71 results about "Infrared image segmentation" patented technology

The invention relates to an image fusion of sequence infrared and visible light based on region segmentation. The invention is characterized in that infrared images are segmented into different regions according to the interframe objective change situation and gray scale change degree thereof; the images are dissembled to different frequency domains in the different directions by utilizing non-sub-sampling Contourlet transform; different fusion rules are selected based on the characteristics of different regions in the different frequency domains; and image reformation is carried out for the processed coefficients to obtain the final fusion result. The method takes the information of the characteristics in one of the regions into consideration, so the algorithm is capable of effectively reducing the rate of fusion image error due to the noise and low matching precision, and has stronger robustness.

The invention discloses an infrared image segmentation method based on multiple threshold values and self-adaptation fuzzy clustering. The main problem that according to an existing multi-threshold-value segmentation method, in the segmentation process, because false peak interference exists, the segmentation result is not ideal is solved. The method includes the steps that (1), an original infrared image is input; (2), a multi-threshold-value algorithm to which a control factor is introduced is used for coarsely segmenting the original infrared image; (3), morphology smoothing is performed on the coarsely segmented image; (4), a clustering center needed for fine segmenting the image is randomly selected, and the clustering number is determined according to a self-adaptation function; (5), fuzzy clustering is performed on pixel points of the smoothed image, so that a final segmentation result image is obtained. While the segmentation efficiency is guaranteed, the segmentation accuracy can be improved, and the infrared image segmentation method has the advantages of achieving a clear outline of the segmentation result and a complete target, and can be effectively applied to precise infrared guiding and target recognizing and tracking.

The invention discloses an infrared image dividing method and system for power system equipment based on wavelet analysis. Firstly, wavelet conversion, fuzzy entropy, a genetic algorithm and a mathematical morphology are used for carrying out image processing on an infrared image of the power system equipment. The method comprises the following steps: firstly, carrying out thermal failure detection on the power system equipment by an infrared thermal imager to obtain a thermal image; eliminating mixed noise of the thermal image by using the wavelet conversion to inhibit background interferences and enhance a target; carrying out combined optimization operation by applying the fuzzy entropy and the genetic algorithm to determine an optimal threshold value; extracting a target; solving a discontinuous boundary problem by using a waterline region dividing method of the mathematical morphology and dividing the image so as to find a largest communication region; and separating a target region. Finally, a position of a failure point of the power equipment and an element with a failure can be judged clearly according to the separated target region; an accident is prevented form occurring and overhauling under poweroff is not blindly carried out; the operation reliability of a power system is improved.

The embodiment of the invention discloses a histogram equalization method for an infrared image. The histogram equalization method for the infrared image comprises the following steps: reading the infrared image; segmenting the infrared image into at least two sub-images; performing histogram equalization processing on each sub-image to obtain at least two histogram equalization sub-images; stitching the histogram equalization sub-images. According to the method provided by the embodiment of the invention, segmentation processing is performed on the infrared image to obtain the multiple sub-images, and moreover, platform histogram equalization enhancement is performed on each sub-image, and then, the sub-images after being subjected to histogram equalization are stitched to be fused into the final histogram equalized infrared image, and thus, when the histogram equalization processing is performed, parallel processing can be performed on the multiple sub-images; then, fusion processing is performed on the image, so that the processing speed of the image is greatly quickened.

The invention provides a small object detection method based on random sampling and sparse matrix restoration in infrared scene. The method includes the following steps: conducting random sampling at the position of each pixel of a single frame infrared image, acquiring an infrared image having random characteristics; conducting patch transformation on the infrared image after random sampling, segmenting the infrared image after random sampling into a plurality of small images having no overlapped regions, and conducting 1 dimensional vectorization processing to obtain a 2 dimensional matrix after patch transformation; analyzing main components of the 2 dimensional matrix after patch transformation to obtain a sparse matrix and a low-rank matrix; conducting image restoration on the sparse matrix by applying patch inverse transformation, and separately obtaining a corresponding infrared small object image and an infrared image background; using the low-rank matrix to determine a segmentation threshold of the infrared small object detection, and conducting image segmentation on the infrared small object image based on the segmentation threshold to detect the infrared small object. According to the invention, the method is simple and requires short time to operate.

