504 results about "Hough transforms" patented technology

The invention discloses a bend target identification system and method based on multi-sensor fusion, which mainly aims at a problem of front target detection of a vehicle at the bend of expressway. The lane line is divided into a straight line portion of the near view field and a curve portion of the far view field. For information acquisition of a camera, lane line fitting and tracking at the near view field are completed by using Hough transform and Kalman filtering, and curve fitting at the far view field is completed by using a BP neural network. For information acquisition of radar, information of a static object group is extracted, and curve fitting is performed by using the BP neural network. through space-time alignment, the lane line information acquired by vision and the lane line information acquired by the radar are fused to determine a travelable region of the lane where the vehicle is located, and finally a bend target identification algorithm based on the fusion of the camera and the millimeter wave radar is provided by combining the travelable region and the lane line type so as to realize detection for targets at the bend.

The invention discloses a round small target accurate detection method based on an image, and solves the problems that the target is hard to be detected or the detection precision is low due to small target imaging. The method comprises the following steps of fast positioning a round small target by utilizing improved random Hough transform at the initial detection stage, and extracting a small range ROI (region of interest) in the image; and at the secondary accurate detection stage, obtaining sample points of boundary in the target through a centering eight-directional searching method, and solving the accurate detection result of the round small target by adopting a least squares fitting method. The method provided by the invention can be used for accurately identifying and detecting the round small target from the complex background, ensuring the fastness of the initial detection by utilizing the improved random Hough transform, and further improving the precision through a secondary detection way.

The invention discloses a road marking detection method in the field of vehicle aided driving or automatic driving. Relative position information of a vehicle and a road can be provided for a driver or a vehicle-mounted system so as to reduce the traffic accidents caused by deviation. A plurality of marked image sensors are mounted around the vehicle to be obtaining the information, an image edge point is extracted by using an edge detection method based on a self-adaptive gradient threshold value, and then the straight line is fast extracted from the edge information by using improved method in Hough transform, the reverse solving is performed on the straight line according to the marking information so as to fuse the multi-angle information into a coordinate system using the vehicle as the centre, and finally the position relation of the road marking and the vehicle can be obtained by carrying out statistics on the straight line information based on a parallel marking model. The method disclosed by the invention has the characteristics that the detection angel range can be selected, the algorithm calculated amount is few, and the method is easy to realize.

The invention provides a method for real-time lane line detection based on vision under complex lighting conditions and belongs to the field of computer vision and unmanned intelligent driving. During image preprocessing, light estimation and light color correction are conducted on different light images so that the images can recover under standard white light; noise which is introduced in the image acquisition process is eliminated through Gaussian filtering, and then the images are binarized and subjected to edge extraction; the original images are divided in the extraction process; improved Hough transform is used for obtaining a lane candidate line, and a dynamic interesting region (ROI) is built; through Hough transform based on the dynamic interesting region (ROI) and kalman filtering, a lane line is tracked in real time to realize constraint and update of a lane line model. An algorithm is added into a lane line detection failure judgment module to improve the reliability of detection. The method is high in speed and good in robustness, a good lane line detection effect is obtained under the complex lighting conditions, the dynamic lane line identifying ability of a vehicle is improved, and the safety of unmanned vehicle driving is improved.

The present invention relates to the methods of estimation and recovering of general affine geometrical transformations which were applied to data, extensible to any other defined class of geometrical transformations, according to the preamble of the dependent claims. The parameters of the undergone deformation are robustly estimated based on maxima given by a parametric transform such as Hough transform or Radon transform of some embedded information with periodical or any other known regular structure. The main applications of this invention are robust digital still image / video watermarking, document authentication, and detection of periodical or hidden patterns. In the case of periodical watermarks, the watermark can also be predistorted before embedding based on a key to defeat block-by-block removal attack.

The invention, which relates to the field of metering equipment employing an optical method as a feature, provides a machine vision positioning method for automatic screw assembling. The method comprises steps of inputting and recording data, obtaining an image, carrying out image pretreatment, carrying out Hough transform treatment, calibrating a camera, and calculating a position. According to the method, a circle in an image is searched by using the Hough transform method and the center of the circle is used as a positioning point, so that defects of high selection requirement and low precision for feature point selection according to the existing visual positioning method can be overcome; and a defect of an assembling failure caused by inaccurate product position fixation by a clamp during the automatic screw assembling process in the prior art can be overcome.

