Motion trajectory information detection method and system based on intelligent AI decoding algorithm

Abstract

The present invention discloses a motion trajectory information detection method based on an intelligent AI decoding algorithm, including data synchronization processing, motion feature analysis, intermediate frame position prediction and interpolation calculation; according to the average speed of the object and the image resolution, a bilinear or bicubic interpolation algorithm is selected to generate an intermediate frame to fill in the missing information; further pixel-level comparison is performed to identify and classify the information missing area, and pixel values ​​are obtained and filled in the continuous missing area to achieve accurate detection and prediction of motion information. This method is applicable to a variety of scenarios that require precise motion analysis, and improves the accuracy and efficiency of detection.

Description

Technical Field

[0001] The present invention relates to the field of image analysis and processing technology, and more specifically, to a motion trajectory information detection method, system, electronic device and storage medium based on an intelligent AI decoding algorithm. Background Art

[0002] In the field of intelligent monitoring and motion analysis, existing technologies mainly rely on high-speed cameras to capture the motion trajectory of target objects. These cameras are usually equipped with advanced image processing algorithms to analyze and record the motion state of objects in real time. The technical principles mainly include image acquisition, data synchronization, motion feature extraction, and further analysis and processing based on these data.

[0003] Existing systems have limitations in many ways. First, they often have difficulty maintaining high-precision time synchronization when dealing with high-speed moving objects, which may lead to inconsistent timestamps during data acquisition. Second, existing interpolation algorithms may not accurately reflect the actual motion state of objects when filling in missing information in image data, especially when the object moves quickly or the image resolution is low. In addition, existing technologies often lack a deep understanding and prediction of the object's motion pattern when filling in areas where information is missing, which limits the system's adaptability and accuracy to complex motion scenes.

[0004] In the process of implementing the embodiments of the present invention, the inventors found that there are at least the following problems or defects in the prior art: Although the existing motion trajectory detection methods can track and analyze the motion of the target object to a certain extent, they often cannot provide sufficiently accurate and reliable data when dealing with complex motion in a high-dynamic environment. In particular, in terms of data fusion and time synchronization of multi-angle and multi-camera devices, the prior art has not yet been able to fully meet the needs of high-precision motion trajectory detection. In addition, the prior art lacks sufficient intelligence and adaptability when predicting motion information and filling in information missing areas, which to a certain extent affects the accuracy and reliability of motion trajectory detection. Summary of the invention

[0005] The present invention provides a motion trajectory information detection method, system, medium and computing device based on an intelligent AI decoding algorithm.

[0006] In a first aspect of the present invention, a motion trajectory information detection method based on an intelligent AI decoding algorithm is provided, comprising:

[0007] Using at least three high-speed camera devices at different positions to collect multi-angle image data of the target object, each high-speed camera device is equipped with a time synchronization device, a positioning device and an atomic clock;

[0008] Sending a unified trigger signal to each high-speed camera device according to a preset time accuracy through a time synchronization device, wherein the trigger signal includes a start pulse and a stop pulse, which are used to mark the start and end of data acquisition respectively;

[0009] Collecting image data collected by each high-speed camera device and its corresponding timestamp, and using the timestamp as a reference to perform data synchronization processing, including sorting and aligning the image data based on the timestamp;

[0010] Acquire image data with inconsistent timestamps, select at least two frames of image data with adjacent timestamps from the image data with inconsistent timestamps, determine the time interval Δt and motion features between the image data with adjacent timestamps, the motion features including the speed of the target object and acceleration

[0011] According to the time interval Δt and acceleration The position of the target object in the middle frame is calculated by formula (1): Based on the position of the target object in the intermediate frame Get the target pixel point Ptarget;

[0012]

[0013] In the formula, is the position of the target object in the middle frame; is the position of the target object in the first frame; the average speed of the target object is calculated by formula (2):

[0014]

[0015] In the formula, v i is the speed of the target object in the i-th frame, and n is the number of acquisition frames;

[0016] Determining a resolution R of the image data, the resolution including the number of pixels or the physical size of the image;

[0017] Set the speed threshold Vthreshold and the resolution threshold Rthreshold;

[0018] like Or R<Rthreshold, interpolation calculation is performed based on the bilinear interpolation algorithm to generate an intermediate frame;

[0019] like If R≥Rthreshold, the interpolation calculation is performed based on the bicubic interpolation algorithm to generate an intermediate frame;

[0020] Based on the generated intermediate frames, the information missing in the time series is filled and the motion information is detected.

[0021] Furthermore, the interpolation calculation based on the bilinear interpolation algorithm and the generation of the intermediate frame include:

[0022] The value of each pixel in the intermediate frame is determined by the weighted average of the four nearest neighboring pixels P1, P2, P3, and P4 to the target pixel Ptarget. The weight of the pixel is determined by the position of the target object in the intermediate frame. The distance to the nearest neighbor pixel is determined by the weight W i It is inversely proportional to the distance, as shown in formula (3);

[0023]

[0024] Where P i is the pixel; W i is the pixel point P i The weight of .

