CN117474959B - Target object motion trail processing method and system based on video data - Google Patents
Target object motion trail processing method and system based on video data Download PDFInfo
- Publication number
- CN117474959B CN117474959B CN202311744298.2A CN202311744298A CN117474959B CN 117474959 B CN117474959 B CN 117474959B CN 202311744298 A CN202311744298 A CN 202311744298A CN 117474959 B CN117474959 B CN 117474959B
- Authority
- CN
- China
- Prior art keywords
- frame
- target object
- image sequence
- motion
- image
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000003672 processing method Methods 0.000 title claims abstract description 10
- 230000009466 transformation Effects 0.000 claims abstract description 60
- 238000000034 method Methods 0.000 claims abstract description 52
- 239000011159 matrix material Substances 0.000 claims abstract description 48
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 29
- 238000012545 processing Methods 0.000 claims abstract description 11
- 239000003086 colorant Substances 0.000 claims abstract description 8
- 238000004458 analytical method Methods 0.000 claims description 28
- 230000003287 optical effect Effects 0.000 claims description 26
- 230000000694 effects Effects 0.000 claims description 22
- PXFBZOLANLWPMH-UHFFFAOYSA-N 16-Epiaffinine Natural products C1C(C2=CC=CC=C2N2)=C2C(=O)CC2C(=CC)CN(C)C1C2CO PXFBZOLANLWPMH-UHFFFAOYSA-N 0.000 claims description 15
- 238000001514 detection method Methods 0.000 claims description 8
- 239000013598 vector Substances 0.000 claims description 8
- 239000006002 Pepper Substances 0.000 claims description 5
- 238000004590 computer program Methods 0.000 claims description 5
- 238000006073 displacement reaction Methods 0.000 claims description 5
- 235000002566 Capsicum Nutrition 0.000 claims description 4
- 235000016761 Piper aduncum Nutrition 0.000 claims description 4
- 235000017804 Piper guineense Nutrition 0.000 claims description 4
- 235000008184 Piper nigrum Nutrition 0.000 claims description 4
- 150000003839 salts Chemical class 0.000 claims description 4
- 230000001131 transforming effect Effects 0.000 claims description 4
- 238000011179 visual inspection Methods 0.000 claims description 3
- 238000012937 correction Methods 0.000 claims description 2
- 244000203593 Piper nigrum Species 0.000 claims 1
- 230000008569 process Effects 0.000 description 10
- 230000008859 change Effects 0.000 description 8
- 238000005516 engineering process Methods 0.000 description 5
- 230000006870 function Effects 0.000 description 4
- 241000722363 Piper Species 0.000 description 3
- 238000013178 mathematical model Methods 0.000 description 3
- 230000001105 regulatory effect Effects 0.000 description 3
- 230000005540 biological transmission Effects 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 230000002349 favourable effect Effects 0.000 description 2
- 230000002159 abnormal effect Effects 0.000 description 1
- 230000006399 behavior Effects 0.000 description 1
- 238000007906 compression Methods 0.000 description 1
- 230000006835 compression Effects 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000013144 data compression Methods 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 230000001066 destructive effect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 238000000691 measurement method Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000012216 screening Methods 0.000 description 1
- 230000035945 sensitivity Effects 0.000 description 1
- 238000012163 sequencing technique Methods 0.000 description 1
- 238000010008 shearing Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 230000001360 synchronised effect Effects 0.000 description 1
- 238000000844 transformation Methods 0.000 description 1
- 238000013519 translation Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/20—Analysis of motion
- G06T7/269—Analysis of motion using gradient-based methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/20—Analysis of motion
- G06T7/207—Analysis of motion for motion estimation over a hierarchy of resolutions
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/20—Analysis of motion
- G06T7/254—Analysis of motion involving subtraction of images
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/30—Determination of transform parameters for the alignment of images, i.e. image registration
- G06T7/38—Registration of image sequences
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/74—Image or video pattern matching; Proximity measures in feature spaces
- G06V10/761—Proximity, similarity or dissimilarity measures
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/40—Scenes; Scene-specific elements in video content
- G06V20/41—Higher-level, semantic clustering, classification or understanding of video scenes, e.g. detection, labelling or Markovian modelling of sport events or news items
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/40—Scenes; Scene-specific elements in video content
- G06V20/46—Extracting features or characteristics from the video content, e.g. video fingerprints, representative shots or key frames
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/40—Scenes; Scene-specific elements in video content
- G06V20/48—Matching video sequences
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/40—Scenes; Scene-specific elements in video content
- G06V20/49—Segmenting video sequences, i.e. computational techniques such as parsing or cutting the sequence, low-level clustering or determining units such as shots or scenes
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10016—Video; Image sequence
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Multimedia (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Computing Systems (AREA)
- Software Systems (AREA)
- Computational Linguistics (AREA)
- Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Databases & Information Systems (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Image Analysis (AREA)
Abstract
The invention provides a target object motion trail processing method and system based on video data, which relate to the technical field of motion data processing and comprise the following steps: the method comprises the steps of obtaining a video file, decomposing the video file into an initial image sequence through a decoder, identifying noise and denoising, generating a frame image sequence, adjusting RGB curves of the frame image sequence through a linear tool, and correcting colors to obtain a standard image sequence; identifying a key frame in a standard image sequence, analyzing the motion trail of a target object in the key frame and the speed difference between a current frame and a previous frame, and determining the motion difference of the target object; according to the motion difference degree, matching the characteristic points corresponding to the target object in the key frame with other images in the standard image sequence through the adjacent image matcher to obtain matching point pairs, determining a transformation matrix corresponding to the matching point pairs through a fitting iterative algorithm, summarizing all transformation matrices, determining an optimal transformation matrix, and obtaining the target object state estimation.
Description
Technical Field
The invention relates to the technical field of motion data processing, in particular to a target object motion trail processing method and system based on video data.
Background
With the development of society and the advancement of technology, the acquisition of video data becomes easier and easier. A large amount of video data is generated in the fields of monitoring, transportation, entertainment, etc., and contains abundant information. Wherein, the motion trail analysis of the target object has important significance for many applications,
in the prior art, CN113808167a discloses a volleyball motion trail detection method based on video data, which effectively detects and calculates volleyball motion trail from multi-view synchronous volleyball video captured by a camera by using a target detection model based on deep learning. The method comprises the following steps: volleyball target detection, motion trail matching screening and motion trail three-dimensional reconstruction.