The image processing ECU periodically acquires road-surface images and extracts edge points in the acquired road-surface image. Subsequently, the ECU determines the operating mode and extracts the edge line when the operating mode is either a dotted mode or a frame-accumulation mode. The edge points are transformed e.g. Hough transform, to extract an edge line that most frequently passes through the edge points. The extracted edge line denotes the lane marking. The ECU outputs a signal to activate a buzzer alert when determining the vehicle may depart from the lane.

The invention discloses an adverse state detection method of transverse pins of rotation double lugs of high-speed rail overhead contact line equipment. The method includes following steps: in a positioning stage, firstly, according to line characteristics of a connecting position at which a cotter pin is located and inclined cantilevers in an image, detecting lower edges of the inclined cantilevers by employing Hough transform and extracting slope characteristics thereof to realize primary positioning of the cotter pin, and then classifying primary positioning pictures by employing a PBoW model; and in a detection stage, performing coarse positioning by employing target geometric distribution characteristics, performing image enhancement by employing a Retinex method according to characteristics of small size, darkness, and weakness of the cotter pin, extracting a circular characteristic of a screw at the connecting position by employing a Hough accumulation matrix, realizing accurate extraction of a circular ring area in which the cotter pin is located, and finally giving detection criteria of deletion of the cotter pin and insufficient field angle according to a in-circular-ring non-communication area distribution rule. According to the method, positioning accuracy of the connecting position of double lugs and the connecting position of upper ends of short diagonal braces and fault detection of falling of the cotter pins and insufficient field angle are realized.

Methods and apparatus for determining the skew angle of a digitized form scanned at an angle to the original page are described. The method determines a rotation angle and a boundary rectangle of best fit. Sections of the digital image in the form of thin strips are examined for the boundary between the page and the region of the scanned image beyond the original page. The Hough transform is employed to determined candidate edge line segments for the page from the sets of perimeter points. These line segments are then combined to select the best rectangle enclosing the page, from which the skew angle is determined. The algorithm also determines a rotated bounding box enclosing the page. An innovation of the invention is the use of fuzzy logic, whereby several candidates for perimeter points, candidate edges, and bounding rectangles are determined in each step of the computation with associated confidence values and the final skew angle is selected by choosing the bounding rectangle with the highest confidence values.

The invention discloses a lane line historical frame recognition method. The method comprises performing Gaussian filter preprocessing on acquired road images and performing adverse projection on the images to be converted into airscapes; performing binarization processing on the airscapes through an adaptive threshold binarization method and performing Hough transformation on the airscapes to extract lines; determining which lines in Hough transformation results are lane lines based on the line angle LX.theta, the line distance LX.rho, the line ticket number LX.V and the line initial point distance LX.S of each line LX, wherein X =1, 2, ......N and the space and positions of the lane lines obtained in the previous period. By the aid of the method, interferences of complex road conditions of characters, multiple interference lines, shadow shielding, lane line damage, stain coverage and the like are removed effectively, and the recognition rate and the stability of lane lines are improved greatly; the method can be widely applied to car safety auxiliary driving systems to assist drivers for the maintaining of traveling of cars in the lanes in the monotonous driving environment.

The invention relates to the field of machine vision, and discloses a pointer type instrument reading automatic identification method based on Faster R-CNN and U-Net, and the method comprises: S1), making a Faster R-CNN data set; S2) establishing a Faster R-CNN network model, and respectively training and testing the Faster R-CNN network model; S3) constructing a U-Net network model, establishinga loss function L, and respectively training and testing the U-Net network model; S4) fitting a scale line contour by utilizing a scale line segmentation result; S5) calibrating the dial image by using perspective transformation; S6) detecting a pointer area by utilizing a Faster R-CNN network model; and S7) obtaining a pointer inclination angle of the dial calibration image and a final result. According to the invention, the Faster R-CNN model is adopted to replace ORB and other traditional corner detection algorithms, so that the accuracy of detecting the area where the instrument dial and the instrument pointer are located is improved; and a Hough transform algorithm is replaced by an image segmentation and contour fitting method, a U-Net model and a corresponding loss function are redesigned for the characteristics of the electric power instrument, and the automatic identification accuracy is high.

The invention discloses a hexagonal bolt looseness detection method based on deep learning and Hough transform, and belongs to the technical field of computer vision. According to the method, a deep learning technology is used, and detection of bolt looseness is realized through an area-based full convolutional neural network target detection algorithm. The depth characteristics of an input picture is calculated to obtain a bolt position, a bolt image is extracted from an original picture according to the calculated bolt position, edge information is obtained through a series of image processing algorithms, and a bolt angle is calculated through Hough transform. By adopting the designed detection scheme, whether the bolt is loosened or not can be accurately detected.