[0025] Furthermore, performing interpolation calculation on image data with inconsistent timestamps based on a bicubic interpolation algorithm and generating an intermediate frame comprises:

[0026] Determine the 16 pixels around the target pixel Ptarget to form a 4x4 grid;

[0027] Use a third-order polynomial fit to determine the coefficient c corresponding to each pixel in the 4x4 grid ij , where i and j range from 0 to 3;

[0028] Construct a bicubic polynomial function f(x, y) as shown in formula (4) to calculate the gray value of the target pixel point Ptarget;

[0029]

[0030] Where x and y are coordinate variables in the image space.

[0031] Furthermore, the method further comprises:

[0032] For each pair of adjacent known frames and generated intermediate frames, pixel-level comparison is performed to identify the difference areas;

[0033] Through comparative analysis, the parts in the difference area that do not belong to background changes or normal motion features are marked as information missing areas.

[0034] Furthermore, the method further comprises:

[0035] For the identified information missing regions, the information missing types are classified according to their distribution patterns and durations in consecutive frames, and the information missing types include random missing, periodic missing, and specific event missing.

[0036] Furthermore, the method further comprises:

[0037] For persistent information missing areas, the pixel values ​​of the missing areas in the intermediate frames are obtained by combining the pixel changes of adjacent frames and the motion characteristics of the target object.

[0038] Based on the pixel values ​​of the missing areas in the intermediate frames, pixel data for filling the missing areas is generated to ensure the consistency of the filled content in time series and space;

[0039] Based on the filled intermediate frames, motion information detection is performed, including but not limited to target object recognition, positioning, tracking and motion trajectory construction;

[0040] By analyzing the motion pattern and speed parameters of the target object in the filled frame, the motion state of the target object can be predicted.

[0041] Furthermore, the method further comprises:

[0042] Based on the position of the target object in the intermediate frame As the search center, a local search is performed in the middle frame to determine the target pixel point Ptarget.

[0043] In a second aspect of the present invention, a motion trajectory information detection system based on an intelligent AI decoding algorithm is provided, comprising:

[0044] At least three high-speed camera devices, each of which is equipped with a time synchronization device, a positioning device and an atomic clock, for collecting multi-angle image data of the target object;

[0045] A time synchronization controller, used to send a unified trigger signal to each high-speed camera device according to a preset time accuracy;

[0046] A data collection unit, used to collect image data collected by each camera device and its corresponding time stamp, and perform data synchronization processing;

[0047] Motion feature analyzer, used to determine the time interval Δt and the motion features of the target object, including speed, from the image data with adjacent time stamps and acceleration

[0048] Position prediction unit, used to predict the position of the vehicle based on the time interval Δt and acceleration Calculate the position of the target object in the intermediate frame

[0049] Interpolation calculation unit, used to calculate the average speed of the target object The resolution R of the image data is interpolated and an intermediate frame is generated;

[0050] An information filling unit, used to fill in missing information in the time series based on the generated intermediate frames;

[0051] Motion information detector, used to detect motion information of the filled intermediate frames, including target object recognition, positioning, tracking and motion trajectory construction;

[0052] The data processing unit is used to analyze the motion pattern and speed parameters of the target object in the filled frame, predict the motion state of the target object, and provide corresponding data output.

[0053] In a third aspect of the present invention, an electronic device is provided, comprising: at least one processor, a memory and an input-output unit; wherein the memory is used to store a computer program, and the processor is used to call the computer program stored in the memory to execute any one of the methods described in the first aspect.

[0054] In a fourth aspect of the present invention, a computer-readable storage medium is provided, which includes instructions, and when the instructions are executed on a computer, the computer executes any one of the methods in the first aspect.

[0055] According to the above-mentioned embodiments of the present invention, at least the following beneficial effects are achieved: the motion trajectory information detection method based on the intelligent AI decoding algorithm provided by the present invention collects multi-angle image data by using at least three high-speed camera devices, and combines time synchronization equipment, positioning equipment and atomic clocks to ensure high precision and synchronization of data acquisition. The method effectively solves the problem of inconsistent timestamps in the prior art through time synchronization processing, and improves the accuracy and reliability of data. The bilinear and bicubic interpolation algorithms are also used to intelligently select the interpolation method according to the motion speed and image resolution of the target object, optimize the generation process of the intermediate frame, and enhance the ability to fill the missing information, thereby ensuring the continuity and integrity of the motion information while maintaining high resolution. In addition, the method of the present invention also includes intelligent identification and classification of information missing areas, and pixel value recovery based on the motion characteristics of the target object, which further improves the accuracy of the filled content and the reliability of the prediction. Through these technical means, the present invention not only improves the accuracy of motion trajectory detection, but also enhances the adaptability and intelligence level of the system, so that it can better cope with complex and changeable motion scenes and provide users with more accurate and comprehensive motion analysis results. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] The above and other objects, features and advantages of the exemplary embodiments of the present invention will become readily understood by reading the detailed description below with reference to the accompanying drawings. In the accompanying drawings, several embodiments of the present invention are shown in an exemplary and non-limiting manner, in which:

[0057] Figure 1 A schematic diagram of a flow chart of a motion trajectory information detection method based on an intelligent AI decoding algorithm provided by an embodiment of the present invention;

[0058] Figure 2 A schematic diagram of the structure of a motion trajectory information detection system based on an intelligent AI decoding algorithm provided by an embodiment of the present invention;

[0059] Figure 3 The schematic diagram schematically shows the structure of an electronic device according to an embodiment of the present invention. DETAILED DESCRIPTION

[0060] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0061] According to the first aspect of the present invention, see Figure 1 , a motion trajectory information detection method based on intelligent AI decoding algorithm, comprising:

[0062] Using at least three high-speed camera devices at different positions to collect multi-angle image data of the target object, each high-speed camera device is equipped with a time synchronization device, a positioning device and an atomic clock;

[0063] Sending a unified trigger signal to each high-speed camera device according to a preset time accuracy through a time synchronization device, wherein the trigger signal includes a start pulse and a stop pulse, which are used to mark the start and end of data acquisition respectively;

[0064] Collecting image data collected by each high-speed camera device and its corresponding timestamp, and using the timestamp as a reference to perform data synchronization processing, including sorting and aligning the image data based on the timestamp;

[0065] Acquire image data with inconsistent timestamps, select at least two frames of image data with adjacent timestamps from the image data with inconsistent timestamps, determine the time interval Δt and motion features between the image data with adjacent timestamps, the motion features including the speed of the target object and acceleration

[0066] According to the time interval Δt and acceleration The position of the target object in the middle frame is calculated by formula (1): Based on the position of the target object in the intermediate frame Get the target pixel point Ptarget;

[0067]

[0068] In the formula, is the position of the target object in the middle frame; is the position of the target object in the first frame; the average speed of the target object is calculated by formula (2):

[0069]

[0070] In the formula, v i is the speed of the target object in the i-th frame, and n is the number of acquisition frames;

[0071] Determining a resolution R of the image data, the resolution including the number of pixels or the physical size of the image;

[0072] Set the speed threshold Vthreshold and the resolution threshold Rthreshold;

[0073] like Or R<Rthreshold, interpolation calculation is performed based on the bilinear interpolation algorithm to generate an intermediate frame;

[0074] like If R≥Rthreshold, the interpolation calculation is performed based on the bicubic interpolation algorithm to generate an intermediate frame;

[0075] Based on the generated intermediate frames, the information missing in the time series is filled and the motion information is detected.

[0076] It should be noted that the timestamp of the present invention refers to an accurate time mark associated with each image data, which is used to ensure the acquisition order and synchronization of the image data. Image data with adjacent timestamps refers to image data pairs that are adjacent in sequence in the time series. For example, if the data acquisition frequency is 30 frames per second, the time difference between any two frames of data is about 1 / 30 second. The position of the intermediate frame Refers to the position of the target object at a certain moment between two frames of images predicted by the physical motion model. The target pixel point Ptarget refers to the specific pixel or pixel set of the target object identified in the image, which is usually used as a reference point for analysis and tracking. The position of the target object in the first frame Refers to the position of the target object determined in the initial image frame, which serves as the starting point for motion analysis.

[0077] In some embodiments, multiple high-speed cameras can be deployed to capture multi-angle images of the target object. These cameras are installed at strategic locations that can fully cover the target object's motion area to ensure that high-quality image data is obtained from different perspectives. Image data refers to the digital image sequence captured by the high-speed camera, which is the basis for subsequent motion analysis and trajectory detection.

[0078] It should be noted that in some embodiments, the target object refers to the object being tracked and analyzed in the multi-angle image data. In practical applications, this can be an athlete in a sports event, a mechanical part in industrial automation, or a vehicle in traffic monitoring.

[0079] Specifically, the selection of the camera device takes into account its image resolution, frame rate and dynamic range to adapt to different monitoring environments and the possible movement speed of the target object. Image resolution refers to the pixel density of the image, usually expressed in pixels per inch (PPI) or number of pixels (such as 1920x1080). In practical applications, resolution determines the level of detail of the image and the accuracy of the analysis.

[0080] For example, a camera may use a 24-megapixel sensor that can capture images at 120 frames per second, ensuring clear image data even in high-speed motion. At the same time, each camera integrates a high-precision time synchronization module that uses an atomic clock to synchronize time with other cameras with a synchronization accuracy of microseconds.