In summary, although the prior art can analyze the target object based on the motion track in the video, the state prediction cannot be performed through the track of the target object in the video and the corresponding feature points, so a solution is needed to solve the problems in the prior art.
Disclosure of Invention
The embodiment of the invention provides a target object motion trail processing method and system based on video data, which are used for identifying the motion trail of a target to be identified according to video images and estimating the state of the target object.
In a first aspect of an embodiment of the present invention, a method for processing a motion trajectory of a target object based on video data is provided, including:
the method comprises the steps of obtaining a video file, decomposing the video file into an initial image sequence through a decoder, identifying noise in the initial image sequence, denoising, generating a frame image sequence, adjusting RGB curves of the frame image sequence through a linear tool, and correcting colors to obtain a standard image sequence;
identifying a key frame in the standard image sequence, analyzing a motion track of a target object in the key frame and the speed difference degree of the target object between a current frame and a previous frame by an optical flow method, and determining the motion difference degree of the target object between the current frame and the previous frame according to the motion track and the speed difference degree;
and according to the motion difference degree, matching the characteristic points corresponding to the target object in the key frame with the rest images in the standard image sequence through a preset adjacent image matcher to obtain matching point pairs, determining a transformation matrix corresponding to the matching point pairs through a fitting iterative algorithm, summarizing all transformation matrices, and determining an optimal transformation matrix to obtain the target object state estimation.
In an alternative embodiment of the present invention,
the obtaining the video file, decomposing the video file into an initial image sequence through a decoder, identifying noise in the initial image sequence and denoising, and generating a frame image sequence comprises the following steps:
acquiring a video file to be detected, identifying the format of the video file, selecting a corresponding decoder, initializing the decoder, distributing memory for the decoder, and setting the resolution, the frame rate and the color space of the decoder;
reading each frame of the video file, acquiring the type and the time stamp corresponding to each frame in the video file, inputting compressed frame data into a decoder to restore to an original image, performing inter-frame prediction according to the type corresponding to the original image, removing redundant information in the original image, and arranging the original image according to the time stamp to obtain an initial image sequence;
and carrying out pixel value analysis on each frame of original image in the initial image sequence, identifying the noise type of the original image through frequency domain analysis according to the result of the pixel value analysis, removing the corresponding noise, and removing the salt and pepper noise in the original image through visual inspection to obtain the frame image sequence.
In an alternative embodiment of the present invention,
the adjusting the RGB curve of the frame image sequence by a linear tool and correcting the color, the obtaining the standard image sequence comprises the following steps:
loading a frame image sequence, adding an adjustment layer in the layer, determining an RGB effect curve in the adjustment layer, adjusting the contrast of images in the frame image sequence by adjusting the shape of the RGB effect curve, modifying the brightness of the images in the frame image sequence by adjusting the top and bottom of the RGB effect curve, selecting a color channel in the RGB effect curve, adding a control point in the color channel, and adjusting the color intensity of the images in the frame image sequence by adjusting the position of the control point to obtain an adjustment image;
and evaluating the adjustment image through a histogram tool, determining whether the adjustment image meets the system requirement, if so, arranging the adjustment image according to the time stamp, outputting a standard image sequence, and if not, modifying the adjustment image according to the system requirement.
In an alternative embodiment of the present invention,
the step of identifying the key frame in the standard image sequence, analyzing the motion trail of the target object in the key frame and the speed difference degree of the target object between the current frame and the last frame by an optical flow method, and determining the motion difference degree of the target object between the current frame and the last frame according to the motion trail and the speed difference degree comprises the following steps:
Applying a difference analysis algorithm to each frame in the standard image sequence, extracting image characteristics of each frame image in the standard image sequence, judging characteristic similarity between continuous frames according to the image characteristics, and taking a frame with the highest similarity with a later frame image as a key frame;
applying an optical flow method to the standard image sequence, extracting characteristic points of each frame in the standard image sequence, calculating an optical flow field of the target object between continuous frames, connecting initial characteristic points corresponding to the target object and characteristic points in key frames, and drawing a motion track of the target object;
calculating a displacement vector of each characteristic point corresponding to the target object in the optical flow field, determining the speed of the target object between frames by combining the time interval between frames in the standard image sequence, and comparing the speed corresponding to the key frame with the speed corresponding to the last frame to obtain a speed difference degree;
and calculating the motion difference degree of the target object between the current frame and the previous frame according to the speed difference degree and the motion trail.
In an alternative embodiment of the present invention,
the calculating the motion difference degree of the target object between the current frame and the previous frame according to the speed difference degree and the motion trail comprises the following steps:
；
Wherein,Dthe degree of difference in motion is indicated,w 1 representing the weight of the velocity component in the horizontal direction,V x representing a velocity component in the horizontal direction,d x indicating the difference in position in the horizontal direction,αrepresenting the horizontal-direction scaling factor,w 2 representing the weight of the velocity component in the vertical direction,V y representing the velocity component in the vertical direction,d y a position difference in the vertical direction is indicated,βrepresenting the vertical scaling factor.
In an alternative embodiment of the present invention,
according to the motion difference, matching the feature points corresponding to the target object in the key frame with the rest images in the standard image sequence through a preset adjacent image matcher to obtain matching point pairs, wherein the step of obtaining the matching point pairs comprises the following steps:
applying a feature point detection algorithm to the key frame, extracting a first feature point corresponding to a target object in the key frame, calculating a feature descriptor corresponding to the first feature point, and adding information corresponding to the motion difference degree to the feature descriptor corresponding to the first feature point to obtain a motion feature descriptor;
initializing the adjacent image matcher according to the motion feature descriptors, correspondingly modifying indexes of the adjacent image matcher, dividing the feature points of other images except key frames in the standard image sequence into two parts based on division values according to the motion feature descriptors, wherein one part is the feature point with the division value, the other part is the feature point with the division value, repeating the division until the maximum depth is reached, obtaining a multi-dimensional tree subtree, and connecting the multi-dimensional tree subtrees to obtain a multi-dimensional tree;
For each first feature point, starting the first feature point from a root node of the multi-dimensional tree through the adjacent image matcher, comparing the value of the first feature point in a dividing dimension with the dividing value of a current node, determining a searching direction, and repeating searching until a leaf node of the multi-dimensional tree is reached;
and tracing back upwards from the leaf node, checking whether the feature points of the current node are closer to each other, updating nearest neighbor feature points and distances, and repeating the operation until all the first feature points are matched to obtain a matching point pair.