The invention discloses a detection method for radar multi-target Hough transform target-by-target elimination, and belongs to the field of tracking processing before radar detection. The conventional tracking processing method before radar multi-target detection has the defects of high operational complexity, great mutual influence between flight paths, simple setting of a second threshold and the like. Therefore, each parameter unit is detected and judged one by one by adopting Hough transform target-by-target elimination technology. By utilizing the capacity of the Hough transform of simultaneously forming peak values through multiple targets and adopting a target-by-target elimination strategy, the technology greatly reduces the mutual influence between target flight paths; meanwhile,different second thresholds are set to detect different parameter units, the technology does not need prior information in a target number and multi-hypothesis test and has obviously lower operational complexity; meanwhile, target motion information, flight path feasibility and other factors are considered in the process of the setting of the second thresholds, the false-alarm probability is reduced and the detection probability of a detector is improved.

The invention provides an image processing method for automatically measuring the thickness of a fetal nuchal translucency from an ultrasonic image. The image processing method comprises the following steps that an ultrasonic instrument collects the input of an image and the importing of shooting parameters; quality analysis of an ultrasonic image to be detected of the fetal nuchal translucency based on Gabor feature matching is carried out; the ultrasonic image which accords with a measurement condition is screened according to a Gabor feature matching degree of the images in an image region and a training set; with regard to the ultrasonic image which accords with the measurement requirements, the position of the NT in the ultrasonic image is detected; a partial image of a clear NT line can be measured by a final output of the step through gradually-iterative threshold segmentation, canny operator edge detection and hough transform line detection; an NT target in the ultrasonic image can be measured; a K-MEANS method is used for classifying lines of NT target positions and upper and lower two clusters of line sections are aggregated to find out two measurement lines of the NT positions, and measuring a maximum pixel distance between the two lines; and the pixel distance of the NT is converted into a real length value according to a measuring scale of the shooting parameters.

The invention provides a visual inspection method for step defects of bearing rivets. The method comprises the following steps: step 1, bearing rivet image acquisition; step 2, bearing rivet image preprocessing; step 3, inner step defect detection; step 4, outer step defect detection; step 5, result analysis for judging qualification of rivets, wherein unqualified phenomena comprise inner step defects, outer step defects, inner and outer step defects or other types of defects. The method has the following advantages: 1) rivet regions are positioned reasonably and the defects are detected with an adaptive detection method, and the algorithm robustness is good; 2) in the aspect of detection of rivet inner step defects, an region labeling method is good in stability; 3) for inner steps, the inner step judging accuracy is enhanced with a secondary region labeling algorithm; 4) during outer step detection, outer ring parameters are detected firstly through Hough Transform, then whether outer step regions are effective or not is judged according to a geometric constraint relation between outer rings and rivet regions, judgment is effective, and misjudgment is avoided.

The invention relates to the technical field of image recognition, and especially relates to a character recognition method. The method comprises the following steps: obtaining an image to be recognized; preprocessing the obtained image: performing image tilt correction to correct the image and performing threshold processing to get an image with single foreground information and background information; analyzing and processing the image: analyzing the inter-row texture features of the image and obtaining the character matrix parameters of the image; segmenting the image: segmenting the image based on the character matrix parameters to form a plurality of sub images, and obtaining character blocks of the image; and performing recognition: separately processing the character blocks, obtaining the image features of the character blocks, and recognizing the image features, wherein image correction includes the sub steps of expanding the image, detecting the edge of the image after expansion, performing Hough transform on the edge points to find the angle of the longest line, getting the angle of the straight line and rotating the angle of the image. The method is simple and easy to implement, and is of high recognition rate.

The invention discloses a method for automatically identifying a sulfur hexafluoride pressure instrument image. The method comprises the following steps: an image obtained through instrument video monitoring is pretreated and converted into a gray image; the OTSU is utilized to find a proper threshold value of the image, a target pointer in the instrument image is distinguished from a disc background; sobel operator edge detection is performed on the gray image, and then Hough Transform is utilized to obtain the coordinate and radius of a central point of a circular area of the image; according to the features of the instrument image, the reference point position and the reference terminal point coordinate of a dial plate are obtained; according to the obtained coordinate parameters, the deflection included angel of the pointer is calculated, and the pointer read is calculated by combining the reference point position of the dial plate to realize the automatic identification of instrument image read. According to the invention, the pointer read of the instrument can be accurately and rapidly identified, a template image database is not needed to be established in advance, which is remarkably different from other image identification technologies, and automatic read identification of the image pointer of the sulfur hexafluoride pressure instrument can be realized through feature separation and identification of the image.