[0081] Preferably, the camera also has automatic focus and exposure adjustment functions to adapt to different lighting conditions and distance changes of the target object. In addition, the field of view and focal length of the camera can be adjusted according to the size and range of motion of the target object to optimize the quality and coverage of the image.

[0082] In some embodiments, a centralized time synchronization controller sends a unified trigger signal to all camera devices to achieve accurate time synchronization. The trigger signal includes a start pulse and a stop pulse to ensure that all camera devices start and end image data acquisition at the same time.

[0083] Specifically, the time synchronization controller can be an independent device or integrated into one of the camera devices. It communicates with all camera devices by wired or wireless means and sends precise synchronization signals. The sending of trigger signals can be periodic to meet the needs of continuous monitoring, or based on specific event triggering to achieve the capture of specific motion events.

[0084] Preferably, the time synchronization controller can also be dynamically adjusted according to the feedback of the camera device to compensate for possible transmission delays or clock drift. In addition, a redundant synchronization mechanism can be adopted, such as using multiple time synchronization controllers, to improve the reliability and fault tolerance of synchronization. The system can also dynamically adjust the timing of sending the trigger signal according to the motion state of the target object to achieve more flexible and efficient data collection. For example, when the target object enters a key area or starts a specific action, the high-speed camera device can be automatically triggered to collect data.

[0085] In some embodiments, the captured image data is analyzed for motion features to determine the speed and acceleration of the target object. These motion features are essential for predicting the position of the target object in subsequent frames.

[0086] Specifically, motion feature analysis can be achieved by calculating the image differences between consecutive frames. For example, by calculating the position change of the target object in consecutive frames, its velocity vector can be obtained. Further, by analyzing the change of the velocity vector, acceleration information can be obtained.

[0087] Preferably, image processing techniques, such as background subtraction or feature point tracking, can also be used to improve the accuracy of motion feature analysis. In addition, appropriate algorithm parameters, such as search window size, threshold setting, etc., can be selected according to the motion characteristics of the target object and environmental conditions to optimize the extraction effect of motion features.

[0088] In some embodiments, according to the average speed of the target object and the resolution of the image data, an appropriate interpolation algorithm is selected by setting a speed threshold Vthreshold and a resolution threshold Rthreshold to generate an intermediate frame. This helps to fill in the missing information in the time series and improve the continuity and accuracy of motion information detection. When the average speed of the target object is low or the image resolution is high, a bilinear interpolation algorithm can be selected because it is computationally simpler and faster. On the contrary, when the average speed of the target object is high or the image resolution is low, a bicubic interpolation algorithm can be selected to obtain a smoother and higher-quality image.

[0089] It should be noted that determining the resolution of image data is a key step in achieving accurate motion trajectory detection. The resolution R can refer to the number of pixels in the image, such as 1920×1080 pixels, or the physical size of the image, such as 3000×2000 pixels. The resolution can be determined based on the hardware specifications of the camera device and the degree of detail required for the target object. For example, high-resolution images can provide more details and are suitable for application scenarios that require fine analysis, such as medical imaging or satellite image analysis. In practical applications, the resolution setting can also take into account storage and processing efficiency. In some real-time processing systems, it may be necessary to balance image quality and processing speed and select a moderate resolution to ensure that sufficient image details can be obtained while maintaining the real-time responsiveness of the system.

[0090] Specifically, the speed threshold Vthreshold can be set based on the average speed of the target object. For example, if the average speed of the target object exceeds 5 meters per second, a higher speed threshold may need to be set, such as 10 meters per second, to ensure that the interpolation algorithm can adapt to fast motion. The setting of the resolution threshold Rthreshold depends on the required level of image detail. For example, for application scenarios that require high definition, a higher resolution threshold can be set, such as 2 million pixels (Megapixel) or more. These thresholds can be dynamically adjusted according to the specific needs of actual applications. For example, in some monitoring scenarios, the threshold can be automatically adjusted according to the motion state of the target object or environmental changes to optimize the interpolation effect. Specifically, if Or R<Rthreshold, then interpolation calculation is performed based on the bilinear interpolation algorithm and an intermediate frame is generated; if And R ≥ Rthreshold, then interpolation calculation is performed based on the bicubic interpolation algorithm and an intermediate frame is generated; if the average speed of the target object If the speed is lower than the speed threshold Vthreshold, or the image resolution R is lower than the resolution threshold Rthreshold, the bilinear interpolation algorithm will be selected. This algorithm is suitable for situations with slow speed or low resolution, and can provide faster calculation speed and reasonable interpolation effect. On the contrary, if the average speed of the target object is If the speed threshold Vthreshold is greater than or equal to the image resolution R, the system will select the bicubic interpolation algorithm, which is suitable for faster speeds or higher resolutions and can provide smoother images and higher visual quality.