In an alternative embodiment of the present invention,
determining a transformation matrix corresponding to the matching point pair through a fitting iterative algorithm, summarizing all transformation matrices, determining an optimal transformation matrix, and obtaining a target object state estimation comprises the following steps:
randomly selecting one of the matching point pairs as an initial matching point pair, defining the iteration times of the fitting iterative algorithm, randomly selecting three matching point pairs in each iteration to construct an affine transformation matrix, solving, calculating the predicted position of the initial matching point pair under the affine transformation matrix condition, calculating the distance between the actual position of the initial matching point pair and the predicted position, and dividing the initial matching point pair into inner points if the distance is smaller than a preset distance threshold;
And for each matching point pair, determining whether the current matching point pair is an inner point, summarizing all the inner points, and re-estimating an affine transformation matrix by using the inner points through a least square method to obtain an optimal transformation matrix, and transforming the key frame according to the optimal transformation matrix to obtain the target object state estimation.
In a second aspect of the embodiment of the present invention, there is provided a target object motion trajectory processing system based on video data, including:
the first unit is used for acquiring a video file, decomposing the video file into an initial image sequence through a decoder, identifying noise in the initial image sequence, denoising, generating a frame image sequence, adjusting RGB curves of the frame image sequence through a linear tool, and correcting colors to obtain a standard image sequence;
the second unit is used for identifying key frames in the standard image sequence, analyzing the motion trail of a target object in the key frames and the speed difference degree of the target object between the current frame and the previous frame through an optical flow method, and determining the motion difference degree of the target object between the current frame and the previous frame according to the motion trail and the speed difference degree;
And the third unit is used for matching the characteristic points corresponding to the target object in the key frame with other images in the standard image sequence through a preset adjacent image matcher according to the motion difference degree to obtain matching point pairs, determining a transformation matrix corresponding to the matching point pairs through a fitting iterative algorithm, summarizing all transformation matrices and determining an optimal transformation matrix to obtain the target object state estimation.
In a third aspect of an embodiment of the present invention,
there is provided an electronic device including:
a processor;
a memory for storing processor-executable instructions;
wherein the processor is configured to invoke the instructions stored in the memory to perform the method described previously.
In a fourth aspect of an embodiment of the present invention,
there is provided a computer readable storage medium having stored thereon computer program instructions which, when executed by a processor, implement the method as described above.
In the invention, the video file is decomposed into the initial image sequence by the decoder, noise is identified and removed on the basis, the image quality and accuracy can be improved, RGB curves are adjusted by the linear tool, color correction is carried out, the contrast, brightness and color accuracy of images are improved, a more standard image sequence is generated, the motion track of a target object in a key frame is analyzed by identifying the key frame and applying an optical flow method, the motion information of the target object can be captured more accurately, the matching of the target object between the key frame and other images in a standard image sequence is realized by the adjacent image matcher and fitting iterative algorithm, and the transformation matrix is estimated.
Drawings
FIG. 1 is a flow chart of a target object motion trail processing method based on video data according to an embodiment of the invention;
fig. 2 is a schematic structural diagram of a target object motion trail processing system based on video data according to an embodiment of the present invention.
Detailed Description
For the purpose of making the objects, technical solutions and advantages of the embodiments of the present invention more apparent, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention, and it is apparent that the described embodiments are only some embodiments of the present invention, not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
The technical scheme of the invention is described in detail below by specific examples. The following embodiments may be combined with each other, and some embodiments may not be repeated for the same or similar concepts or processes.
Fig. 1 is a flow chart of a target object motion trail processing method based on video data according to an embodiment of the invention, as shown in fig. 1, the method includes:
S1, acquiring a video file, decomposing the video file into an initial image sequence through a decoder, identifying noise in the initial image sequence, denoising, generating a frame image sequence, adjusting RGB curves of the frame image sequence through a linear tool, and correcting colors to obtain a standard image sequence;
the RGB curve refers to the brightness or color value of the red, green and blue channels in the image at different brightness levels, each point on the curve represents a brightness level, the shape of the curve affects the hue, contrast and saturation of the image, the color effect of the image can be changed to be brighter, more contrasted and brighter by adjusting the curve, the linear tool refers to software or an application program for image processing, the function of adjusting the image attribute is provided, and the function of adjusting the image attribute is usually included, the frame image sequence refers to an image sequence composed of a series of image frames, usually used for representing video, each frame is a single image in the video, and the frame image sequence is formed by arranging the images according to time sequence.
In an alternative embodiment of the present invention,
The obtaining the video file, decomposing the video file into an initial image sequence through a decoder, identifying noise in the initial image sequence and denoising, and generating a frame image sequence comprises the following steps:
acquiring a video file to be detected, identifying the format of the video file, selecting a corresponding decoder, initializing the decoder, distributing memory for the decoder, and setting the resolution, the frame rate and the color space of the decoder;
reading each frame of the video file, acquiring the type and the time stamp corresponding to each frame in the video file, inputting compressed frame data into a decoder to restore to an original image, performing inter-frame prediction according to the type corresponding to the original image, removing redundant information in the original image, and arranging the original image according to the time stamp to obtain an initial image sequence;
and carrying out pixel value analysis on each frame of original image in the initial image sequence, identifying the noise type of the original image through frequency domain analysis according to the result of the pixel value analysis, removing the corresponding noise, and removing the salt and pepper noise in the original image through visual inspection to obtain the frame image sequence.
The color space is a mathematical model describing and representing the colors of an image, each pixel in the image can be represented as a vector in the color space, the inter-frame prediction is a technology in video compression, data compression is achieved by encoding the difference between adjacent frames in a video sequence, inter-frame prediction is generally used for utilizing the space-time correlation in the video sequence, so as to reduce the data quantity required to be stored or transmitted, the pixel value analysis refers to the process of analyzing the numerical value of each pixel in the image, and comprises the research on the characteristics of brightness, color intensity and the like of the pixel, and the salt-and-pepper noise is a common noise type in the image and is represented as a black-and-white point suddenly appearing in the image, which is possibly caused by faults in the image acquisition process, errors in transmission or problems in image storage.
