43 results about "Top-hat transform" patented technology

Disclosed is a copper sheet and strip surface defect detection method based an on-line sequential extreme learning machine. The method includes the following steps that a copper sheet and strip surface image is captured through an image capturing module; the captured copper sheet and strip surface image is enhanced according to the median filtering method with the masking size of 7*7 to reduce noise in the copper sheet and strip surface image and the effect of the noise on the quality of the surface image; the copper sheet and strip surface image is subject to tophat transform treatment to reduce the effect of uneven illumination; a copper sheet and strip surface image pre-detection method based on eight-neighborhood difference values is adopted; defects in the surface image are segmented according to an image segmentation method, wherein it is judged that the copper sheet and strip surface image has the surface defects after pre-detection; geometrical characteristics, gray characteristics, shape characteristics, texture characteristics and other characteristics of each defect are extracted, and copper sheet and strip surface defect characteristic dimensions are subject to optimization and dimensionality reduction according to the principal component analysis method; a copper sheet and strip surface defect classifier based on the on-line sequential extreme learning machine is designed, and samples are used for training; characteristics of the copper sheet and strip surface image to be detected are extracted to identify types of the surface defects.

The present invention relates to a system for automatically evaluating oxygen saturation of the optic nerve and retina, said system comprising: image capturing system further comprising: a fundus camera (26), a four wavelengths beam splitter (27), a digital image capturing device (28), a computer system, image processing software performing in real-time the steps of: registering set of multi-spectral images (1) by, binarizing the multi-spectral image (7), find the all the border regions of each image by finding the region including the straight line that passes the most number of points in the region (8), use the orientation of the borders to evaluate the orientation of each spectral image (9), equalize the orientation of each spectral image by rotating the spectral image (10), edge detect each spectral image (11), estimate the translation between the spectral images based on the edges of adjacent images (12), transform the images to a stack of registered images (13), locating blood vessels (2) in each of the multi-spectral images by, retrieving registered spectral images (14), for each spectral image, remove defective pixels (15), enhance image contrast (16), perform top-hat transform using structuring element larger than the largest vessel diameter (17), binarize each image (18), apply filter for smoothing the image (19), combine all the spectral images using binary AND operator (20), skeletonize the resulting image (21), prune the skeleton image (22), re-grow the pruned image (23), locate junction points (24), evaluating the width of the blood vessels (3) by calculating the normal vector to the vessel direction to, estimate the direction and position of the vessel profile, evaluate positions of vessel walls based on the vessel profile obtained, evaluate vessel width based on the position of the vessel walls, selecting samples for calculating optical density, calculating the optical density ratio (4), evaluating oxygen saturation level (5), and presenting the results (6), wherein presenting the results may include presenting numerical and visual representation of the oxygen saturation.

The invention relates to an image fusion technique, in particular to an infrared intensity image and infrared polarization image enhancement and fusion method, which can solve the inherent problems of edge distortion and low contrast in conventional infrared fused images. The method is carried out according to the following steps that support value transform is respectively carried out on an infrared intensity image and an infrared polarization image, so that low-frequency images and support value sequence images are obtained; the top-hat transform of mathematical morphology is used for respectively extracting bright information and dark information; the bright information images and the dark information images are enhanced; the two last-layer low-frequency average images and the two enhanced images are fused, so that a low-frequency enhanced fused image is obtained; inverse support value transform is carried out on the low-frequency enhanced fused image and a support value sequence image fused by a maximum method, so that a final fused image is obtained, the contrast between the targets and the background in the fused image is enhanced, and edge distortion is little.

The invention discloses a segmenting method of a fingerprint image. The segmenting method specifically comprises the following steps: a, reading in the fingerprint image; b, quickly cutting the fingerprint image; c, transforming in reverse colour; d, equalizing; e, carrying out top-hat transformation; f, calculating the characteristic quantity in blockst; g, segmenting in blocks; h, processing the image according to morphology; and i, obtaining the segmented fingerprint image. The segmenting method of the finger image is applicable to the fingerprint image poorer in quality, can correctly segment the fingerprint, and is higher in reliability; the ISODATA (Iterative Self-organizing Data Analysis Techniques Algorithm) clustering algorithm is adopted and used for segmenting in blocks, and the clustering speed is higher; the quick cutting way is carried out, thereby the processing time is reduced; the block segmenting mode and pixel segmenting mode are combined, thus the segmented fingerprint is relatively smooth in contour; and the image equalizing mode and the top-hat transformation mode are adopted to enhance the image, and as a result, the segmenting is more efficient.

