing algorithm and thus strabismus diagnosis and treatment of the patientcan be facilitated for the doctor.

The invention discloses a method of extracting substation information in a power distribution system based on hyperspectral remote sensing images. The method comprises the steps of firstly classifying hyperspectral remote sensing images using a decision tree algorithm, and then conducting edge detection of the classified hyperspectral remote sensing images to obtain vector edge lines; secondly, extracting straight line segments from the vector edges, removing the shorter straight line segments and retaining the longer ones; thirdly, extracting right angle points according to the adjacency relation among the straight line segments and included angle constraints, conducting statistics on the density features of the right angle points in a certain spatial region, and generating a right angle point density feature image; and fourthly, binarizing the density feature image using an otsu method, extracting substation vector pattern spots through connect component analysis, and thus achieving extraction of the substation information. The invention solves the classification problem of hyperspectral remote sensing images, satisfies the requirements for the extraction of the substation information in the hyperspectral remote sensing images, and improves the robustness and universality in extraction of the substation information.

The invention discloses a video intrusion detection method used for a power line external force damage-prevention system. The method comprises the following steps that: an Otsu's method is utilized to acquire difference image binaryzation segmentation results of current video image information; an improved SURENDRA algorithm is utilized to build a background model of a scene based on the difference image binaryzation segmentation results; and a foreground movement region is generated based on the built background model, and at the same time, the background model is updated. With the video intrusion detection method used for the power line external force damage-prevention system of the invention, defects of low processing efficiency, poor timeliness and high misoperation rate can be eliminated. The video intrusion detection method used for the power line external force damage-prevention system of the invention has the advantages of high processing efficiency, excellent timeliness and low misoperation rate.

The invention discloses a bat algorithm and Otsu algorithm-based infrared image segmentation method. The method comprises the steps of S10: obtaining a to-be-segmented infrared image; S20: taking an interclass variance function in an Otsu algorithm as a target function of a bat algorithm, and taking a segmentation threshold variable corresponding to the interclass variance function as a bat position variable in the bat algorithm; and S30: based on the bat algorithm, calculating a corresponding bat position when the target function is maximal, and segmenting the to-be-segmented infrared image into a background region and a target region by taking the bat position as an optimal segmentation threshold. In addition, the invention discloses a corresponding system. The to-be-segmented infrared image can be segmented quickly. Compared with an existing infrared image threshold segmentation method, the threshold selection speed is effectively increased and a foundation is laid for real-time extraction and analysis of temperature field characteristics of subsequent equipment.

The invention discloses a seabed geomorphologic boundary extraction method based on an OTSU algorithm. The method comprises the following steps: 1) calculating maximum slope and gradient of each measurement point according to water depth distribution in a target area; 2) calculating similar curvature of each measurement point according to gradient distribution in the target area; 3) normalizing the similar curvature value of each measurement point in the target area to a specified range; 4) according to the nonlinear transformation relation, expanding the dynamic range of a local area of the normalized curvature value; 5) estimating threshold of curvature segmentation through the OTSU algorithm; and 6) dividing the measurement points, the curvature values of which are not smaller than thethreshold, into boundary points. The geomorphologic boundary extraction method can effectively identify complex geomorphologic boundary details, suppresses course noise, greatly reduces work intensityof scientific research and engineering personnel and improves work efficiency thereof; and the method is suitable for boundary extraction of geomorphologic units in relevant fields of morphological analysis, marine surveying and mapping, seabed resource survey and oceanographic engineering.

The invention provides a lane line point cloud extraction method, electronic equipment and a storage medium, and the method comprises: obtaining a road marking point cloud, and extracting the lane line point cloud from the road marking point cloud according to the shape features of the road marking point cloud cluster; after the lane line point cloud is segmented, counting the frequency distribution of the intensity of each segment of lane line point cloud data, and generating a frequency distribution histogram; calculating a segmentation threshold of the point cloud intensity according to thefrequency distribution histogram and an Otsu algorithm; and carrying out binary segmentation on the lane line point cloud data of each segment by utilizing the segmentation threshold of the point cloud intensity so as to extract a highlighted lane line point cloud. By means of the scheme, the problems that an existing lane line point cloud extraction method consumes much time and is difficult toguarantee accuracy are solved, lane line point cloud extraction can be completed accurately and rapidly, and edge point data on the two sides of a lane line are reserved to the maximum extent.