Acquiring a video file to be detected, identifying the format of the video file by using a video processing library, selecting a corresponding video decoder according to the video format, initializing the selected video decoder, distributing necessary memory space for the decoder, and setting parameters such as resolution, frame rate, color space and the like of the decoder according to the requirement;
And circularly reading each frame of the video file, inputting the compressed frame data into a decoder, decoding and restoring the frame data into original images, obtaining the type and the time stamp corresponding to each frame, carrying out inter-frame prediction on the original images of each frame, reducing the storage space of the images by residual information obtained by differential calculation, removing the redundant information, and sequencing the frame images according to the time stamp to obtain an initial image sequence.
In this embodiment, by selecting a suitable decoder and setting related parameters, successfully decoding a video file, restoring compressed frame data to an original image is helpful to preserve important information in a video and provide clear input for subsequent track analysis, and by using inter-frame prediction technology, space-time correlation between adjacent frames in a video sequence is effectively utilized, so that the amount of data required for storage and transmission is reduced, the efficiency of a system is improved, and by using pixel value analysis and frequency domain analysis, the noise type in an image can be detected and identified, the image quality is improved, and especially in target object track analysis, the influence on the target position can be reduced by removing salt and pepper noise and the like.
In an alternative embodiment of the present invention,
the adjusting the RGB curve of the frame image sequence by a linear tool and correcting the color, the obtaining the standard image sequence comprises the following steps:
loading a frame image sequence, adding an adjustment layer in the layer, determining an RGB effect curve in the adjustment layer, adjusting the contrast of images in the frame image sequence by adjusting the shape of the RGB effect curve, modifying the brightness of the images in the frame image sequence by adjusting the top and bottom of the RGB effect curve, selecting a color channel in the RGB effect curve, adding a control point in the color channel, and adjusting the color intensity of the images in the frame image sequence by adjusting the position of the control point to obtain an adjustment image;
and evaluating the adjustment image through a histogram tool, determining whether the adjustment image meets the system requirement, if so, arranging the adjustment image according to the time stamp, outputting a standard image sequence, and if not, modifying the adjustment image according to the system requirement.
The adjustment layer is a special layer added in the image editing software, and is mainly used for adjusting the layer or the image below without directly changing the pixel data of the original image, and the bottom image is affected in a non-destructive manner through the self adjustment parameters, wherein the color intensity refers to the saturation or intensity degree of the color in the image.
Loading the frame image sequence, creating a new adjustment layer, finding RGB effect curve adjustment options in the adjustment layer, determining the shape of the RGB effect curve, adjusting the contrast of images in the frame image sequence by adjusting the shape of the RGB effect curve, modifying the brightness of images in the frame image sequence by adjusting the top and bottom of the RGB effect curve, selecting color channels including red, green and blue channels in the RGB effect curve, adding control points in the selected color channels, and adjusting the color intensity of the images in the frame image sequence by adjusting the positions of the control points;
and evaluating the adjustment image by using a histogram tool, displaying the distribution conditions of different brightness levels in the image through the histogram, judging whether the adjustment image meets the system requirement according to the information of the histogram, arranging the adjustment image according to the time stamp if the adjustment image meets the system requirement, and modifying adjustment parameters such as RGB effect curves according to the system requirement if the adjustment image does not meet the system requirement, and finally arranging the adjustment image meeting the system requirement according to the time stamp to form a standard image sequence.
In this embodiment, the contrast of the images in the frame image sequence may be adjusted by adjusting the shape of the RGB effect curve, so as to help to highlight or smooth the brightness difference in the images, so that the features of the target object are more highlighted, the brightness of the images in the frame image sequence may be modified by adjusting the top and bottom of the RGB effect curve, so as to help to adapt to different environmental conditions or the characteristics of the target object, and by histogram analysis, whether the adjusted images meet the system requirements may be evaluated.
S2, identifying a key frame in the standard image sequence, analyzing a motion trail of a target object in the key frame and a speed difference degree of the target object between a current frame and a previous frame through an optical flow method, and determining the motion difference degree of the target object between the current frame and the previous frame according to the motion trail and the speed difference degree;
the optical flow method is a technology for estimating the motion of pixels in an image sequence, is generally used for tracking objects, analyzing motion scenes, or providing motion information for object detection, the speed difference degree refers to the speed change between adjacent frames, and generally represents the difference of the motion speeds of objects or pixels between two continuous frames, the motion difference degree refers to the overall motion difference between adjacent frames, and not only comprises the speed change of the objects, but also considers the factors such as the shape, the track and the like of the objects, and is used for comparing the motion characteristics between different objects or scenes, or is used for detecting the change of the overall motion.
In an alternative embodiment of the present invention,
the step of identifying the key frame in the standard image sequence, analyzing the motion trail of the target object in the key frame and the speed difference degree of the target object between the current frame and the last frame by an optical flow method, and determining the motion difference degree of the target object between the current frame and the last frame according to the motion trail and the speed difference degree comprises the following steps:
Applying a difference analysis algorithm to each frame in the standard image sequence, extracting image characteristics of each frame image in the standard image sequence, judging characteristic similarity between continuous frames according to the image characteristics, and taking a frame with the highest similarity with a later frame image as a key frame;
applying an optical flow method to the standard image sequence, extracting characteristic points of each frame in the standard image sequence, calculating an optical flow field of the target object between continuous frames, connecting initial characteristic points corresponding to the target object and characteristic points in key frames, and drawing a motion track of the target object;
calculating a displacement vector of each characteristic point corresponding to the target object in the optical flow field, determining the speed of the target object between frames by combining the time interval between frames in the standard image sequence, and comparing the speed corresponding to the key frame with the speed corresponding to the last frame to obtain a speed difference degree;
and calculating the motion difference degree of the target object between the current frame and the previous frame according to the speed difference degree and the motion trail.
The difference analysis algorithm is used to compare differences between images or image sequences to extract features or detect changes, the optical flow field being a field describing the motion of each pixel point in an image between two adjacent frames, the goal being to estimate for each pixel a motion vector representing the displacement of that pixel from one frame to the next.
Considering the characteristics of an image sequence and an application scene, selecting a proper difference analysis algorithm, such as color difference analysis, texture difference analysis, structure difference analysis and the like, applying the difference analysis algorithm to each frame in a standard image sequence, extracting the characteristics of each frame image, calculating the characteristic similarity between continuous frames through a similarity measurement method, judging the similarity between the continuous frames according to the calculated characteristic similarity, and taking the frame with the highest similarity with the image of the next frame as a key frame;
applying an optical flow method to each frame in a standard image sequence, extracting feature points, calculating an optical flow field of a target object between continuous frames, connecting the track of each feature point corresponding to the target object in the continuous frames by matching the feature points, including an initial frame and a key frame, drawing the motion track of the target object, and visualizing the motion condition of the target object in the image sequence by connecting paths among the feature points;
calculating displacement vectors of each characteristic point of the target object in the optical flow field, determining the speed of the target object between frames by combining the time intervals between frames in the standard image sequence, comparing the speed corresponding to the key frame with the speed corresponding to the last frame, and calculating the speed difference degree;
And calculating the motion difference degree of the target object between the current frame and the previous frame according to the speed difference degree and the motion trail.