The invention relates to a method for measuring image definition by utilizing multi-scale morphological characteristics. The method comprises the following steps of: (1) respectively extracting multi-scale bright characteristics and dark characteristics of an image through multi-scale structuring elements by utilizing top-hat transform; (2) calculating final bright characteristics and dark characteristics synthesized with multi-scale characteristics by utilizing the extracted multi-scale bright characteristics and dark characteristics; (3) obtaining the final multi-scale characteristics for the calculation of definition according to the final bright characteristics and dark characteristics by an maximum value operation; and (4) calculating gray average of the final multi-scale characteristics and taking the gray average as the measurement of the image definition. According to the method disclosed by the invention, the definition of images can be correctly measured, the change of definition is more sensitive, and the definition resolving capability is better; and in addition, the method disclosed by the invention can be widely applied to various image-based application systems and has wide market and higher application value.

The present invention relates to an infrared weak target detection method of utilizing sample characteristic learning classification. The method is characterized by setting an area containing a target as a positive sample and a background area as a negative sample, and obtaining enough positive and negative samples with labels from an actual infrared image, and comprises the steps of firstly analyzing the samples at the four aspects of gray distribution, edge, information entropy and energy, and extracting seven characteristics of fitting residual, center surrounding contrast ratio, the radius of an edge fitting circle, the center offset amount of the edge fitting circle, the center distance variance of the edge fitting circle, the reference information entropy contrast ratio and the texture energy contrast ratio; then selecting an optimal characteristic subset from all characteristics by the modes of package type selection and forward search and by taking the area below a subject performance curve as an evaluation index; and then extracting the optimal characteristic subset of the positive and negative samples to train a support vector machine classifier, and supervising the learning; and finally, carrying out the top-hat transform pre-processing on the image to obtain a candidate target, and discriminating and screening via the classifier to obtain a final detection result.

The invention discloses a ship wake detection method based on top hat transformation and Radon transformation, and the method comprises the steps: obtaining an original image, processing the originalimage based on the top hat transformation, and obtaining an image with improved contrast; and performing Radon transformation on the image with the improved contrast, extracting a peak point, obtaining wake position information, and marking a wake in the image. The method provided by the invention has a contrast improvement effect on both the bright wake and the dark wake, and the dark line feature of the dark wake can be converted into the bright line feature; radon transformation is carried out on the image with the improved contrast ratio, and a peak point in a Radon transformation domain is highlighted, so that the detection performance of the ship wake is improved; the method can be applied to ship wake detection in Tiangong-2 interferometric three-dimensional imaging microwave altimeter images.

The invention provides an infrared complicated background inhibiting method based on combined filtering. The method comprises the following steps of: S1, determination of the dimensions of morphological structural elements; S2, spatial background inhabitation: spatial primary inhabitation of an infrared complicated background is realized by utilizing the gray-scale morphological top-hat transform on the basis of the self-adapting structural elements; S3, non-subsampled contourlet transform: the non-subsampled contourlet first-level factoring is carried out on infrared images subjected to background primary inhibition by a space domain method, wherein a bandpass subband is factored into four high-frequency directions; S4, high-frequency coefficient reconstruction: the low-frequency influence is eliminated, and each high-frequency direction coefficient is calculated again on the basis of neighborhood average values; S5, high-frequency coefficient central vector construction: the average value of each high-frequency coefficient in each spatial position and in the four directions is calculated, and the high-frequency coefficient central vector is formed; and S6, non-subsampled contourlet domain background inhibition. The infrared complicated background inhibiting method solves the problem of infrared complicated background inhibition in typical environments, and lays a good foundation for the subsequent target detection.

The invention relates to a three-dimensional CT defect extracting and marking method for a solid engine and belongs to the field of nondestructive detection. In terms of a series of solid engine propellant hole defects, when defect segmentation is conducted, reconstructed volume data are layered firstly, operations such as filtering, morphological processing, threshold segmentation, defect extracting, defect marking and three-dimensional visualization are conducted on a set of obtained two-dimensional slice data, and defect information is extracted and marked. During morphological processing, top-hat transformation is conducted on each image layer by adopting a square structure operator which is slightly smaller than a central star hole 10. During defect extracting, a defect area is extracted by adopting the maximum connected region method. Through the three-dimensional CT defect extracting and marking method for the solid engine, hole defects in an image can be effectively segmented, extracted and marked.

The invention provides a detection algorithm for grid line offset on the surface of a cell. The detection algorithm includes the steps of performing preprocessing of image correction and image conversion on each frame of acquired image to obtain a target image of a photovoltaic cell; carrying out threshold segmentation of the target image, filling a grid line part, then carrying out shape conversion to eliminate the influence of defects at welding spots of the cell on the outline extraction, and accurately extracting the outline 1 of the edge of the cell; processing the target image by using grayscale top-hat transform to obtain a cell image only containing the grid line part; enhancing the image to eliminate the influence of a discontinuous grid line caused by the welding spots; carryingout threshold segmentation, and filling a gate line part to extract the outline 2 of the outermost peripheral gate line of the cell; and calculating the minimum distance between the outline 1 of the edge and the outline 2 of the outermost peripheral gate line of the cell, setting a threshold and comparing the minimum distance with the threshold to realize the detection of the grid line offset on the surface of the photovoltaic cell.