The invention discloses a neutral point identifying method based on polarization degree ellipse fitting. The neutral point identifying method is characterized by being performed as the following process:1, n sky polarization images are obtained; 2, an Otsu algorithm is used for extracting effective areas; 3, polarization degree images are calculated; 4, binarization processing is performed on the polarization degree images; 5, the areas where neutral points are located are searched for in the effective areas; 6, an ellipse fitting algorithm is used for performing fitting on the edges of the areas where the neutral points are located; 7, the neutral points are identified. The neutral point identifying method can effectively identify and demarcate the neutral points in the sky polarization images and therefore can provide neutral point characteristic information for atmospheric environment detection.

The invention discloses a path identification and centerline optimization method based on a circular road. The method comprises the following steps of S1, obtaining a grayscale image of the circular road; S2, selecting an optimal threshold value by adopting an Otsu algorithm to obtain a binarized image with an optimal pixel point of 0xff or 0x00; S3, determining an image road side boundary of thebinarized image by adopting an edge line hunting method; S4, extracting sideline characteristics of a road in the binary image, judging whether the sideline characteristics are circular rings or not,and if the sideline characteristics are circular rings, optimizing the sideline by adopting a primary linear interpolation method to obtain an optimized secondary sideline so as to obtain a primary road center line; and S5, optimizing the first road center line to obtain a second road center line, namely a final road center line. The circular road image features are identified and compared, the circular road can be accurately identified, and accurate and efficient road path identification in the method has profound significance for development of unmanned driving.

The invention discloses a method for automatically segmenting the human lateral geniculate nucleus through prior knowledge. According to the method, field bias correction is performed on structural image data of the human brain, and then minor structures of the brain are segmented through the corrected image so as to obtain a template in the ventral diencephalon region. Next, the corrected structural image data and the template of the ventral diencephalon region are in registration into MNI standard space. The region limit in the MNI space of the LGN is obtained according to the prior knowledge of the anatomic structure of the lateral geniculate nucleus, then in the region, the LGN is segmented out though the region growing method, the k average value method, the Ostu method and the image segmenting method. Then, several segmenting results are fused to obtain an estimated value of the LGN region to serve as a segmenting result. Finally, the segmenting result is changed into original space of a structural image, and it is that the final segmenting result of the LGN is obtained.

The invention discloses a visual quick response (QR) code generation method based on a Sobel operator. The method comprises the steps of (1) performing Gaussian smoothing processing on an original image I to obtain a processed image Q1; (2) performing edge detection on the image Q1 by use of a Sobel algorithm to obtain a gradient modular diagram T; (3) calculating a threshold value XT of the gradient modular diagram T by use of an Otsu's method, and processing the gradient modular diagram T to generate a color contour diagram C; (4) performing corrosion processing on the color contour diagramC; (5) generating an original quick response code Q which includes corresponding data information, wherein a version number and an error-correction level need to be determined by a user, and meanwhilethe size of the original quick response code Q is determined; ( 6) adjusting the size of a color contour diagram C1 to obtain a color contour diagram C2; and (7) performing visual encoding on the original quick response code Q and the color contour diagram C2 to generate a visual quick response code image QR. The visual quick response code generated by the method has high anti-counterfeiting performance and better visual effect.

An improved multi-sequence magnetic resonance image registration method based on self-similarity comprises the following steps: 1) de-noising all three-dimensional magnetic resonance images layer by layer using a non-local mean fast de-noising algorithm; 2) extracting the foreground part by using the cascading Otsu algorithm, that is, using the Otsu algorithm for many times to superimpose the foreground part on the previous image; 3) using local three-dimensional median filter to optimize the result; 4) encoding that preprocessed image by use self-similarity; 5) using a discrete-based optimization method to calculate similarity measure between feature images, and judge whet that similarity measure reaches the optimal state; 6) using the final transformation matrix to transform the originalfloating image to obtain the final result map. The invention has strong anti-noise interference ability and high registration accuracy.

The invention relates to a method for extracting a cornea contour from a cornea dynamic deformation video based on an Otsu algorithm. The method comprises the following steps: step 1, preprocessing a cornea contour image; step 2, carrying out image segmentation on the image preprocessed in the step 1 by adopting an Otsu algorithm, and introducing mathematical morphological operation to reduce the possibility of occurrence of holes and fine noise; and step 3, calculating an image gradient by using a Gaussian function, and detecting the edge of the image by setting double thresholds. According to the method, the accuracy and integrity of the extracted cornea contour can be greatly improved.