In this embodiment, through a difference analysis algorithm, key frames with significant changes or features can be identified in the standard image sequence, which is helpful for capturing the motion state and key events of the target object, and the speed of the target object between consecutive frames is calculated by using the optical flow field. The speed difference degree is obtained by comparing the speed corresponding to the key frame with the speed corresponding to the previous frame, the speed change of the target object in the motion process can be captured, so that the difference of dynamic characteristics is reflected, an optical flow method is used for extracting the characteristic points of each frame in a standard image sequence, the track of the characteristic points is connected, the motion track of the target object can be formed to provide target motion information with finer granularity, the specific motion mode of the target object is captured, and in conclusion, the embodiment realizes comprehensive and detailed analysis on the motion of the target object by combining various computer vision technologies, and provides a powerful basis for subsequent application such as target tracking, behavior analysis and the like.
In an alternative embodiment of the present invention,
the calculating the motion difference degree of the target object between the current frame and the previous frame according to the speed difference degree and the motion trail comprises the following steps:
；
Wherein,Dthe degree of difference in motion is indicated,w 1 representing the weight of the velocity component in the horizontal direction,V x representing a velocity component in the horizontal direction,d x indicating the difference in position in the horizontal direction,αrepresenting the horizontal-direction scaling factor,w 2 representing the weight of the velocity component in the vertical direction,V y representing the velocity component in the vertical direction,d y a position difference in the vertical direction is indicated,βrepresenting the vertical scaling factor.
In the function, the speed components and the position differences in the horizontal direction and the vertical direction are integrated, the speed and the position information are regulated through weights, so that the calculation of the motion difference degree is more comprehensive, meanwhile, the influence of the speed and the position in different directions is considered, scaling and logarithmic transformation are carried out on the speed components to help regulate the contribution of the speed components to the motion difference degree, the weights of the speed components in the horizontal direction and the vertical direction can be flexibly controlled through regulating the weight coefficients, the possible difference of the importance of the motion in the horizontal direction and the importance of the speed in the possibility of the importance of the speed in the vertical direction in different application scenes are considered, the method is more flexible in different situations, in conclusion, the function comprehensively considers the speed and the position information through a reasonable mathematical model, and the comprehensive and flexible evaluation method for the motion change of a target object is provided through regulating the weights and scaling factors.
S3, according to the motion difference degree, matching the characteristic points corresponding to the target object in the key frame with other images in the standard image sequence through a preset adjacent image matcher to obtain matching point pairs, determining a transformation matrix corresponding to the matching point pairs through a fitting iterative algorithm, summarizing all transformation matrices, and determining an optimal transformation matrix to obtain the target object state estimation.
The close-by image matcher is a tool commonly used in the field of image processing and computer vision for finding similar feature points or image areas between different images, and the fitting iterative algorithm is typically used to estimate a mathematical model to best fit a set of data. In target object trajectory processing, fitting iterative algorithms are often used to estimate the motion transformation matrix of the target object between different frames.
In an alternative embodiment of the present invention,
according to the motion difference, matching the feature points corresponding to the target object in the key frame with the rest images in the standard image sequence through a preset adjacent image matcher to obtain matching point pairs, wherein the step of obtaining the matching point pairs comprises the following steps:
applying a feature point detection algorithm to the key frame, extracting a first feature point corresponding to a target object in the key frame, calculating a feature descriptor corresponding to the first feature point, and adding information corresponding to the motion difference degree to the feature descriptor corresponding to the first feature point to obtain a motion feature descriptor;
Initializing the adjacent image matcher according to the motion feature descriptors, correspondingly modifying indexes of the adjacent image matcher, dividing the feature points of other images except key frames in the standard image sequence into two parts based on division values according to the motion feature descriptors, wherein one part is the feature point with the division value, the other part is the feature point with the division value, repeating the division until the maximum depth is reached, obtaining a multi-dimensional tree subtree, and connecting the multi-dimensional tree subtrees to obtain a multi-dimensional tree;
for each first feature point, starting the first feature point from a root node of the multi-dimensional tree through the adjacent image matcher, comparing the value of the first feature point in a dividing dimension with the dividing value of a current node, determining a searching direction, and repeating searching until a leaf node of the multi-dimensional tree is reached;
and tracing back upwards from the leaf node, checking whether the feature points of the current node are closer to each other, updating nearest neighbor feature points and distances, and repeating the operation until all the first feature points are matched to obtain a matching point pair.
The feature descriptors are a way of numerically describing an image or a specific region in an image, are usually used to represent key points or regions in an image, and have the capability of uniquely and stably describing the content of an image, the motion feature descriptors are an extension of feature descriptors and contain data related to motion information, in the present invention, the motion feature descriptors contain information about the motion of a target object between successive frames, such as speed, direction, etc., by embedding the motion information in the feature descriptors, the dynamic change of the target object can be better captured, the multi-dimensional tree subtree is a branch or substructure of a multi-dimensional tree, the multi-dimensional tree is a data structure, usually a tree, for organizing and storing multi-dimensional data, is usually used to accelerate the searching and matching process, the root node is a top level node in the tree structure, it has no parent node, in the multi-dimensional tree, the root node is a starting point of the whole tree, the data space is divided into subspaces by the branch, the leaf node is a node in the tree structure has no child node, and the final data point is further divided in the multi-dimensional tree contains no data point.