The invention discloses a circular pointer-type dial visual sense positioning and automatic reading method and belongs to the field of power meter automation and visual sense application. The invention comprises the following steps: reading a dial image with a complicated background; graying and equalizing the image; then, automatically locating the central position of the circular dial by random Hough transform; and simultaneously detecting a plurality of circular contours corresponding to the central position; taking the area with the largest contour to extract the accurate dial image; taking the radius corresponding to the smallest circular contour and performing random Huff's accumulation detection outward to obtain two adjacent large circles with the strip area between the two large circles being a dial scale area; performing threshold segmentation and re-expansion and the connected domain filtering to find out the initial position and the ending position of the graduation; and using the linear fitting method to find out the position of the pointer and calculating the reading. The method of the present invention can directly locate a circular instrument and multiple rings in the instrument under a complicated background, and at the same time, a new algorithm to locate the position of the pointer is proposed.

The invention discloses a nameplate recognition method based on depth learning. The method comprises the following steps: identifying and determining an area where the nameplate is located by using anobject detection algorithm to identify an object picture; then edge detection is carried out by using Hough transform, and the slanted image is automatically corrected by affine transform; cutting the picture of the nameplate into text to determine the position information of each line of text; the segmented text is recognized and the recognition result is obtained. The invention also includes anameplate recognition system based on depth learning. The invention adopts the depth learning algorithm to segment the nameplate at the line level, solves the problems of low Chinese character segmentation efficiency and low segmentation precision, and improves the recognition efficiency. CRNN algorithm is used to recognize line-level text, which improves the accuracy of text recognition.

A method for restoring video data of a pipe based on computer vision is provided, including: performing gray stretching on pipe image / video collected by a pipe robot; processing noise interference by smoothing filtering; extracting an iron chain from the center of a video image as a template for location; performing target recognition on the center of video data by an SIFT corner detection algorithm; detecting ropes on left and right sides of a target by Hough transform; performing gray covering on the iron chain at the center of the video image and the ropes on two sides; and restoring data by an FMM image restoration algorithm.

A system and method for tracking multiple targets using distributed linear sensor arrays. The system includes a plurality of arrays of sensors for receiving signals from a target; a receiver for receiving signals received by the plurality of sensor arrays; an analog / digital converter for converting the signals received from the sensor arrays to a digital format, if signals are received in an analog format; a digital storage device for storing the digitized data from the sensor arrays; and a computer system for retrieving the stored digitized data from the plurality of sensor arrays and processing the data through the use of a composite Hough transform to determine a track of the target. Embodiments include a computer system and method for processing target data from sensors through the use of a composite Hough transform to determine a track of the target.

The invention discloses a vehicle collision prevention early warning method based on vision. The method comprises steps that two determination conditions including a lane line length and a gradient are added to a detection lane line to carry out Hough transform, and a lane line of a self vehicle is extracted; at a detection target vehicle portion, in combination with a normal distribution function, an OTSU method is employed to combine an improved method of a genetic algorithm (GA) to separate a target and the background; at a distance measurement portion, a method combining geometric relation projection with camera calibration is employed, danger determination is carried out according to a relative distance, and collision prevention early warning is realized. According to the method, the distance between the self vehicle and a front vehicle can be more accurately estimated, a driver is prompted through alarming to adopt measures for processing, and road traffic accidents can be greatly reduced.

The invention relates to a three-dimensional attitude estimation method in disordered grabbing based on SHOT and ICP algorithms. The method is used for object pose estimation in the disordered grabbing process. A good initial position is provided for ICP fine registration through an SHOT coarse matching method, an M matrix in an SHOT algorithm is constructed, eigenvalue decomposition is carried out, and then a local reference system is constructed; topological characteristics under a corresponding reference system are calculated, a result is stored in a histogram, Hough transform is performedon a characteristic extraction result corresponding to the model and the scene point cloud, wrong point pairs are eliminated, and resolving is performed to obtain a final coarse registration attitudeestimation result; and nearest point searching is carried out by using an ICP algorithm, a covariance matrix construction is constructed, and a rotation matrix and a translation vector for minimizingan objective function are calculated. The problem that the registration precision is reduced when the traditional ICP is applied to the collection of point clouds with holes due to the reflection of workpieces is solved.