The invention relates to a foam infrared image segmentation method based on NSST saliency detection and image segmentation, and the method comprises the steps: firstly carrying out the NSST decomposition of a foam infrared image, and obtaining a low-frequency sub-band image and a multi-scale high-frequency sub-band; secondly, performing saliency detection on the low-frequency sub-band image by adopting a GBVS algorithm to obtain a saliency value and a visual saliency map; thirdly, calculating a threshold value and a scale correlation coefficient for each high-frequency direction sub-band coefficient, and removing a noise coefficient, a non-linear enhanced edge coefficient and a weak edge coefficient; and finally, performing image segmentation on the NSST reconstructed image in combinationwith visual saliency to obtain a segmentation result. The method is strong in anti-interference capability and high in segmentation precision.

An insulator image segmentation method based on distance transformation and a marker watershed algorithm relates to an insulator infrared image segmentation method. A conventional threshold segmentation method has a high error rate and is not suitable for image segmentation of insulators. The method includes: a watershed algorithm is adopted for image segmentation after preprocessing operation; Euclidean distance transform is performed on that complemented binary image, each pixel is assigned a distance from its nearest background pixel, the processed image is inverted and the local minimum value of the string region is found as a mark, and the region between two adjacent marks is taken as the insulator piece division line, so that only a unique internal mark and a partial background are ensured in each region, and the insulator piece region is effectively prevented from being separated by over-segmentation. The technical scheme is simple and convenient to realize. At the same time, the method adopting marker control and distance transform is used to effectively alleviate noise interference, prevent the image from being over-segmented into many tiny regions, and significantly improve the efficiency and accuracy of image segmentation.

The invention discloses an infrared image segmentation method based on improved FCM (fuzzy C-means) and mean drift, and mainly solves the problems over-segmentation of a segmentation result due to the fact that local convergence is easily caused in a segmentation process in a conventional mean drift segmentation method. The infrared image segmentation method comprises steps as follows: (1), an original infrared image is input; (2), the original infrared image is subjected to primary segmentation with a mean drift algorithm; (3), a clustering center and a clustering number which are required by secondary image segmentation are determined with a minimum / maximum method; (4), the result image after primary segmentation is converted into an initial value of secondary segmentation; (5), pixel points of the initial value of secondary segmentation are subjected to fuzzy classification; and (6), boundaries of different areas are sketched, and an image segmentation result is output. According to the method, the segmentation accuracy is improved while the segmentation efficiency is guaranteed, and the method has the advantages of smooth edges and clear contour of the segmentation result and can be effectively applied to military or civil aspects of infrared precision guide, target recognition and tracking and the like.

The invention provides an atmospheric scattering model based infrared image enhancing method. The method includes steps of performing inversion operation on an infrared image so as to convert the infrared image into a virtual foggy image by utilizing the similarity between an inversion infrared image and a foggy image; utilizing a quadtree decomposition technology to decompose the inversion infrared image into a series of sub blocks and performing independent defogging treatment within each sub block; performing inversion on the inversion infrared image after defogging treatment again and obtaining an enhanced infrared image; based on an orientation total variation model, performing edge repairing on the infrared image subjected to enhancement; performing luminance adjustment on the infrared image subjected to enhancement based on a Retainex model, and thus the final enhanced infrared image is obtained. By adopting the method provided by the invention, a large amount of scene details hidden in the infrared image can be restored and no negative effect is caused.