Applying a feature point detection algorithm to a key frame, extracting a first feature point corresponding to a target object, forming a feature vector by extracting information of a local image area around the key point, calculating a feature descriptor of the first feature point, and adding motion difference information to the feature descriptor corresponding to the first feature point by embedding a numerical value or other representation forms of the motion difference into the feature vector to obtain a motion feature descriptor;
initializing a nearby image matcher by using a motion feature descriptor, modifying an index structure of the nearby image matcher to adapt to a new descriptor, dividing feature points of other images except key frames in a standard image sequence into two parts based on division values according to the motion feature descriptor, repeatedly dividing until the maximum depth is reached, forming a subtree of a multi-dimensional tree, connecting the subtrees of the multi-dimensional tree to form a complete multi-dimensional tree structure, wherein the construction and connection process of the multi-dimensional tree can be performed based on the similarity of the motion feature descriptor so as to ensure that feature points with similar motion features are closer in the tree;
searching the first feature point from the root node of the multi-dimensional tree by using an adjacent image matcher, comparing the value of the first feature point in the dividing dimension with the dividing value of the current node, determining the searching direction, and repeating the searching until the leaf node of the multi-dimensional tree is reached;
And (3) starting from the leaf node, tracing back the multidimensional tree upwards, checking whether the feature points of the current node are closer to each other, if so, updating the nearest neighbor feature points and the distances, and repeating the process until all the first feature points are matched to obtain the nearest neighbor feature points of each first feature point to form a matching point pair.
In the embodiment, the feature points of other images except the key frames in the standard image sequence are organized by using the multidimensional tree, so that the multidimensional tree subtrees are favorable for quickly positioning similar feature points during matching, the found nearest feature points are ensured to be globally optimal through a backtracking process, the matching point pairs are more accurate and reliable, the motion feature descriptors are introduced to enable the matching process to pay more attention to motion information, so that the method is better suitable for the change of the target object in the track, and in sum, the method is favorable for improving the accuracy and the robustness of the target object track processing method, and can better cope with the motion and the change of the target object in the track.
In an alternative embodiment of the present invention,
determining a transformation matrix corresponding to the matching point pair through a fitting iterative algorithm, summarizing all transformation matrices, determining an optimal transformation matrix, and obtaining a target object state estimation comprises the following steps:
Randomly selecting one of the matching point pairs as an initial matching point pair, defining the iteration times of the fitting iterative algorithm, randomly selecting three matching point pairs in each iteration to construct an affine transformation matrix, solving, calculating the predicted position of the initial matching point pair under the affine transformation matrix condition, calculating the distance between the actual position of the initial matching point pair and the predicted position, and dividing the initial matching point pair into inner points if the distance is smaller than a preset distance threshold;
for each matching point pair, determining whether the current matching point pair is an interior point, summarizing all the interior points, re-estimating an affine transformation matrix by using the interior points through a least square method to obtain an optimal transformation matrix, and transforming the key frame according to the optimal transformation matrix to obtain a target object state estimation.
The affine transformation matrix is a linear transformation matrix used for describing mathematical representations of linear transformations including translation, rotation, scaling, and shearing, and the optimal transformation matrix is generally an affine transformation matrix that is found by some optimization method to minimize the error between the transformed image and the target image, and the interior points are pairs of feature points considered as being correctly matched in the matching process, so as to exclude the exterior points (mismatching), thereby improving the accuracy of the estimation.
Randomly selecting one of the matching point pairs as an initial matching point pair, defining the iteration times of an iterative algorithm, randomly selecting three matching point pairs for each iteration, constructing an affine transformation matrix, solving, calculating the predicted position of the initial matching point pair under the condition of the obtained affine transformation matrix, calculating the distance between the actual position and the predicted position of the initial matching point pair according to the predicted position, comparing the calculated distance with a preset distance threshold, and dividing the initial matching point pair into inner points if the distance is smaller than the preset distance threshold;
for each matching point pair, determining whether the matching point pairs are interior points, summarizing all the interior points, re-estimating an affine transformation matrix by using the interior points through a least square method to obtain an optimal transformation matrix, and transforming the key frames according to the optimal transformation matrix to obtain transformed key frames, namely, state estimation of the target object.
In this embodiment, by randomly selecting matching point pairs to iterate, robustness of an algorithm is improved, sensitivity to initial selection is reduced, matching points which do not meet preset conditions can be removed through setting of a distance threshold and re-estimation of inner points, tolerance to noise and abnormal values is improved, affine transformation matrix is re-estimated by using a least square method, fitting effect to the inner points is further improved, a more accurate optimal transformation matrix is obtained, and in conclusion, fitting precision to the matching points is effectively improved, so that estimation of a target object track is more reliable and accurate.
Fig. 2 is a schematic structural diagram of a target object motion trajectory processing system based on video data according to an embodiment of the present invention, as shown in fig. 2, the system includes:
the first unit is used for acquiring a video file, decomposing the video file into an initial image sequence through a decoder, identifying noise in the initial image sequence, denoising, generating a frame image sequence, adjusting RGB curves of the frame image sequence through a linear tool, and correcting colors to obtain a standard image sequence;
the second unit is used for identifying key frames in the standard image sequence, analyzing the motion trail of a target object in the key frames and the speed difference degree of the target object between the current frame and the previous frame through an optical flow method, and determining the motion difference degree of the target object between the current frame and the previous frame according to the motion trail and the speed difference degree;
and the third unit is used for matching the characteristic points corresponding to the target object in the key frame with other images in the standard image sequence through a preset adjacent image matcher according to the motion difference degree to obtain matching point pairs, determining a transformation matrix corresponding to the matching point pairs through a fitting iterative algorithm, summarizing all transformation matrices and determining an optimal transformation matrix to obtain the target object state estimation.
In a third aspect of an embodiment of the present invention,
there is provided an electronic device including:
a processor;
a memory for storing processor-executable instructions;
wherein the processor is configured to invoke the instructions stored in the memory to perform the method described previously.
In a fourth aspect of an embodiment of the present invention,
there is provided a computer readable storage medium having stored thereon computer program instructions which, when executed by a processor, implement the method as described above.
The present invention may be a method, apparatus, system, and/or computer program product. The computer program product may include a computer readable storage medium having computer readable program instructions embodied thereon for performing various aspects of the present invention.
Finally, it should be noted that: the above embodiments are only for illustrating the technical solution of the present invention, and not for limiting the same; although the invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical scheme described in the foregoing embodiments can be modified or some or all of the technical features thereof can be replaced by equivalents; such modifications and substitutions do not depart from the spirit of the invention.