[0091] Preferably, the parameters of the interpolation algorithm can be dynamically adjusted in combination with the motion characteristics of the target object and the complexity of the image content, such as the size and shape of the interpolation kernel. In addition, the system can also use multi-scale interpolation technology to achieve more refined image reconstruction at different resolution levels.

[0092] In some embodiments, the information missing in the time sequence is filled based on the generated intermediate frames, and motion information detection is performed on the filled image sequence, which helps to achieve complete and accurate tracking of the motion trajectory of the target object.

[0093] Specifically, information filling can be achieved by inserting the calculated intermediate frames into the missing regions in the time series. These intermediate frames not only fill the information gaps, but also maintain the coherence and smoothness of the target object's motion.

[0094] In some embodiments, performing interpolation calculation based on a bilinear interpolation algorithm and generating an intermediate frame includes:

[0095] The value of each pixel in the intermediate frame is determined by the weighted average of the four nearest neighboring pixels P1, P2, P3, and P4 to the target pixel Ptarget. The weight of the pixel is determined by the position of the target object in the intermediate frame. The distance to the nearest neighbor pixel is determined by the weight W i It is inversely proportional to the distance, as shown in formula (3);

[0096]

[0097] Where P i is the pixel; W i is the pixel point P i The weight of .

[0098] In some embodiments, a bilinear interpolation algorithm is used to generate an intermediate frame, and the value of a pixel in the intermediate frame is determined by calculating a weighted average of four nearest neighboring pixel points P1, P2, P3, and P4 of a target pixel point Ptarget.

[0099] The pixel weight W i The position of the target object in the middle frame The distance to the nearest neighbor pixel is determined by calculating the distance between Ptarget and each nearest neighbor pixel P i The distance can be calculated using the Euclidean distance to ensure that the weight reflects the proximity of the pixel to Ptarget. In addition, in order to avoid the weight being affected by extreme values, the weight can be normalized so that the sum of the weights of all four pixels is 1.

[0100] In some embodiments, performing interpolation calculation on image data with inconsistent timestamps based on a bicubic interpolation algorithm and generating an intermediate frame comprises:

[0101] Determine the 16 pixels around the target pixel Ptarget to form a 4x4 grid;

[0102] Use a third-order polynomial fit to determine the coefficient c corresponding to each pixel in the 4x4 grid ij , where i and j range from 0 to 3;

[0103] Construct a bicubic polynomial function f(x, y) as shown in formula (4) to calculate the gray value of the target pixel point Ptarget;

[0104]

[0105] Where x and y are coordinate variables in the image space.

[0106] The bicubic interpolation algorithm can be used to interpolate image data with inconsistent timestamps to generate intermediate frames. First, determine the 16 pixels around the target pixel Ptarget to form a 4x4 grid. Then, use a third-order polynomial fit to determine the coefficient c corresponding to each pixel in the grid. ij , where the values ​​of i and j range from 0 to 3.

[0107] The fitting of the third-order polynomial can be achieved by the least squares method to ensure that the value of the polynomial at the grid pixel point is as close as possible to the actual pixel value. The constructed bicubic polynomial function f(x,y) is used to calculate the grayscale value of the target pixel point Ptarget, where x and y are the coordinate variables in the image space.

[0108] More specifically, the bicubic polynomial provides a smooth transition of grayscale changes around the target pixel, so that the visual quality and edge smoothness of the image can be maintained when generating intermediate frames. In practical applications, this interpolation method is suitable for occasions that require high-quality image processing, such as image enlargement or rotation.

[0109] It should be noted that the selection of the above two interpolation algorithms and the setting of parameters are aimed at improving the quality of the intermediate frame generation and ensuring the accuracy and reliability of motion information detection. Through precise weight calculation and polynomial fitting, these implementations can effectively fill in the information missing in the time series while maintaining image quality.

[0110] In some embodiments, the method further comprises:

[0111] For each pair of adjacent known frames and generated intermediate frames, pixel-level comparison is performed to identify the difference areas;

[0112] Through comparative analysis, the parts in the difference area that do not belong to background changes or normal motion features are marked as information missing areas.

[0113] In some embodiments, the pixel-level comparison is performed to identify differences between the known frame and the generated intermediate frame. This involves analyzing the differences between the two frames at each pixel to find areas of change that may be caused by motion or other factors.

[0114] Specifically, pixel-level comparison can be performed by calculating the difference between corresponding pixels in two image frames. For example, if the pixel value of the known frame is A and the pixel value of the intermediate frame is B, the comparison operation can be expressed as calculating |AB|. The difference area can be determined by setting a threshold. When the difference exceeds this threshold, the pixel is considered to belong to the difference area.

[0115] In order to improve the accuracy of the comparison, a local threshold method can also be used, which dynamically adjusts the threshold according to the local brightness or color changes of the image. In addition, the background model of the image can be combined to distinguish between background changes and the movement of the target object, so as to more accurately identify the information missing area.

[0116] In some embodiments, the method further comprises:

[0117] For the identified information missing regions, the information missing types are classified according to their distribution patterns and durations in consecutive frames, and the information missing types include random missing, periodic missing, and specific event missing.

[0118] The classification of information loss types is based on the performance of the information loss area in consecutive frames. This helps to understand the cause of the information loss and take appropriate measures to deal with it. Specifically, the classification process can be performed by analyzing the characteristics of the information loss area in the time series. For example, random loss may appear and disappear randomly in different frames, while periodic loss may recur at fixed time intervals.

[0119] More specifically, pattern recognition techniques can be used to automate the classification process, such as by training a support vector machine (SVM) classifier to distinguish different types of information loss. In addition, multiple thresholds and feature parameters can be set to improve the accuracy and robustness of the classification.

[0120] In some embodiments, the method further comprises:

[0121] For persistent information missing areas, the pixel values ​​of the missing areas in the intermediate frames are obtained by combining the pixel changes of adjacent frames and the motion characteristics of the target object.

[0122] Based on the pixel values ​​of the missing areas in the intermediate frames, pixel data for filling the missing areas is generated to ensure the consistency of the filled content in time series and space;

[0123] Based on the filled intermediate frames, motion information detection is performed, including but not limited to target object recognition, positioning, tracking and motion trajectory construction;

[0124] By analyzing the motion pattern and speed parameters of the target object in the filled frame, the motion state of the target object can be predicted.

[0125] In some embodiments, for areas with persistent information loss, the system combines pixel changes in adjacent frames and the motion characteristics of the target object to obtain pixel values ​​of the missing area in the intermediate frame. This helps fill in the missing area and ensure the integrity of the image sequence.

[0126] The position of the target object in the missing area can be predicted by analyzing its motion trajectory in the known frames and interpolating the values ​​of the surrounding pixels. For example, if the target object shows uniform linear motion, the pixel values ​​of the missing area can be estimated by linear interpolation.

[0127] More advanced interpolation techniques, such as deep learning-based generative models, can be used to generate more natural and accurate filled pixels. In addition, the filling process can combine the texture and color information of the image to ensure that the filling result is consistent with the surrounding environment.

[0128] In some embodiments, the method further comprises:

[0129] Based on the position of the target object in the intermediate frame As the search center, a local search is performed in the middle frame to determine the target pixel point Ptarget.

[0130] In some embodiments, based on the position of the target object in the intermediate frame As the search center, the system performs a local search in the intermediate frame to determine the target pixel Ptarget. This helps to accurately locate the target object even when some information is missing. The local search can use a cross or spiral search pattern to start from the predicted position. Start by checking the surrounding pixels step by step until you find a pixel that matches the target object's features.

[0131] The search algorithm can incorporate the expected shape and size information of the target object to narrow the search range and improve the search efficiency. In addition, the termination condition of the search, such as the maximum search radius or the number of search steps, can be set to avoid infinite search and improve the practicality of the algorithm.

[0132] The above-mentioned embodiments of the present invention have the following beneficial effects: First, multi-angle image data of the target object is collected by at least three high-speed camera devices, and the time synchronization device, positioning device and atomic clock are used to achieve accurate synchronization of data, which greatly improves the accuracy and reliability of data collection. Secondly, the method sorts and aligns the image data by timestamps, effectively solves the problem of inconsistent timestamps, and ensures the consistency of the data. In addition, by calculating the position of the target object in the intermediate frame, combining the bilinear and bicubic interpolation algorithms, the interpolation method suitable for the current motion speed and image resolution is intelligently selected, the generation of the intermediate frame is optimized, and the accuracy of filling the missing information is improved. Finally, the method also performs motion information detection, including the identification, positioning, tracking and construction of motion trajectories of the target object, by analyzing the filled intermediate frames, thereby achieving accurate prediction of the motion state of the target object and providing users with more comprehensive and in-depth motion analysis results.

[0133] like Figure 2 As shown, in some embodiments, a motion trajectory information detection system 200 based on an intelligent AI decoding algorithm includes:

[0134] At least three high-speed camera devices 201, each camera device is equipped with a time synchronization device, a positioning device and an atomic clock, and is used to collect multi-angle image data of the target object;

[0135] A time synchronization controller 202, used to send a unified trigger signal to each high-speed camera device according to a preset time accuracy;

[0136] A data collection unit 203, used to collect image data collected by each camera device and its corresponding timestamp, and perform data synchronization processing;

[0137] The motion feature analyzer 204 is used to determine the time interval Δt and the motion features of the target object from the image data with adjacent time stamps, including the speed and acceleration

[0138] The position prediction unit 205 is used to predict the position of the vehicle according to the time interval Δt and the acceleration. Calculate the position of the target object in the intermediate frame

[0139] The interpolation calculation unit 206 is used to calculate the average speed of the target object. The resolution R of the image data is interpolated and an intermediate frame is generated;

[0140] An information filling unit 207, used to fill in missing information in the time series based on the generated intermediate frames;

[0141] The motion information detector 208 is used to perform motion information detection on the filled intermediate frames, including target object identification, positioning, tracking and motion trajectory construction;

[0142] The data processing unit 209 is used to analyze the motion mode and speed parameters of the target object in the filled frame, predict the motion state of the target object, and provide corresponding data output.

[0143] It can be understood that the modules recorded in the motion trajectory information detection system 200 based on the intelligent AI decoding algorithm are similar to those in the reference Figure 1 The steps in the motion trajectory information detection method based on the intelligent AI decoding algorithm described above correspond to each other. Therefore, the operations, features and beneficial effects described above for the motion trajectory information detection method based on the intelligent AI decoding algorithm are also applicable to the motion trajectory information detection system 200 based on the intelligent AI decoding algorithm and the modules contained therein, and will not be repeated here.

[0144] Reference below Figure 3 , which shows a schematic diagram of the structure of an electronic device 300 suitable for implementing some embodiments of the present invention. The electronic devices in some embodiments of the present invention may include but are not limited to mobile terminals such as mobile phones, notebook computers, digital broadcast receivers, PDAs (personal digital assistants), PADs (tablet computers), PMPs (portable multimedia players), vehicle terminals (such as vehicle navigation terminals), etc., and fixed terminals such as digital TVs, desktop computers, etc. Figure 3 The terminal device shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present invention.

[0145] like Figure 3 As shown, the electronic device 300 may include a processing device (e.g., a central processing unit, a graphics processing unit, etc.) 301, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 302 or a program loaded from a storage device 308 into a random access memory (RAM) 303. In the RAM 303, various programs and data required for the operation of the electronic device 300 are also stored. The processing device 301, the ROM 302, and the RAM 303 are connected to each other via a bus 304. An input / output (I / O) interface 305 is also connected to the bus 304.

[0146] Typically, the following devices may be connected to the I / O interface 305: input devices 306 including, for example, a touch screen, a touch pad, a keyboard, a mouse, a camera, a microphone, an accelerometer, a gyroscope, etc.; output devices 307 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, etc.; storage devices 308 including, for example, a magnetic tape, a hard disk, etc.; and communication devices 309. The communication devices 309 may allow the electronic device 300 to communicate with other devices wirelessly or by wire to exchange data. Although Figure 3 The electronic device 300 is shown with various devices, but it should be understood that it is not required to implement or possess all the devices shown. More or fewer devices may be implemented or possessed instead. Figure 3 Each block shown in the figure may represent one device, or may represent multiple devices as required.

[0147] Furthermore, the storage medium of the embodiment of the present application stores program instructions that can implement all the above methods, wherein the program instructions can be stored in the above storage medium in the form of a software product, including several instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) or a processor to execute all or part of the steps of the methods described in each embodiment of the present application. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, or terminal devices such as a computer, a server, a mobile phone, and a tablet.

[0148] The above descriptions are only some preferred embodiments of the present invention and an explanation of the technical principles used. Those skilled in the art should understand that the scope of the invention involved in the embodiments of the present invention is not limited to the technical solutions formed by a specific combination of the above technical features, but should also cover other technical solutions formed by any combination of the above technical features or their equivalent features without departing from the above inventive concept. For example, the above features are replaced with (but not limited to) technical features with similar functions disclosed in the embodiments of the present invention to form a technical solution.

Claims

1. A motion trajectory information detection method based on intelligent AI decoding algorithm, characterized in that: The method comprises: Using at least three high-speed camera devices at different positions to collect multi-angle image data of the target object, each high-speed camera device is equipped with a time synchronization device, a positioning device and an atomic clock; Sending a unified trigger signal to each high-speed camera device according to a preset time accuracy through a time synchronization device, wherein the trigger signal includes a start pulse and a stop pulse, which are used to mark the start and end of data acquisition respectively; Collecting image data collected by each high-speed camera device and its corresponding timestamp, and using the timestamp as a reference to perform data synchronization processing, including sorting and aligning the image data based on the timestamp; Acquire image data with inconsistent timestamps, select at least two frames of image data with adjacent timestamps from the image data with inconsistent timestamps, determine the time interval Δt and motion features between the image data with adjacent timestamps, the motion features including the speed of the target object and acceleration According to the time interval Δt and acceleration The position of the target object in the middle frame is calculated by formula (1): Based on the position of the target object in the intermediate frame Get the target pixel point Ptarget; In the formula, is the position of the target object in the middle frame; is the position of the target object in the first frame; The average speed of the target object is calculated by formula (2): In the formula, v i is the speed of the target object in the i-th frame, and n is the number of acquisition frames; Determining a resolution R of the image data, the resolution including the number of pixels or the physical size of the image; Set the speed threshold Vthreshold and the resolution threshold Rthreshold; like Or R<Rthreshold, interpolation calculation is performed based on the bilinear interpolation algorithm to generate an intermediate frame; like If R≥Rthreshold, the interpolation calculation is performed based on the bicubic interpolation algorithm to generate an intermediate frame; Fill in the missing information in the time series based on the generated intermediate frames and detect motion information; The interpolation calculation based on the bilinear interpolation algorithm and the generation of the intermediate frame include: The value of each pixel in the intermediate frame is determined by the weighted average of the four nearest neighboring pixels P1, P2, P3, and P4 to the target pixel Ptarget. The weight of the pixel is determined by the position of the target object in the intermediate frame. The distance to the nearest neighbor pixel is determined by the weight W i It is inversely proportional to the distance, as shown in formula (3); Where P i is the pixel point; W i is the pixel point P i The weight of The interpolation calculation based on the bicubic interpolation algorithm and the generation of the intermediate frame comprises: Determine the 16 pixels around the target pixel Ptarget to form a 4x4 grid; Use a third-order polynomial fit to determine the coefficient c corresponding to each pixel in the 4x4 grid ij , where i and j range from 0 to 3; Constructing a bicubic polynomial function As shown in formula (4), the gray value of the target pixel Ptarget is calculated; Where x and y are coordinate variables in the image space.

2. According to claim 1, a motion trajectory information detection method based on intelligent AI decoding algorithm is characterized in that: The method further comprises: For each pair of adjacent known frames and generated intermediate frames, pixel-level comparison is performed to identify the difference areas; Through comparative analysis, the parts in the difference area that do not belong to background changes or normal motion features are marked as information missing areas.

3. According to claim 2, a motion trajectory information detection method based on intelligent AI decoding algorithm is characterized in that: The method further comprises: For the identified information missing regions, the information missing types are classified according to their distribution patterns and durations in consecutive frames, and the information missing types include random missing, periodic missing, and specific event missing.

4. According to claim 3, a motion trajectory information detection method based on intelligent AI decoding algorithm is characterized in that: The method further comprises: For persistent information missing areas, the pixel values ​​of the missing areas in the intermediate frames are obtained by combining the pixel changes of adjacent frames and the motion characteristics of the target object. Based on the pixel values ​​of the missing areas in the intermediate frames, pixel data for filling the missing areas is generated to ensure the consistency of the filled content in time series and space; Based on the filled intermediate frames, motion information detection is performed, including target object recognition, positioning, tracking and motion trajectory construction; By analyzing the motion pattern and speed parameters of the target object in the filled frame, the motion state of the target object can be predicted.

5. The motion trajectory information detection method based on the intelligent AI decoding algorithm according to claim 1 is characterized in that: The method further comprises: Based on the position of the target object in the intermediate frame As the search center, a local search is performed in the middle frame to determine the target pixel point Ptarget.

6. A motion trajectory information detection system based on an intelligent AI decoding algorithm, used to execute the method according to any one of claims 1 to 5, characterized in that: The system comprises: At least three high-speed camera devices, each of which is equipped with a time synchronization device, a positioning device and an atomic clock, for collecting multi-angle image data of the target object; A time synchronization controller, used to send a unified trigger signal to each high-speed camera device according to a preset time accuracy; A data collection unit, used to collect image data collected by each camera device and its corresponding time stamp, and perform data synchronization processing; Motion feature analyzer, used to determine the time interval Δt and the motion features of the target object, including speed, from the image data with adjacent time stamps and acceleration Position prediction unit, used to predict the position of the vehicle based on the time interval Δt and acceleration Calculate the position of the target object in the intermediate frame Interpolation calculation unit, used to calculate the average speed of the target object The resolution R of the image data is interpolated and an intermediate frame is generated; An information filling unit, used to fill in missing information in the time series based on the generated intermediate frames; Motion information detector, used to detect motion information of the filled intermediate frames, including target object recognition, positioning, tracking and motion trajectory construction; The data processing unit is used to analyze the motion pattern and speed parameters of the target object in the filled frame, predict the motion state of the target object, and provide corresponding data output.

7. An electronic device, characterized in that: The method comprises a memory and a processor, wherein a computer program is stored in the memory, and when the computer program is executed by the processor, the method according to any one of claims 1 to 5 is executed.

8. A storage medium, characterized in that: The computer program stored in the storage medium can be executed by one or more processors and can be used to implement the method as claimed in any one of claims 1 to 5.

Images