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 relates to an infrared image segmentation and fusion method based on an inspection robot, which comprises the steps of enhancing and marking equipment edges in an infrared image by usingan MSR algorithm to obtain an enhanced infrared image; performing segmentation on the enhanced infrared image by adopting MUL-PCNN to obtain a segmented infrared image; detecting and extracting equipment edges in the segmented infrared image and a visible image by using Wavelet-Canny; performing feature point purification on the equipment edges by adopting an improved ASURF algorithm, and constructing an affine transformation matrix; and fusing grayscale information in the infrared image into the visible image by adopting bicubic interpolation based on the reliability so as to realize accurate fusion of the infrared image and the visible image. Compared with the prior art, the infrared image segmentation and fusion method has the characteristics of accurate positioning, complete segmentation, high matching accuracy and the like on the aspects such as fault point temperature reading and positioning, equipment segmentation and fault point matching and fusion, and has certain applicationvalues.

The invention discloses a superspeed impact damage area feature strengthening method. The method comprises the following steps: extracting typical transient thermal response of a defect; obtaining an infrared reconstruction image; separating a background area and a defect area of the infrared reconstructed image; constructing an infrared image segmentation function under the guidance of noise removal, detail reservation and edge maintenance; combining a multi-objective optimization algorithm with the segmentation model to realize one-time segmentation of test piece defects in the infrared reconstructed image; and performing category division on the pixel points according to the distance from the pixel points in the infrared image to the clustering center to obtain a segmented image of the damage defect in the infrared detection image. According to the method, defect segmentation in the infrared reconstructed image is carried out by utilizing a multi-objective optimization theory, objective functions are respectively constructed aiming at a noise problem and an edge blurring problem, the segmentation precision of the damage defect area is improved, the defect detection rate is high, the false detection rate is low, the ultra-high-speed impact damage defect area in the infrared reconstructed image is convexly strengthened, and quantitative research of complex defects is facilitated.

The invention provides an improved infrared image segmentation algorithm based on the Otsu method, and the algorithm can solve a problem of poor segmentation effect caused by sky exposure to the greatest extent, and can enable the segmented target to maintain a more complete shape. Because the contrast of an infrared image is low relatively and the gray scale range is relatively small, the Otsu method cannot achieve the good segmentation of a target. Through giving consideration to the impact on the segmentation from the image gray scale, the number of background pixels of an image and the number of target pixels, the algorithm achieves the improvement of a formula for solving a variance in a conventional Otsu method, irons out the defects that an infrared image segmentation algorithm based on the Otsu method cannot obtain a better segmentation result under the condition of sky exposure and smaller contrast between the sky and the target, and enables the segmentation effect to be goodin a region where the contrast between the target and the sky background is larger.

The invention discloses a segmentation strengthening method for a damage detection image of an aerospace composite material. The segmentation strengthening method comprises the following steps: extracting typical transient thermal response of a defect; obtaining an infrared reconstruction image; utilizing multi-target to measure segmentation performance of three aspects ofnoise removal, detail reservation and edge maintenance, and solving the weight coefficient of each segmentation performance; constructing infrared image segmentation function data under the guidance of three purposes of noise removal, detail reservation and edge maintenance; obtaining a weight coefficient of the target function; carrying out image segmentation on the full-pixel infrared image obtained through reconstruction; and achieving infrared full-pixel image segmentation on the image segmentation layer to obtain a segmented image of the defect. According to the method, the multi-objective optimization theory is utilized to segment the damage defect area in the infrared reconstructed image, objective functions are respectively constructed for the noise problem and the edge blur problem, the area segmentation precision is improved, the false detection rate is reduced, the readability of the damage defect area is effectively enhanced, and quantitative research of complex defects is facilitated.

The invention relates to a PSO based infrared image segmentation method of Niblack power equipment. The method comprises the following steps that 1) an infrared image is obtained and divided into q non-overlapped continuous rectangular neighborhoods, and a grayscale mean value and a grayscale standard deviation of each rectangular neighborhood are calculated; 2) a threshold optimization searching interval corresponding to each rectangular neighborhood is obtained according to a set step length, a q-dimensional particle swarm solution space is formed, an inter-class deviation serves as a PSO fitness function, and the q-dimensional particle swarm solution space is searched automatically for an optimal segmentation threshold T*, which enables a maximal inter-class deviation, corresponding to each rectangular neighborhood; and 3) and each rectangular neighborhood is binarized according to the optimal segmentation threshold, obtained in the step 2), of the rectangular neighborhood. Compared with the prior art, the problem of infrared image over-segmentation caused by a traditional global threshold segmentation method is solved.

The invention discloses a method for carrying out infrared image segmentation by virtue of an active outline. According to the method disclosed by the invention, the grey level distribution information of an image is adequately utilized by constructing an active outline model based on the combined drive of a local entropy and a local standard deviation, meanwhile, the infrared image is segmented by introducing a Gaussian kernel function to much accurately driving outline evolution in a local area. The method disclosed by the invention is capable of effectively realizing segmentation for the infrared image, so as to obtain a complete and correct target outline.

The invention discloses a Mask-RCNN-based electric power equipment infrared image segmentation method, and the method comprises the following steps: S1, building a data set of an electric power equipment infrared image, and marking a training set and a test set; S2, constructing a vertical deep learning model; S3, setting initial hyper-parameters and the number of iterations of the model; S4, using the training set marked in the step S1, and inputting the training set into the constructed model for training; s5, evaluating the performance of the model obtained by the training in the step S4 byadopting the test set marked in the step S1 every 2000-3000 iterations; s6, when the number of iterations reaches a set value, stopping training, and screening out a deep learning model with the optimal performance; and S7, inputting the infrared image of the to-be-tested power equipment into the trained optimal deep learning model for processing, and obtaining a segmentation result. According tothe method, the segmentation precision is remarkably improved, the original color information of the target equipment is reserved, the temperature information can be obtained, and a basis is providedfor fault diagnosis.

The invention discloses a transformer substation infrared image segmentation method based on an improved two-dimensional Otsu algorithm, solves the problem that the conventional two-dimensional Otsu algorithm has error segmentation and incomplete target segmentation; the method lies in that the weight adaptive algorithm of the morphology is adopted for performing denoising on an input image, and the improved two-dimensional Otsu algorithm is finally adopted for performing image segmentation. According to the transformer substation infrared image segmentation method provided by the invention, target regions can be divided completely and effectively, and a good foundation is laid for characteristic extraction of subsequent target regions.

The invention discloses an automatic identification method for a damage area of an aerospace composite material. The method comprises the following steps: extracting typical transient thermal response of each type of defects; obtaining an infrared reconstruction image; obtaining a low-quality infrared reconstructed image; solving a weight coefficient of three segmentation performances of noise removal, detail reservation and edge maintenance; constructing an infrared image segmentation function; obtaining a weight coefficient of the objective function for realizing each segmentation performance; constructing a full-pixel infrared image segmentation target function, and performing image segmentation on the reconstructed full-pixel infrared image by using the segmentation model; and achieving infrared full-pixel image segmentation on the image segmentation layer to obtain a segmented image. According to the method, defect segmentation in the infrared reconstructed image is carried out by utilizing a multi-objective optimization theory, objective functions are respectively constructed for a noise problem and an edge blurring problem so as to improve segmentation precision, a high defect detection rate is ensured, a false detection rate is reduced, a damage defect area in the reconstructed image is effectively extracted, and quantitative research of complex defects is facilitated.

An embodiment of the invention discloses an infrared image segmentation method. The method comprises: obtaining a frame of infrared image data; performing linear stretch processing and histogram equalization processing on the infrared image data to obtain histogram equalization image data; performing clustering analysis on the histogram equalization image data to obtain a plurality of clusters; constructing a measure function, and obtaining a corresponding cluster when the value of the measure function is maximal from the clusters; calculating a threshold according to the corresponding cluster when the value of the measure function is maximal; and binarizing the infrared image data by using the threshold. The infrared image segmentation method provided by the embodiment of the invention is simple to operate and quick in clustering.

The invention discloses an infrared image segmentation method based on Otsu and improved Bernsen, and the method specifically comprises the steps: firstly carrying out the contrast extension transformation preprocessing of an inputted original image, wherein the preprocessed image meets the histogram equalization; calculating a segmentation threshold GT by adopting an Otsu method, and segmenting the image into a background region and a target region; performing Gaussian smoothing filtering on g (i, j) in the target area in a (2w + 1) * (2w + 1) window; pre-judging whether the contrast in a local window of the image exceeds a threshold range or not by adopting an improved Bernsen method to obtain a threshold T (i, j) of the image; and finally, performing point-by-point binarization on the image by using T (i, j) to obtain a binarized image b (i, j). According to the method, the problem that over-segmentation occurs when an Otsu method is used for processing images with uneven illumination and complex backgrounds is solved; the problems of serious image noise and edge loss after segmentation caused by forced binarization of a Bernsen method are overcome, and the two methods are combined to improve the efficiency and accuracy of image segmentation.

The invention discloses an improved region growing method for infrared image segmentation of power equipment, and the method comprises the steps: converting an infrared image of power equipment into agray level image, employing an edge detection segmentation method to carry out the transverse and longitudinal differential operation of the gray level image, obtaining edge pixel points of the powerequipment, carrying out the matching, and preliminarily determining a possible region of the equipment; traversing the gray scale image of the area where the equipment is possibly located by using aneffectiveness detection operator to obtain an effective area, and completing preliminary division of the equipment area; automatically selecting seed points; and taking the access points obtained bydifferential operation in the transverse and longitudinal directions and the preliminarily divided equipment regions as constraints, carrying out equipment region growth operation, and removing burrsand gaps for the segmented image by utilizing morphological operation after growth is completed, thereby obtaining a final power equipment segmentation result. The method provided by the invention issimple in calculation process and good in effect, and can realize high-quality rapid automatic segmentation of large-scale infrared images in a power system.

The invention relates to an image fusion of sequence infrared and visible light based on region segmentation. The invention is characterized in that infrared images are segmented into different regions according to the interframe objective change situation and gray scale change degree thereof; the images are dissembled to different frequency domains in the different directions by utilizing non-sub-sampling Contourlet transform; different fusion rules are selected based on the characteristics of different regions in the different frequency domains; and image reformation is carried out for the processed coefficients to obtain the final fusion result. The method takes the information of the characteristics in one of the regions into consideration, so the algorithm is capable of effectively reducing the rate of fusion image error due to the noise and low matching precision, and has stronger robustness.

The invention discloses a segmentation identification method for an ultra-high-speed impact damage detection image. The method comprises the following steps: extracting typical transient response of each type of defects; obtaining an infrared reconstruction image; combining a multi-objective optimization algorithm with the segmentation model to separate a background area and a defect area of the infrared reconstructed image; constructing an infrared image segmentation function under the guidance of three purposes of noise removal, detail retention and edge retention; combining a multi-objective optimization algorithm with the segmentation model to realize one-time segmentation of test piece defects in the infrared reconstructed image, and adjusting weight vectors according to preferences; and carrying out category division on the pixel points to obtain a final segmented image. According to the method, the collision damage area in the infrared reconstructed image is segmented according to the multi-objective optimization theory, the objective function is constructed for the noise problem and the edge blurring problem so as to improve the segmentation precision, the defect detection rate is high, the false detection rate is reduced, the damage defect area in the reconstructed image is effectively extracted, and quantitative research of complex ultra-high-speed impact damage is facilitated.

The invention discloses an infrared and visible light image fusion method based on saliency fusion. Three extraction methods including salient target extraction based on infrared image segmentation and Bessel background reconstruction based on a quadtree, sparse salient target extraction based on structure low-rank coding and salient target multi-scale detail extraction based on Laplace are adopted, an infrared salient target is determined, and the infrared image and the visible light image are fused, so that the visual saliency characteristic of the target is obtained, and meanwhile, the detail information of the scene is reserved. According to the image fusion method provided by the invention, the defects of the infrared image can be overcome, and the advantages of visible light and infrared light can be exerted, so that the fused image has the advantages of infrared light and visible light at the same time, and the detection and reconnaissance capability of the system is improved.

The invention discloses an infrared image segmentation algorithm based on human eye visual characteristics, and belongs to the technical field of image segmentation algorithms. The objective of the invention is to solve the problems of limitation of an existing segmentation algorithm, influence of image noise and illumination, difficulty in human eye observation of a segmentation effect and the like. Firstly, human eye brightness perception characteristics and Weber-are utilized; And according to the Fecner law, properly processing an over-bright area and an over-dark area in the whole image,so that the target information is concentrated in a Weber area. Secondly, simulating a human eye vision system by utilizing the characteristic that the background intensity and the light intensity gradient are in a local linear relation in a logarithmic domain, establishing an HVS model, and segmenting the image into three regions; The segmented infrared picture effect conforms to the human eye perception characteristic, and the processing time is also improved to a certain extent.

The invention discloses a maximum entropy combined improved Niblet infrared image segmentation method, which is specifically implemented according to the following steps: step 1, inputting an originalimage, and performing median filtering denoising preprocessing on the original image; step 2, improving a neighborhood window and a correction coefficient in a Niblet method to obtain a local threshold TNiblack; step 3, carrying out threshold weighting on the optimal threshold TKapur and the local threshold TNiback obtained by the maximum entropy method, and determining the optimal threshold of segmentation; step 4, segmenting the image subjected to denoising preprocessing in the step 1 through an optimal threshold value. The problem of image over-segmentation or under-segmentation caused bydifficulty in threshold selection under the condition of complex illumination is solved.

The invention discloses a power equipment infrared image segmentation method based on deep learning and image gradient, and the method comprises the steps: carrying out the feature mining of an infrared image through deep learning, completing the first segmentation, combining the temperature features of the infrared image, completing the second segmentation through the gradient of the image, and effectively improving the boundary segmentation precision. Areas of key parts of the electrical equipment can be separated, original infrared image information of the electrical equipment is reserved,and direct and accurate data support is provided for diagnosis and analysis of infrared images; and the diagnosis efficiency and the intelligent level of the infrared image can be improved, the laborcost is reduced, and the efficiency and the accuracy of power equipment fault analysis work are improved.

The invention provides an improved infrared image segmentation algorithm based on the Otsu method, and the algorithm can solve a problem of poor segmentation effect caused by sky exposure to the greatest extent, and can enable the segmented target to maintain a more complete shape. Because the contrast of an infrared image is low relatively and the gray scale range is relatively small, the Otsu method cannot achieve the good segmentation of a target. Through giving consideration to the impact on the segmentation from the image gray scale, the number of background pixels of an image and the number of target pixels, the algorithm achieves the improvement of a formula for solving a variance in a conventional Otsu method, irons out the defects that an infrared image segmentation algorithm based on the Otsu method cannot obtain a better segmentation result under the condition of sky exposure and smaller contrast between the sky and the target, and enables the segmentation effect to be goodin a region where the contrast between the target and the sky background is larger.

The invention relates to an infrared image processing method and device, a computer device and a storage medium, and relates to the technical field of image processing. The infrared image processing method comprises the following steps: acquiring a to-be-processed initial infrared image, and segmenting the initial infrared image into a plurality of sub-interval images; splicing the sub-interval images to form a plurality of sub-images with different semantic tags; and performing gray scale enhancement on each sub-image, and performing linear weighted superposition on each enhanced sub-image toobtain an output infrared image. According to the invention, the occurrence of a grey degree supersaturation phenomenon can be avoided.