The invention discloses a two-phase flow pattern feature extraction method. The invention belongs to the field of flow pattern recognition in metallurgy and chemical engineering. the method comprisesthe following steps: firstly, by adopting an existing visualization technology, for example, acquiring a two-phase flow mixed color image by a high-speed camera; preprocessing a color image by utilizing top hat transformation and image filtering in existing software matlab; after preprocessing, extracting RGBvalues of a representative target and a representative background respectively by adoptingthe existing Colorpix software; adopting a support vector machine (SVM) algorithm to segment a target and a background in a two-phase flow image, extracting the target or the background, and finallyrealizing target image recognition through an rgb2gram function in Matlab software and binarization threshold processing, so that two-phase flow distribution is reflected more directly. According to the method, the space-time distribution diagram of the two-phase flow can be obtained more directly. The flow pattern feature extraction method is simple and short in operation time, the image recognition precision is improved, and the subjectivity of selection of an existing threshold method is also avoided.

The invention discloses a method and a system for automatically extracting a defect area of an X-ray image of an automobile hub. The method comprises the steps: firstly, obtaining an X-ray image of anautomobile hub, and constructing structural elements; carrying out top-hat transformation and top-hat reconstruction transformation on the X-ray image, and carrying out expansion reconstruction operation by taking a top-hat reconstruction transformation result as a mark and a top-hat transformation result as a template; performing binarization processing on the expansion reconstruction result toobtain a preliminary defect area of the hub; performing feature elimination on the preliminary defect area to obtain a real defect area of the hub; and obtaining the minimum enclosing rectangle of thereal defect area, namely the defect area of the hub X-ray image. When the method is used for extracting the defect area of the hub, an operator does not need to specify an interested area in advance,the defect area can be directly and automatically determined on the X-ray image of the detected hub, the method and system are not influenced by links such as wheel type recognition and an area tracking matching algorithm, and the extraction efficiency and accuracy of the defect area can be greatly improved.

The invention discloses an interactive extraction method and system for a defect area of an X-ray image of an automobile hub. The method comprises the following steps: firstly, acquiring an X-ray image of an automobile hub and a minimum enclosing rectangle of a mark area on the image; constructing structural elements according to the minimum bounding rectangle; carrying out top-hat transformationand top-hat reconstruction transformation on the X-ray image by adopting structural elements, carrying out expansion reconstruction operation of mathematical morphology according to a transformation result, and carrying out binarization processing on an obtained expansion reconstruction operation result; and analyzing the binarization processing result to obtain a defect area to be extracted. By adopting the defect area interactive extraction method provided by the invention, rapid and accurate separation of the whole defect area can be achieved as long as an operator randomly marks the defectposition by using a mouse, so that the speed and the efficiency of quantitative analysis of defects by detection personnel are greatly improved, and the method and system have the advantages of simplicity and convenience in operation and accuracy in identification.

The invention discloses a small target detection method based on joint edge filtering morphology in the technical field of image processing, and aims to solve the technical problems of high small target identification difficulty and low accuracy in the prior art. The method comprises the following steps: operating local pixels of an original image by using a joint edge filtering function to generate a structural element matrix of the original image; performing joint edge filtering morphological corrosion on the original image by using the structural element matrix of the original image to obtain a corroded image as a background image; performing top-hat transformation based on the original image and the background image to obtain a result image; and performing target recognition on the result image by using the adaptive threshold.

The invention discloses a sea surface infrared image enhancement method, which comprises the following steps of: inputting a sea surface infrared image to be enhanced, and performing top-hat transformation on the sea surface infrared image to be enhanced to obtain a region of interest; performing adaptive segmentation on the region-of-interest image to obtain a radiation source binarization mark image of the sea surface infrared image to be enhanced, determining a radiation source region image by using the radiation source binarization mark image, and constructing a radiation source suppression image according to the radiation source binarization mark image; determining a land-air area local illumination map and a seawater area local illumination map of the sea surface infrared image by taking a sea-sky line as a reference; constructing a local guide map according to the local illuminance map of the land-air area and the local illuminance map of the seawater area, and performing background enhancement on the sea surface infrared image to be enhanced to obtain a background enhancement map of the sea surface infrared image; and superposing the radiation source area image to the sea surface infrared image background enhancement image to obtain an enhancement result of the sea surface infrared image.