Methods for characterizing atmospheric turbulence along an optical path from a laser transmitter to a laser receiver can include the steps of counting the number of laser speckles at the receiver imaging plane, and then finding Fried's parameter r0 using the counting result to characterize the turbulence along the path. Before counting speckles, images at the receiver image plane can be preprocessed by capturing the images. The captured images at the image plane can then be blurred and a threshold can be chosen so that only certain pixels in the image are further processed. The thresholding can be via Otsu's methods or via variants of a Gaussian fit. Kostelec's method can then be used to count speckles in the portions of the image that have made it through the thresholding step. Other counting methods could be used. Fried's can then be found using the speckle count.

The invention discloses a method for automatically segmenting human lateral geniculate nucleus (LGN) by using prior knowledge. The method comprises the steps: firstly, carrying out bias field correction on data of a structure image of a human brain, and then, carrying out brain substructure segmentation on the corrected image, so as to obtain a template of a LGN; then, registering the corrected structure image data and the template of the LGN to a MNI (Montreal Neurological Institute) standard space; acquiring the region boundary of the LGN in the MNI space according to prior knowledge of an LGN anatomical structure, and then, segmenting the LGN in the region respectively by a region-growing method, a k-means method, an Otsu algorithm and an image segmentation method; then, fusing the segmented results, so as to obtain an estimated value of the LGN region, which serves as a segmented result; finally, transforming the segmented result into an original space of the structure image, thereby obtaining a final LGN segmented result.

A method is described to obtain a binary image from the print-and-scan process to best match the known original. A point-spread function (PSF) of the PAS process is first obtained from its knife-edge responses, and deblurring is carried out on the scanned images using deconvolution. After image deskewing and preliminary registration, a supervised adaptive thresholding procedure is utilized to binarize the scanned image such that a measure of difference (e.g. the Euclidean distance) between the original and binarized images is minimized. The supervised adaptive thresholding procedure divides the scanned images into many rectangular sub-images. Otsu's method is used to find a starting threshold for each scanned sub-image. An optimal threshold is found around the Otsu's threshold via iterative search to minimize the measure of difference between the original sub-image and scanned sub-image. The sub-images are binarized using the optimal threshold. This method may be used in document authentication.

The invention provides a scanned payment method based on mobile payment equipment. After payment information sent by the server is received, the payment two-dimensional code is intercepted as a payment image, the payment image is stored in a picture library, and the payment information is generated by the server according to the payment information obtained by the payment terminal equipment through the payment two-dimensional code displayed by the mobile payment equipment; when it is monitored that the payment event is triggered, payment operation is completed. The method is directed to payment images, an Otsu algorithm is adopted to perform coarse segmentation on a payment image, segmentation operation suitable for the payment image is completed through shape testing on a result obtainedafter coarse segmentation operation, after payment information sent by a server is received, a payment two-dimensional code is intercepted to serve as the payment image, and when it is monitored thata payment event is triggered, payment operation is completed. Therefore, the image reading operation is efficiently, accurately and quickly realized for the payment image, and the method has usabilityand applicability. The invention further provides a scanned payment device based on the mobile payment equipment.

The present invention relates to a device and method for classifying seeds, for example coffee seeds. The device is characterised by: a seed-feeding mechanism; a seed-containing mechanism connected to the seed-feeding mechanism; an electronic seed-viewing system operationally disposed in the seed-containing mechanism; a seed-ejecting mechanism connected to the outlet of the seed-containing mechanism, the electronic seed-viewing system having a central processing unit that implements methods for classifying seeds. The method is characterised by the steps of: a) obtaining a digital image of the seed; b) storing the RGB components of the image obtained in step a); c) generating a histogram for each colorimetric and luminosity component of the histogram of step b); d) determining the thresholding point according to Otsu's method; e) obtaining a binary image; f) removing areas; g) obtaining the edges; h) obtaining the vectors corresponding to the seeds; i) identifying black seeds; j) identifying seeds with fermentation- and immaturity-related defects; k) identifying seeds with mechanical damage; and l) actuating a seed-ejecting mechanism, activating actuators for black seeds, actuators for seeds with fermentation- and immaturity-related defects and actuators for seeds with mechanical damage.

The invention discloses an adaptive head lighter turning control algorithm based on lane line detection. The algorithm comprises the following steps that (1) an image is preprocessed, regions of interest are determined according to structural characteristics of the image, the image is divided than, and contrast of different regions is improved via linear grayscale transformation, and an improved Otsu's method is used to binarize the image; (2) an improved Hough transformation algorithm is used to detect lane lines; (3) a curve fitting and curvature radius is calculated; and (4) according to the parking stadia and the geometric relation between the curve illumination distance and the curvature radius, a head lighter angle adjusting model is established and solved. A head lighter is rotatedto eliminate a visual blind area in the inner side of a curve during driving at night, and the driving safety in the curve at night is ensured.

The invention relates to an ultraviolet discharge image segmentation method and a computer readable medium, and the method comprises the steps: 1, obtaining a to-be-segmented ultraviolet discharge image, and carrying out the preprocessing of the image; step 2, carrying out superpixel segmentation on the ultraviolet image; step 3, performing saliency detection on the image; and 4, performing binarization threshold selection on the saliency detection result by using an Otsu algorithm optimized by a bat algorithm with Levy flight characteristics to obtain an ultraviolet image segmentation result. Compared with the prior art, the method has the advantages of good segmentation effect, accuracy, rapidness and the like.

The invention discloses a rural resident point information extraction method based on a high-resolution remote-sensing image. The rural resident point information extraction method comprises the following steps: firstly, carrying out Canny edge detection on the high-resolution remote-sensing image, and obtaining a vector edge line through chain code tracking; then, carrying out Douglas-Peucker abbreviation on a vector edge to extract straight-line segments, and rejecting shorter straight-line segments and keeping longer straight-line segments according to length constraints; secondly, according to adjacency relation between the straight-line segments and included angle constraints, extracting straight angle points, calculating the density characteristics of the straight angle points within a certain space region range, and generating a straight angle point density characteristic image, wherein the characteristic image and the input high-resolution remote-sensing image have same spatial range and spatial resolution; and finally, carrying out binarization processing on the density characteristic image by utilizing Otsu, extracting rural resident point vector pattern spots through connected component analysis so as to extract rural resident point information. Precision and effects are improved, and robustness and universality are better.

The invention discloses a method adopting image segmentation for automatically testing the LED indicator light of automobile instruments and provides a region-growing method based on the maximum between-cluster variance (Otsu's Method). The method comprises the following steps: automatically selecting the centroid in the light-emitting region of the LED light as the seed point to carry out the region-growing; realizing the region segmentation and extraction; and judging whether the brightness and color of the light are correct upon the seed point and the area of the region. The invention can effectively improve the testing accuracy rate and testing efficiency of the LED indicator light of the automobile instruments and provide effective means for the automatic production of the automobileinstruments.

The invention provides a drivable area segmentation method and system, electronic equipment and a storage medium. The method comprises the steps that road surface information is collected through a vehicle-mounted camera and a laser radar; constructing a lane line equation, marking obstacle information, and segmenting a drivable area in a lane line through an Otsu algorithm; and when no lane line exists on the road surface, matching the current road through a predetermined road model, extracting road edge information, and virtualizing the lane line to obtain a drivable area of the lane line-free road. Therefore, the path process can be simplified, vehicle calculation performance requirements are reduced, and automatic driving safety is guaranteed.

The invention relates to an online wear particle image data processing method. The method comprises the steps of S1, performing graying processing on a wear particle image; S2, adding random noise tothe grayed wear particle image, then denoising the wear particle image through a 3M-WVNID denoising algorithm, and integrating wavelet analysis denoising, two-dimensional variational mode decomposition and non-local mean filtering fusion by the 3M-WVNID denoising algorithm; S3, performing binarization image segmentation on the denoised wear particle image by adopting an Otsu algorithm; S4, carrying out boundary feature detection by adopting a canny operator and carrying out wear particle image segmentation by adopting a watershed algorithm; and S5, extracting characteristic parameters from thewear particle image after image processing. The extracted wear particle image characteristic parameters can more accurately represent the health state monitored by the equipment through the online wear particle monitoring method in time and accurately.