Claims (8)
1. The target object motion trail processing method based on the video data is characterized by comprising the following steps of:
the method comprises the steps of obtaining a video file, decomposing the video file into an initial image sequence through a decoder, identifying noise in the initial image sequence, denoising, generating a frame image sequence, adjusting RGB curves of the frame image sequence through a linear tool, and correcting colors to obtain a standard image sequence;
identifying a key frame in the standard image sequence, analyzing a motion track of a target object in the key frame and the speed difference degree of the target object between a current frame and a previous frame by an optical flow method, and determining the motion difference degree of the target object between the current frame and the previous frame according to the motion track and the speed difference degree;
according to the motion difference degree, matching characteristic points corresponding to the target object in the key frame with other images in the standard image sequence through a preset adjacent image matcher to obtain matching point pairs, determining a transformation matrix corresponding to the matching point pairs through a fitting iterative algorithm, summarizing all transformation matrices, and determining an optimal transformation matrix to obtain target object state estimation;
The step of identifying the key frame in the standard image sequence, analyzing the motion trail of the target object in the key frame and the speed difference degree of the target object between the current frame and the last frame by an optical flow method, and determining the motion difference degree of the target object between the current frame and the last frame according to the motion trail and the speed difference degree comprises the following steps:
applying a difference analysis algorithm to each frame in the standard image sequence, extracting image characteristics of each frame image in the standard image sequence, judging characteristic similarity between continuous frames according to the image characteristics, and taking a frame with the highest similarity with a later frame image as a key frame;
applying an optical flow method to the standard image sequence, extracting characteristic points of each frame in the standard image sequence, calculating an optical flow field of the target object between continuous frames, connecting initial characteristic points corresponding to the target object and characteristic points in key frames, and drawing a motion track of the target object;
calculating a displacement vector of each characteristic point corresponding to the target object in the optical flow field, determining the speed of the target object between frames by combining the time interval between frames in the standard image sequence, and comparing the speed corresponding to the key frame with the speed corresponding to the last frame to obtain a speed difference degree;
According to the speed difference and the motion trail, calculating to obtain the motion difference of the target object between the current frame and the previous frame;
the calculating the motion difference degree of the target object between the current frame and the previous frame according to the speed difference degree and the motion trail comprises the following steps:
；
wherein,Dthe degree of difference in motion is indicated,w 1 representing the weight of the velocity component in the horizontal direction,V x representing a velocity component in the horizontal direction,d x indicating the difference in position in the horizontal direction,αrepresenting the horizontal-direction scaling factor,w 2 representing the weight of the velocity component in the vertical direction,V y representing the velocity component in the vertical direction,d y a position difference in the vertical direction is indicated,βrepresenting the vertical scaling factor.
2. The method of claim 1, wherein the acquiring the video file, decomposing the video file into an initial image sequence by a decoder, identifying noise in the initial image sequence and denoising, generating a sequence of frame images comprises:
acquiring a video file to be detected, identifying the format of the video file, selecting a corresponding decoder, initializing the decoder, distributing memory for the decoder, and setting the resolution, the frame rate and the color space of the decoder;
Reading each frame of the video file, acquiring the type and the time stamp corresponding to each frame in the video file, inputting compressed frame data into a decoder to restore to an original image, performing inter-frame prediction according to the type corresponding to the original image, removing redundant information in the original image, and arranging the original image according to the time stamp to obtain an initial image sequence;
and carrying out pixel value analysis on each frame of original image in the initial image sequence, identifying the noise type of the original image through frequency domain analysis according to the result of the pixel value analysis, removing the corresponding noise, and removing the salt and pepper noise in the original image through visual inspection to obtain the frame image sequence.
3. The method of claim 1, wherein adjusting the RGB curves of the sequence of frame images by a linear tool and performing color correction to obtain a sequence of standard images comprises:
loading a frame image sequence, adding an adjustment layer in the layer, determining an RGB effect curve in the adjustment layer, adjusting the contrast of images in the frame image sequence by adjusting the shape of the RGB effect curve, modifying the brightness of the images in the frame image sequence by adjusting the top and bottom of the RGB effect curve, selecting a color channel in the RGB effect curve, adding a control point in the color channel, and adjusting the color intensity of the images in the frame image sequence by adjusting the position of the control point to obtain an adjustment image;
And evaluating the adjustment image through a histogram tool, determining whether the adjustment image meets the system requirement, if so, arranging the adjustment image according to the time stamp, outputting a standard image sequence, and if not, modifying the adjustment image according to the system requirement.
4. The method according to claim 1, wherein the matching, according to the motion difference degree, the feature point corresponding to the target object in the key frame with the rest of the images in the standard image sequence through a preset adjacent image matcher, to obtain a matching point pair includes:
applying a feature point detection algorithm to the key frame, extracting a first feature point corresponding to a target object in the key frame, calculating a feature descriptor corresponding to the first feature point, and adding information corresponding to the motion difference degree to the feature descriptor corresponding to the first feature point to obtain a motion feature descriptor;
initializing the adjacent image matcher according to the motion feature descriptors, correspondingly modifying indexes of the adjacent image matcher, dividing the feature points of other images except key frames in the standard image sequence into two parts based on division values according to the motion feature descriptors, wherein one part is the feature point with the division value, the other part is the feature point with the division value, repeating the division until the maximum depth is reached, obtaining a multi-dimensional tree subtree, and connecting the multi-dimensional tree subtrees to obtain a multi-dimensional tree;
For each first feature point, starting the first feature point from a root node of the multi-dimensional tree through the adjacent image matcher, comparing the value of the first feature point in a dividing dimension with the dividing value of a current node, determining a searching direction, and repeating searching until a leaf node of the multi-dimensional tree is reached;
and tracing back upwards from the leaf node, checking whether the feature points of the current node are closer to each other, updating nearest neighbor feature points and distances, and repeating the operation until all the first feature points are matched to obtain a matching point pair.
5. The method of claim 1, wherein determining the transformation matrix corresponding to the matching point pair by a fitting iterative algorithm, summing all the transformation matrices, and determining an optimal transformation matrix, and obtaining the target object state estimate comprises:
randomly selecting one of the matching point pairs as an initial matching point pair, defining the iteration times of the fitting iterative algorithm, randomly selecting three matching point pairs in each iteration to construct an affine transformation matrix, solving, calculating the predicted position of the initial matching point pair under the affine transformation matrix condition, calculating the distance between the actual position of the initial matching point pair and the predicted position, and dividing the initial matching point pair into inner points if the distance is smaller than a preset distance threshold;
And for each matching point pair, determining whether the current matching point pair is an inner point, summarizing all the inner points, and re-estimating an affine transformation matrix by using the inner points through a least square method to obtain an optimal transformation matrix, and transforming the key frame according to the optimal transformation matrix to obtain the target object state estimation.