The invention puts forward a device for road surface identification in low-definition video streaming. The device comprises a road surface information collecting device with a binocular camera system, wherein the road surface information collecting device uses the binocular camera system to shoot road ahead of a vehicle so as to collect video information of the road ahead; a road surface or roadblock region detection device for converting the video information of the road ahead collected by the road surface information collecting device into a disparity map, converting the disparity map into a series of sub-V disparity maps and U disparity maps, performing adaptive line segment fitting to each sub-V disparity map and each U disparity map by using adaptive Hough transform, fitting mapping line segments of the road surface and the roadblock in each sub-V disparity map, and obtaining regional ranges of the road surface and the roadblock according to the mapping line segments of the road surface and the roadblock, fitting a mapping line segment of the roadblock in the U disparity maps, and obtaining a regional range of the roadblock according to the mapping line segment of the roadblock.

The invention discloses an automatic identification method and system for a pipeline target in ground penetrating radar data. According to the automatic identification method and system for a pipelinetarget in ground penetrating radar data, the positioning of a pipeline feature (hyperbolic curve) region in an image is identified by using a regional convolutional neural network; the positioning isaccurate and rapid, and the training time of a model is short; compared with the traditional Hough transform, the method for extracting the pipeline parameters from the pipeline feature region by using the cross-correlation operation and the Hough transform of the one-dimensional parameter space has the advantages that the operation time is shorter, and the requirement on hardware is low; and compared with a traditional full-graph search and calculation mode, the mode of carrying out region identification on the full-graph lines and then extracting the parameters from the regions has the advantages that the calculation of non-feature regions is avoided, and the calculation efficiency is greatly improved.

The present invention discloses a vehicle-mounted lane line detection system and method. The system includes a processing unit arranged on a traveling vehicle and an image acquisition unit connected with the processing unit; the image acquisition unit is used for acquiring lane pictures in the traveling process of the vehicle in real time and sending the lane pictures to the processing unit; and the processing unit is used for obtaining lane line sample pictures in an offline or online manner to establish a lane line sample library, extracting the RGB pixel values of lane lines in the lane line sample library to establish a G-S color model, inputting the lane pictures in the traveling process on the basis of the G-S color model, detecting the lane lines in the lane pictures through likelihood probability calculation and threshold segmentation, and extracting the lane lines from the lane pictures through Hough transform. With the vehicle-mounted lane line detection system and method of the present invention adopted, influence caused by environmental change or illumination change can be eliminated, and a lane line detection effect can be effectively improved.

The invention provides an image recognition method for the state of a pressure gauge of a fire extinguisher.The method comprises the following steps that a digital image of the pressure gauge of the fire extinguisher is collected and preprocessed; the dial plate area of the pressure gauge is extracted from the preprocessed digital image through a Hough transform mode and preset dial plate color characteristics of the pressure gauge; the dial plate area of the pressure gauge is searched for the current position of a pointer through the HIS color characteristics; the colors of the left side and the right side of the pointer are detected according to the current position of the pointer and preset color judgment conditions to judge the current state of the pressure gauge.Automatic recognition of the state of the pressure gauge is completed, the influence of certain environment factors and self conditions is eliminated, and the recognition accuracy is high.

The embodiment of the invention provides a text detection method and device based on image processing, a medium and electronic equipment, and the method comprises the steps: obtaining a text inclination angle of a to-be-detected image through Hough transform, and rotating the to-be-detected image through the text inclination angle, so as to obtain an inclination correction image; inputting the inclination correction image into a pre-trained text region detection model to obtain a text to-be-selected region in the inclination correction image; determining one or more text to-be-selected imagescorresponding to the to-be-detected image based on the text to-be-selected area, and performing graying processing on each text to-be-selected image; determining a binarization segmentation thresholdvalue according to the gray value of each pixel point in the text to-be-selected image subjected to the gray processing, and performing binarization processing on the text to-be-selected image by utilizing the binarization segmentation threshold value; and identifying the text information in the text candidate image after binarization processing. The method can improve the character recognition efficiency and recognition precision.

A method for converting a portion of an image from a first domain to a second domain. The method may apply a Hough transform on the converted portion of the image, including calculating a range of angles for each tested pixel q relative to a center pixel p, quantizing the range of angles into a plurality of bins, voting each tested pixel q using a range of bins using a weighted voting schema; and detecting one or more features in the portion of the image. The methods may be implemented by program instructions executing in parallel on CPU(s) or GPUs.