6. A target object motion trajectory processing system based on video data for implementing the target object motion trajectory processing method based on video data according to any one of the preceding claims 1 to 5, characterized by comprising:
the first unit is used for acquiring a video file, decomposing the video file into an initial image sequence through a decoder, identifying noise in the initial image sequence, denoising, generating a frame image sequence, adjusting RGB curves of the frame image sequence through a linear tool, and correcting colors to obtain a standard image sequence;
the second unit is used for identifying key frames in the standard image sequence, analyzing the motion trail of a target object in the key frames and the speed difference degree of the target object between the current frame and the previous frame through an optical flow method, and determining the motion difference degree of the target object between the current frame and the previous frame according to the motion trail and the speed difference degree;
And the third unit is used for matching the characteristic points corresponding to the target object in the key frame with other images in the standard image sequence through a preset adjacent image matcher according to the motion difference degree to obtain matching point pairs, determining a transformation matrix corresponding to the matching point pairs through a fitting iterative algorithm, summarizing all transformation matrices and determining an optimal transformation matrix to obtain the target object state estimation.
7. An electronic device, comprising:
a processor;
a memory for storing processor-executable instructions;
wherein the processor is configured to invoke the instructions stored in the memory to perform the method of any of claims 1 to 5.
8. A computer readable storage medium having stored thereon computer program instructions, which when executed by a processor, implement the method of any of claims 1 to 5.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311744298.2A CN117474959B (en) | 2023-12-19 | 2023-12-19 | Target object motion trail processing method and system based on video data |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311744298.2A CN117474959B (en) | 2023-12-19 | 2023-12-19 | Target object motion trail processing method and system based on video data |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117474959A CN117474959A (en) | 2024-01-30 |
| CN117474959B true CN117474959B (en) | 2024-03-08 |
Family
ID=89627772
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311744298.2A Active CN117474959B (en) | 2023-12-19 | 2023-12-19 | Target object motion trail processing method and system based on video data |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117474959B (en) |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102156995A (en) * | 2011-04-21 | 2011-08-17 | 北京理工大学 | Video movement foreground dividing method in moving camera |
| CN105023278A (en) * | 2015-07-01 | 2015-11-04 | 中国矿业大学 | Movable target tracking method and system based on optical flow approach |
| CN105631870A (en) * | 2015-12-25 | 2016-06-01 | 北京理工大学 | Star image registration and target track extraction method based on NoC framework and device thereof |
| CN110968711A (en) * | 2019-10-24 | 2020-04-07 | 湖南大学 | Autonomous unmanned system position identification and positioning method based on sequence image characteristics |
| CN115439424A (en) * | 2022-08-23 | 2022-12-06 | 成都飞机工业（集团）有限责任公司 | Intelligent detection method for aerial video image of unmanned aerial vehicle |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US9277129B2 (en) * | 2013-06-07 | 2016-03-01 | Apple Inc. | Robust image feature based video stabilization and smoothing |
-
2023
- 2023-12-19 CN CN202311744298.2A patent/CN117474959B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102156995A (en) * | 2011-04-21 | 2011-08-17 | 北京理工大学 | Video movement foreground dividing method in moving camera |
| CN105023278A (en) * | 2015-07-01 | 2015-11-04 | 中国矿业大学 | Movable target tracking method and system based on optical flow approach |
| CN105631870A (en) * | 2015-12-25 | 2016-06-01 | 北京理工大学 | Star image registration and target track extraction method based on NoC framework and device thereof |
| CN110968711A (en) * | 2019-10-24 | 2020-04-07 | 湖南大学 | Autonomous unmanned system position identification and positioning method based on sequence image characteristics |
| CN115439424A (en) * | 2022-08-23 | 2022-12-06 | 成都飞机工业（集团）有限责任公司 | Intelligent detection method for aerial video image of unmanned aerial vehicle |
Non-Patent Citations (1)
| Title |
|---|
| 基于最小二乘法B样条曲线的视频去雾研究;刘佳佳;易淑铃;姚佳俊;林春;谢从华;;常熟理工学院学报;20170720(第04期);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117474959A (en) | 2024-01-30 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Liu et al. | Point cloud quality assessment: Dataset construction and learning-based no-reference metric | |
| CN111028213B (en) | Image defect detection method, device, electronic equipment and storage medium | |
| EP1612733B1 (en) | Color segmentation-based stereo 3D reconstruction system and process | |
| US5572258A (en) | Motion compensation estimating device and method achieving improved motion compensation | |
| CN108182421B (en) | Video segmentation method and device | |
| Zuo et al. | Explicit edge inconsistency evaluation model for color-guided depth map enhancement | |
| US20060056679A1 (en) | Full depth map acquisition | |
| EP0634873A2 (en) | Method to determine the motion vectors in small picture segments of a television picture | |
| US9247139B2 (en) | Method for video background subtraction using factorized matrix completion | |
| US20160117573A1 (en) | Method and apparatus for extracting feature correspondences from multiple images | |
| Kordelas et al. | Content-based guided image filtering, weighted semi-global optimization, and efficient disparity refinement for fast and accurate disparity estimation | |
| US11651581B2 (en) | System and method for correspondence map determination | |
| Wang et al. | Reference-free DIBR-synthesized video quality metric in spatial and temporal domains | |
| IT202000016054A1 (en) | METHOD FOR DETERMINING THE CONFIDENCE OF A DISPARITY MAP BY SELF-ADAPTIVE LEARNING OF A NEURAL NETWORK, AND RELATED SENSOR SYSTEM | |
| US20110085026A1 (en) | Detection method and detection system of moving object | |
| CN112037109A (en) | Improved image watermarking method and system based on saliency target detection | |
| CN114419102A (en) | Multi-target tracking detection method based on frame difference time sequence motion information | |
| CN117474959B (en) | Target object motion trail processing method and system based on video data | |
| CN116980549A (en) | Video frame processing method, device, computer equipment and storage medium | |
| Kim et al. | Cnn-based blind quality prediction on stereoscopic images via patch to image feature pooling | |
| CN114519832A (en) | Affine inverse transformation model-based video global motion compensation method | |
| Mouffranc et al. | Colorimetric correction for stereoscopic camera arrays | |
| CN113129332A (en) | Method and apparatus for performing target object tracking | |
| CN117336620B (en) | Adaptive video stitching method and system based on deep learning | |
| Prabavathy et al. | Gradual transition detection in shot boundary using gradual curve point |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |