CN115965654A - High-altitude parabolic tracking method, device, equipment and medium - Google Patents
High-altitude parabolic tracking method, device, equipment and medium Download PDFInfo
- Publication number
- CN115965654A CN115965654A CN202211704001.5A CN202211704001A CN115965654A CN 115965654 A CN115965654 A CN 115965654A CN 202211704001 A CN202211704001 A CN 202211704001A CN 115965654 A CN115965654 A CN 115965654A
- Authority
- CN
- China
- Prior art keywords
- frame
- area
- detection
- prediction
- value
- 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.)
- Pending
Links
Images
Classifications
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A90/00—Technologies having an indirect contribution to adaptation to climate change
- Y02A90/10—Information and communication technologies [ICT] supporting adaptation to climate change, e.g. for weather forecasting or climate simulation
Landscapes
- Image Analysis (AREA)
Abstract
The application discloses a high-altitude parabolic tracking method, a high-altitude parabolic tracking device, high-altitude parabolic tracking equipment and a high-altitude parabolic tracking medium, which relate to the technical field of computers, and the method comprises the following steps: acquiring a high-altitude parabolic current frame detection frame in the current frame image, and acquiring a current frame prediction frame of a previous frame target frame in the current frame image; calculating the area intersection degree of the frame prediction frame and the frame detection frame to obtain an area intersection value, and calculating the distance similarity degree of the frame prediction frame and the frame detection frame to obtain a distance similarity value; calculating the area ratio of the first intersection area of the tracking area corresponding to the frame prediction frame and the frame detection frame to the area of the frame detection frame, and calculating the transverse movement degree of the frame prediction frame and the frame detection frame based on the tracking area to obtain a movement angle cosine value; and selecting the target frame of the frame corresponding to the prediction frame of the frame from the detection frame of the frame based on the area intersection value, the distance similarity value, the area ratio and the cosine value of the movement angle. The method and the device can reduce the influence of various interferences and improve the detection tracking precision.
Description
Technical Field
The invention relates to the technical field of computers, in particular to a high-altitude parabolic tracking method, device, equipment and medium.
Background
At present, each high-rise building has frequent falling objects, serious potential safety hazards exist, and the high-altitude parabolic detection method is taken as an important intelligent monitoring means, so that the importance is self-evident. The detection of the high-altitude parabolic object generally comprises three parts of object detection, tracking matching and parabolic object identification, and generally comprises the steps of tracking the object by using Kalman prediction and Hungarian matching methods, matching the detected object track and then judging the parabolic object according to the tracking track. Due to the fact that the parabolic detection scene is complex, the target is small, the falling speed is high, and meanwhile, a plurality of interference factors such as birds and balcony clothes exist, the detection accuracy of the high-altitude parabolic detection is difficult to improve. In the prior art, after a target is detected, false detection target filtering is performed according to target characteristics to realize final parabolic identification, but actually, in a tracking stage after the target is detected, due to the limitation of the influence factors, tracking errors are very easily caused, the target cannot be correctly matched, and the accuracy of subsequent parabolic identification is directly influenced.
The method is based on various modes such as a traditional machine vision method, a deep learning method and the like to detect and track the moving target, such as a background difference method, a deep network detection method, gaussian background modeling, multi-target tracking and the like, to obtain the moving information of the parabolic target, but in a real detection scene, a plurality of interference factors such as birds, balcony clothes, light and the like exist, so that the high-altitude parabolic detection has a certain false detection problem.
In summary, how to reduce the influence of various types of interference and improve the detection tracking accuracy is a problem to be solved urgently at present.
Disclosure of Invention
In view of this, the present invention provides a high altitude parabolic tracking method, apparatus, device and medium, which can reduce the influence of various interferences and improve the detection and tracking accuracy. The specific scheme is as follows:
in a first aspect, the present application discloses a high altitude parabolic tracking method, including:
acquiring a frame detection frame of a high-altitude parabola in a frame image, and acquiring a frame prediction frame of a previous frame target frame in the frame image; the previous frame target frame is a detection frame of a high altitude parabola in the previous frame image;
calculating the area intersection degree of the frame prediction frame and the frame detection frame to obtain an area intersection value, and calculating the distance similarity degree of the frame prediction frame and the frame detection frame to obtain a distance similarity value;
calculating the area ratio of a first intersection area of the tracking area corresponding to the frame prediction frame and the frame detection frame to the area of the frame detection frame, and calculating the transverse movement degree of the frame prediction frame and the frame detection frame based on the tracking area to obtain a movement angle cosine value;
and selecting the target frame of the current frame corresponding to the prediction frame of the current frame from the detection frame of the current frame based on the area intersection value, the distance similarity value, the area ratio and the moving angle cosine value.
Optionally, the calculating the similarity between the current frame prediction frame and the current frame detection frame to obtain a similarity value includes:
calculating a target distance between the central point of the current frame prediction frame and the central point of the current frame detection frame, and calculating a diagonal distance of the current frame image;
and calculating the ratio of the target distance to the diagonal distance to calculate the distance similarity degree to obtain a distance similarity value between the current frame prediction frame and the current frame detection frame.
Optionally, before calculating an area ratio of a first intersection area of the tracking area corresponding to the current frame prediction frame and the current frame detection frame to an area of the current frame detection frame, the method further includes:
calculating the height and width of the tracking area corresponding to the current frame prediction frame based on the first coordinate of the central point of the current frame prediction frame and the width and height of the current frame prediction frame;
determining the tracking area based on the first coordinates and a height and a width of the tracking area.
Optionally, the calculating the horizontal movement degree of the current frame prediction frame and the current frame detection frame based on the tracking area to obtain a movement angle cosine value includes:
determining a first vector based on the first coordinate and a second coordinate of the center point of the frame detection frame, determining a second vector based on a middle coordinate close to the boundary of the frame detection frame in the tracking area and the first coordinate, and determining a third vector based on the intersection coordinate between the center point of the frame prediction frame and the center point of the frame detection frame and the first coordinate;
calculating a first angular cosine value based on the second vector and the third vector, and calculating a second angular cosine value based on the first vector and the third vector;
and if the first angle cosine value is not smaller than the second angle cosine value, taking the second angle cosine value as the moving angle cosine value, and if the first angle cosine value is smaller than the second angle cosine value, taking a preset value as the moving angle cosine value.
Optionally, the calculating the area intersection degree of the frame prediction box and the frame detection box to obtain the area intersection value includes:
calculating a second intersection area of the frame prediction frame and the frame detection frame, and calculating a difference area obtained by subtracting the second intersection area from the common total area of the frame prediction frame and the frame detection frame;
and calculating the ratio of the second intersection area to the difference area so as to calculate the area intersection degree to obtain the area intersection value between the current frame prediction frame and the current frame detection frame.
Optionally, the selecting a target frame of the current frame corresponding to the prediction frame of the current frame from the detection frame of the current frame based on the area intersection value, the distance proximity value, the area ratio, and the moving angle cosine value includes:
calculating the matching distance between the frame prediction frame and the frame detection frame based on the area intersection value, the distance similarity value, the area ratio and the moving angle cosine value;
and determining the minimum distance corresponding to the frame prediction frame from the matching distances, and determining the frame detection frame corresponding to the minimum distance as the frame target frame corresponding to the frame prediction frame.
Optionally, the obtaining a current frame prediction frame of a previous frame target frame in the current frame image includes:
determining ID information of the previous frame target frame in the previous frame image;
acquiring a frame prediction frame of a previous frame target frame in the frame image, and transmitting the ID information to the frame prediction frame;
correspondingly, after the frame target frame corresponding to the frame prediction frame is selected from the frame detection frame based on the area intersection value, the distance closeness value, the area ratio and the moving angle cosine value, the method further includes:
and transmitting the ID information corresponding to the frame prediction frame to the corresponding frame target frame.
In a second aspect, the present application discloses a high altitude parabolic tracking apparatus, comprising:
the frame detection frame acquisition module is used for acquiring a frame detection frame of a high-altitude parabola in the frame image;
the frame prediction frame obtaining module is used for obtaining a frame prediction frame of a previous frame target frame in the frame image; the previous frame target frame is a detection frame of a high altitude parabola in the previous frame image;
the area intersection value calculation module is used for calculating the area intersection degree of the current frame prediction frame and the current frame detection frame to obtain an area intersection value;
a distance similarity value calculation module, configured to calculate a distance similarity degree between the current frame prediction frame and the current frame detection frame to obtain a distance similarity value;
the area ratio calculation module is used for calculating the area ratio of a first intersection area of the tracking area corresponding to the current frame prediction frame and the current frame detection frame to the area of the current frame detection frame;
a moving angle cosine value calculation module, configured to perform lateral movement degree calculation on the current frame prediction frame and the current frame detection frame based on the tracking area to obtain a moving angle cosine value;
and the selection module is used for selecting the target frame of the current frame corresponding to the prediction frame of the current frame from the detection frame of the current frame based on the area intersection value, the distance similarity value, the area ratio and the cosine value of the movement angle.
In a third aspect, the present application discloses an electronic device comprising a processor and a memory; wherein the processor, when executing the computer program stored in the memory, implements the high altitude parabolic tracking method disclosed above.
In a fourth aspect, the present application discloses a computer readable storage medium for storing a computer program; wherein the computer program when executed by a processor implements the high altitude parabolic tracking method disclosed in the foregoing.
Therefore, the method comprises the steps of obtaining a high-altitude parabolic current frame detection frame in a current frame image, and obtaining a current frame prediction frame of a previous frame target frame in the current frame image; the previous frame target frame is a detection frame of a high altitude parabola in the previous frame image; calculating the area intersection degree of the frame prediction frame and the frame detection frame to obtain an area intersection value, and calculating the distance similarity degree of the frame prediction frame and the frame detection frame to obtain a distance similarity value; calculating the area ratio of a first intersection area of the tracking area corresponding to the frame prediction frame and the frame detection frame to the area of the frame detection frame, and calculating the transverse movement degree of the frame prediction frame and the frame detection frame based on the tracking area to obtain a movement angle cosine value; and selecting the target frame of the current frame corresponding to the prediction frame of the current frame from the detection frame of the current frame based on the area intersection value, the distance similarity value, the area ratio and the moving angle cosine value. Therefore, the target frame of the current frame corresponding to the prediction frame of the current frame is selected by using the area intersection value, the distance similarity value, the area proportion and the moving angle cosine value, the influence of various interferences can be reduced, and the detection tracking precision is improved.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to the provided drawings without creative efforts.
Fig. 1 is a flowchart of a high altitude parabolic tracking method provided in the present application;
FIG. 2 is a schematic diagram of a high altitude parabolic condition provided by the present application;
fig. 3 is a schematic diagram of a first tracking area acquisition provided in the present application;
fig. 4 is a schematic diagram of a second tracking area acquisition provided in the present application;
FIG. 5 is a schematic diagram illustrating cosine value calculation of a shift angle provided herein;
FIG. 6 is a flow chart of a specific high altitude parabolic tracking method provided by the present application;
fig. 7 is a schematic structural diagram of a high altitude parabolic tracking apparatus provided in the present application;
fig. 8 is a block diagram of an electronic device provided in the present application.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
At present, in the reality detection scene, there are interference factors such as a lot of birds, balcony clothing, light for the high altitude is thrown thing and is detected and to have certain false retrieval problem.
In order to overcome the problems, the application provides a high-altitude parabolic tracking scheme which can reduce the influence of various interferences and improve the detection tracking precision.
Referring to fig. 1, an embodiment of the present application discloses a high altitude parabolic tracking method, including:
step S11: acquiring a frame detection frame of a high-altitude parabola in a frame image, and acquiring a frame prediction frame of a previous frame target frame in the frame image; and the target frame of the previous frame is a detection frame of a high-altitude parabola in the image of the previous frame.
In the embodiment of the application, the foreground target is obtained by adopting a traditional image processing Gaussian background modeling or VIBE and other machine vision methods, and any moving target meeting the moving condition in the picture can be obtained. The traditional algorithm has the big defect that the real parabolic target and other interference targets such as balcony clothes, bird lights and the like cannot be distinguished, and besides the falling condition judgment, the tracking module also needs to solve the interference factors and ensure the accuracy of parabolic target detection.
It should be noted that, the current frame image corresponds to a plurality of current frame detection frames, the previous frame target frame has a plurality of current frame prediction frames in the current frame image, and when the calculation between the current frame detection frame and the current frame prediction frame is subsequently performed, the calculation between one current frame detection frame and one current frame prediction frame is performed until the calculation between one current frame detection frame and all current frame prediction frames and the calculation between one current frame prediction frame and all current frame detection frames are completed.
In the embodiment of the application, a current frame prediction frame of a previous frame target frame in the current frame image is obtained through a Kalman filter.
Step S12: and calculating the area intersection degree of the frame prediction frame and the frame detection frame to obtain an area intersection value, and calculating the distance closeness degree of the frame prediction frame and the frame detection frame to obtain a distance closeness value.
In this embodiment of the present application, the calculating the area intersection degree of the frame prediction box and the frame detection box to obtain the area intersection value includes: calculating a second intersection area of the frame prediction frame and the frame detection frame, and calculating a difference area obtained by subtracting the second intersection area from the common total area of the frame prediction frame and the frame detection frame; and calculating the ratio of the second intersection area to the difference area so as to calculate the area intersection degree to obtain the area intersection value between the frame prediction frame and the frame detection frame.
It should be noted that, when the present application performs matching between the frame prediction box and the frame detection box, the area intersection degree of the boxes needs to be considered, that is, the area similarity between the frame prediction box and the frame detection box is considered, the area may be a common area or an IOU area, and a first formula (a first cost matrix) for calculating the area intersection degree to obtain the area intersection value is as follows:
wherein, area (Dbox) is the area of the detection frame of the current frame; area (Tbox) is the prediction frame area of the current frame; area (Dbox:. Tbox) is the intersection area of the frame prediction block and the frame detection block, T can represent the frame prediction block, and D can represent the frame detection block.
In this embodiment of the present application, the calculating a proximity degree between the current frame prediction frame and the current frame detection frame to obtain a proximity value includes: calculating a target distance between the central point of the current frame prediction frame and the central point of the current frame detection frame, and calculating a diagonal distance of the current frame image; and calculating the ratio of the target distance to the diagonal distance to calculate the distance similarity degree to obtain a distance similarity value between the current frame prediction frame and the current frame detection frame.
It should be noted that, when a thrower or balcony clothes shakes, erroneous tracking due to detection of a plurality of targets around the target is easily caused; or under the scenes of jitter and the like, other targets exist around the false detection target, so that the cost matrix value generated according to IOU calculation is small, the cost matrix value is easily matched with the same target, the false detection target is successfully tracked, the falling condition is met, and false alarm is caused, as shown in fig. 2, the target 2 and the target 1 with a large area are the detection frame of the current frame, and the target 1 with a small area is the prediction frame of the current frame; the area, the width-height ratio and the like of the target 1 with a smaller area are the same as those of the target 2, the area intersection value is larger, the target 1 with a smaller area and the similarity are closer to the target 1 with a larger area, and the similarity between the frame detection frame and the frame prediction frame cannot be accurately evaluated only by means of the IOU, the width-height ratio, the area and the like, so that the tracking effect is influenced. For this case, it is important to add a distance factor to evaluate the distance similarity between the detection box and the prediction box.
In this embodiment of the present application, the calculating the proximity degree between the frame prediction block and the frame detection block to obtain a proximity value includes: calculating a target distance between the central point of the frame prediction frame and the central point of the frame detection frame, and calculating a diagonal distance of the frame image; and calculating the ratio of the target distance to the diagonal distance to calculate the distance similarity degree to obtain a distance similarity value between the current frame prediction frame and the current frame detection frame. It should be noted that the second formula (the second cost matrix) for calculating the proximity to obtain the proximity value is:
wherein (x) 2 ,y 2 ) The coordinate of the center point of the frame is detected as (x) 1 ,y 1 ) For the coordinate of the center point of the frame prediction box, (w) img ,h img ) Is the width and height of the image of the frame, R D And calculating to obtain the ratio of the distance between the prediction frame and the detection frame of the current frame to the diagonal distance of the image (distance approximate value), wherein the closer the distance between the prediction frame and the detection frame of the current frame is, the smaller the value is, and the distance cost between the detection frame and the prediction frame is evaluated.
Step S13: calculating the area ratio of the first intersection area of the tracking area corresponding to the current frame prediction frame and the current frame detection frame to the area of the current frame detection frame, and calculating the transverse movement degree of the current frame prediction frame and the current frame detection frame based on the tracking area to obtain a movement angle cosine value.
In this embodiment of the present application, before calculating an area ratio of a first intersection area of the tracking area corresponding to the current frame prediction frame and the current frame detection frame to an area of the current frame detection frame, the method further includes: calculating the height and width of the tracking area corresponding to the frame prediction frame based on the first coordinate of the central point of the frame prediction frame and the width and height of the frame prediction frame; determining the tracking area based on the first coordinates and a height and a width of the tracking area.
It should be noted that, there are light between floors, and different balcony clothes shake to cause the situation of target linkage tracking, as shown in fig. 3, the target 2 and the target 1 with a large area are the detection frame of the present frame, and the target 1 with a small area is the prediction frame of the present frame, relatively speaking, the distance between the target 2 and the target 1 with a small area is closer except for the width, height, and area, and the similarity between the prediction frame and the detection frame cannot be accurately evaluated. For the situation, the actual parabolic law is combined, the position of the current frame detection result is located in the lower area of the previous frame target in the target falling process, namely the target is in the falling process from top to bottom, the limitation on the tracking area is added by combining the actual law, and the problem of false detection of the non-parabolic position area can be effectively reduced.
In one embodiment, the coordinates (x) of the center point of the frame are first predicted according to the current frame 1 ,y 1 ) And width and height (w) of the frame prediction frame 1 ,h 1 ) The width and height of the tracking area are calculated by the following formula:
w TA ＝min(w img -x 1 ,w 1 *n)；
h TA ＝min(h ing -y 1 ,h 1 *n)；
wherein n is an empirical value and represents that the tracking area is several times of the width and the height of the prediction frame of the current frame; (w) img ,h img ) TA may indicate a tracking area for the width and height of the present frame image.
Secondly, the coordinate (x) of the center point of the frame is predicted according to the frame 1 ,y 1 ) And width and height (w) of the frame prediction frame 1 ,h 1 ) The coordinates of the upper left corner of the tracking area are calculated as follows:
x TA ＝max(0,x 1 -w 1 *n)；
y TA ＝y 1 ；
the first tracking area as shown with reference to fig. 3 is obtained from the width and height of the tracking area and the coordinates of the upper left corner of the tracking area. Note that, the frame detection frame used for searching the target ranges n below, left, right, and left of the prediction frame is the minimum range (minimum tracking area) within the dotted line.
In another embodiment, after the first tracking area shown in fig. 3 is obtained, the areas on the left, right, and below the first tracking area are not limited and extend to the boundary of the image of the present frame, so that the second tracking area shown in fig. 4 is obtained.
In the embodiment of the present application, a third formula (a third price matrix) for calculating an area ratio of a first intersection area of the tracking area corresponding to the current frame prediction frame and the current frame detection frame to an area of the current frame detection frame is:
wherein, area (Dbox:TAbox) is the first intersection area of the tracking area and the frame detection frame; area (Dbox) is the area of the detection frame of the current frame. It should be noted that the larger the area ratio of the first intersection area between the tracking area and the current frame detection frame is, the greater the similarity between the current frame detection frame and the current frame prediction frame is. It should be noted that limiting the tracking area can better eliminate the detection frame, and improve the tracking accuracy.
In this embodiment of the application, the calculating the degree of lateral movement of the frame prediction box and the frame detection box based on the tracking area to obtain a cosine value of a movement angle includes: determining a first vector based on the first coordinate and a second coordinate of the center point of the current frame detection frame, determining a second vector based on a middle coordinate close to the boundary of the current frame detection frame in the tracking area and the first coordinate, and determining a third vector based on the intersection point coordinate between the center point of the current frame prediction frame and the center point of the current frame detection frame and the first coordinate; calculating a first angular cosine value based on the second vector and the third vector, and a second angular cosine value based on the first vector and the third vector; and if the first angle cosine value is not smaller than the second angle cosine value, taking the second angle cosine value as the moving angle cosine value, and if the first angle cosine value is smaller than the second angle cosine value, taking a preset value as the moving angle cosine value.
It should be noted that, due to the occurrence of interference targets such as reflection of window glass of some buildings and snowy days, the similarity between the prediction frame of the current frame and the interference detection frame in the current frame is high, and based on the tracking area, a tracking angle factor is added, and the horizontal movement degree is calculated to obtain a movement angle cosine value, so that the tracking accuracy is further increased.
In a specific embodiment, the first moving angle cosine value is calculated according to the first tracking area, as shown in fig. 5, the high middle coordinate P (x) of the first tracking area close to the boundary of the frame detection frame is taken TA ,y TA +h TA And 2), the coordinate of the central point of the frame prediction frame is A (x) 1 ,y 1 ) The coordinate of the central point of the frame detection frame is B (x) 2 ,y 2 ) Coordinate C of intersection point of center point of prediction frame and center point of detection frame of current frame 1 (x 2 ,y 1 ) A first vector determined based on the first coordinate and the second coordinate of the central point of the frame detection frame is a vector V AB A second vector determined based on the high middle coordinate close to the boundary of the current frame detection frame in the tracking area and the first coordinate is a vector V AP Determining a third vector based on the intersection point coordinate between the central point of the current frame prediction frame and the central point of the current frame detection frame and the first coordinate as a vector
α 1 For a first angle obtained based on the second vector and the third vector, a first angle cosine value calculation formula is as follows:
α 2 based on the first directionA second angle obtained by measuring and the third vector, and a second angle cosine value calculation formula is as follows:
a fourth formula (fourth cost matrix) for calculating the cosine value of the movement angle is as follows:
wherein, T is a preset value, and when the angle is larger than a certain value, a larger loss is assigned, generally set to 1.5.
In another specific embodiment, the second movement angle cosine value is calculated according to the second tracking area, only P needs to be modified to the high middle coordinate close to the boundary of the detection frame of the current frame in the second tracking area as shown in fig. 4, the boundary of the second tracking area is the boundary of the image of the current frame, and other steps are the same as the previous embodiment and are not specifically described herein.
In summary, the tracking method of the invention adds the limiting factors such as the tracking area and the angle range besides considering the similarity (area and distance) between the targets, improves the accuracy of tracking matching and the tracking angle range of scene self-adaption, effectively judges the consistency of the targets, and solves the problem of abnormal tracking.
Step S14: and selecting the target frame of the current frame corresponding to the prediction frame of the current frame from the detection frame of the current frame based on the area intersection value, the distance similarity value, the area ratio and the moving angle cosine value.
In this embodiment of the application, the selecting, from the frame detection frame, the frame target frame corresponding to the frame prediction frame based on the area intersection value, the distance closeness value, the area ratio, and the movement angle cosine value includes: calculating the matching distance between the frame prediction frame and the frame detection frame based on the area intersection value, the distance approximation value, the area ratio and the moving angle cosine value; and determining the minimum distance corresponding to the frame prediction frame from the matching distance, and determining the frame detection frame corresponding to the minimum distance as the frame target frame corresponding to the frame prediction frame.
It should be noted that, a fourth formula (a fourth cost matrix) for calculating the matching distance between the current frame prediction box and the current frame detection box based on the area intersection value, the distance proximity value, the area ratio, and the moving angle cosine value is as follows:
Aff D ＝β 1 *(1-R B )+β 2 *R D +β 3 *(1-R A )+β 4 *R C ；
wherein beta is 1 ,β 2 ,β 3 ,β 4 For the weight coefficient of each cost value, the cost matrix weighted value of each matched pair of the detection result (the frame detection frame) and the prediction result (the frame prediction frame) is used for evaluating the similarity of each matched pair; by combining the cost matrix calculation mode, the problems of missed detection caused by tracking errors and matching errors of real-time parabolic targets can be effectively solved, and the accuracy of matching and tracking is improved.
Therefore, the method comprises the steps of obtaining a high-altitude parabolic current frame detection frame in a current frame image, and obtaining a current frame prediction frame of a previous frame target frame in the current frame image; the previous frame target frame is a detection frame of a high altitude parabola in the previous frame image; calculating the area intersection degree of the frame prediction frame and the frame detection frame to obtain an area intersection value, and calculating the distance similarity degree of the frame prediction frame and the frame detection frame to obtain a distance similarity value; calculating the area ratio of a first intersection area of the tracking area corresponding to the frame prediction frame and the frame detection frame to the area of the frame detection frame, and calculating the transverse movement degree of the frame prediction frame and the frame detection frame based on the tracking area to obtain a movement angle cosine value; and selecting the target frame of the current frame corresponding to the prediction frame of the current frame from the detection frame of the current frame based on the area intersection value, the distance similarity value, the area ratio and the moving angle cosine value. Therefore, the target frame corresponding to the frame prediction frame is selected by using the area intersection value, the distance similarity value, the area ratio and the moving angle cosine value, so that the influence of various interferences can be reduced, and the detection tracking precision is improved.
Referring to fig. 6, an embodiment of the present application discloses a specific high altitude parabolic tracking method, including:
step S21: acquiring a high-altitude parabolic current frame detection frame in a current frame image, and determining the ID information of the previous frame target frame in the previous frame image; acquiring a frame prediction frame of a previous frame target frame in the frame image, and transmitting the ID information to the frame prediction frame; and the target frame of the previous frame is a detection frame of a high-altitude parabola in the image of the previous frame.
In the embodiment of the present application, the high-altitude parabolic tracking is performed, that is, a previous frame target frame and a current frame target frame corresponding to the same parabolic are also detected, and the transfer of the ID (Identity document) information of the high-altitude parabolic in the target frames corresponding to the high-altitude parabolic in different frame images is performed, specifically, the transfer between the previous frame target frame and the current frame target frame is performed through the current frame prediction frame. It should be noted that the parabolas in the previous frame object box and the current frame object box with the same ID are the same parabola.
It should be noted that, a tracker may be created for the previous frame target frame in the previous frame image, and the ID information of the high altitude parabola corresponding to the previous frame target frame in the previous frame image of the previous frame may be configured for the tracker, or when the high altitude parabola corresponding to the previous frame target frame does not exist in the previous frame image of the previous frame, new ID information may be configured.
In the embodiment of the application, a current frame prediction frame of a previous frame target frame in the current frame image is obtained through a Kalman filter.
Step S22: and calculating the area intersection degree of the frame prediction frame and the frame detection frame to obtain an area intersection value, and calculating the distance closeness degree of the frame prediction frame and the frame detection frame to obtain a distance closeness value.
For a more specific processing procedure of step S22, reference may be made to corresponding contents disclosed in the foregoing embodiments, and details are not repeated here.
Step S23: and calculating the area ratio of the first intersection area of the tracking area corresponding to the frame prediction frame and the frame detection frame to the area of the frame detection frame, and calculating the transverse movement degree of the frame prediction frame and the frame detection frame based on the tracking area to obtain a movement angle cosine value.
For a more specific processing procedure of step S23, reference may be made to corresponding contents disclosed in the foregoing embodiments, and details are not repeated here.
Step S24: and selecting the target frame of the current frame corresponding to the prediction frame of the current frame from the detection frame of the current frame based on the area intersection value, the distance similarity value, the area ratio and the cosine value of the movement angle, and transmitting the ID information corresponding to the prediction frame of the current frame to the corresponding target frame of the current frame.
In the embodiment of the application, if the frame prediction frame of a certain high-altitude parabolic object is not matched with the corresponding frame detection frame, the next frame detection frame corresponding to the frame prediction frame of the certain high-altitude parabolic object is continuously matched in the tracking process of the next frame image, if the frame prediction frame is not matched for a plurality of times, the target is judged to disappear, and the tracking is finished. It should be noted that, during the tracking process of the next frame image, the tracker for the next frame prediction frame is regenerated, and if the tracker is not matched for several times, the tracker is deleted.
It should be noted that the cost matrix calculation mechanism provided by the present application guarantees the accuracy of matching the same target to the same ID, and effectively solves the problems of false detection and missed detection caused by tracking errors.
In summary, the moving target detection is performed based on the traditional machine vision method, the detected moving target is put into the tracking module, and the Hungarian matching and Kalman tracking methods are adopted to realize the tracking of the same parabolic target. Aiming at the problem of false detection caused by interference targets such as lamplight, balcony clothes, flying birds and inverted images, a tracking module before parabolic behavior judgment is eliminated, factors such as tracking areas, tracking angles and distances are added for the false detection problems of different types, the weighted cost value between a prediction result and a detection result matching pair is calculated, the matching relation between the prediction result and the detection result is accurately evaluated, the false detection problem caused by tracking errors is effectively avoided from the tracking module, in addition, innovation is carried out on how to match real parabolic targets into the same ID, the problems of missing detection and real parabolic target matching errors caused by tracking errors and the problem of interference target tracking errors can be effectively solved, and the parabolic judgment precision is effectively improved.
Therefore, the ID information of the previous frame target frame in the previous frame image is determined; acquiring a current frame prediction frame of a previous frame target frame in the current frame image, transmitting the ID information to the current frame prediction frame, and acquiring the current frame prediction frame of the previous frame target frame in the current frame image; the previous frame target frame is a detection frame of a high-altitude parabola in the previous frame image; calculating the area intersection degree of the frame prediction frame and the frame detection frame to obtain an area intersection value, and calculating the distance similarity degree of the frame prediction frame and the frame detection frame to obtain a distance similarity value; calculating the area ratio of a first intersection area of the tracking area corresponding to the frame prediction frame and the frame detection frame to the area of the frame detection frame, and calculating the transverse movement degree of the frame prediction frame and the frame detection frame based on the tracking area to obtain a movement angle cosine value; and selecting the target frame of the current frame corresponding to the prediction frame of the current frame from the detection frame of the current frame based on the area intersection value, the distance similarity value, the area ratio and the cosine value of the movement angle, and transmitting the ID information corresponding to the prediction frame of the current frame to the corresponding target frame of the current frame. Therefore, the target frame corresponding to the frame prediction frame is selected by using the area intersection value, the distance similarity value, the area ratio and the moving angle cosine value, so that the influence of various interferences can be reduced, and the detection tracking precision is improved.
In summary, the tracking process of the present application mainly includes the following steps:
the method comprises the following steps: obtaining a moving target: because a large amount of training materials are needed for deep learning training detection models, all target types which can appear are traversed as much as possible, but in an actual application scene, the target types are various and cannot be exhausted. The foreground target is obtained by adopting machine vision methods such as traditional image processing Gaussian background modeling or VIBE (fixed machine vision camera), and the like, and any moving target meeting the moving condition in the picture can be obtained. The traditional algorithm has the big defect that a real parabolic target and other interference targets such as balcony clothes, bird light and the like cannot be distinguished, and besides the falling condition judgment, the tracking module also needs to solve the interference factors and guarantee the accuracy of parabolic target detection.
Step two: a new tracker is initialized and created with the detected object in the previous frame image frame and given an ID.
Step three: acquiring a target frame of a moving target detected in a previous frame of image through a Kalman filter to generate state prediction and covariance prediction;
step four: after the detection target in the current frame image is obtained, calculating cost matrixes of all target state prediction target frames and the target frames detected in the current frame, obtaining the only match with the minimum cost matrix through Hungary algorithm, and removing the matching pairs with the cost matrix value larger than a preset threshold value.
Step five: if the matching is successful in the fourth step, acquiring a tracking result of the current frame target; if the detection frames which are not matched exist, a new tracker is established, and the second step is repeated; and maintaining the unmatched targets in a waiting state, if the waiting times are greater than a preset threshold value, disappearing the targets, finishing tracking and deleting the target tracker.
Referring to fig. 7, an embodiment of the present application discloses a high altitude parabolic tracking apparatus, including:
a frame detection frame obtaining module 11, configured to obtain a frame detection frame of a high-altitude parabola in the frame image;
a frame prediction frame obtaining module 12, configured to obtain a frame prediction frame of a previous frame target frame in the frame image; the previous frame target frame is a detection frame of a high-altitude parabola in the previous frame image;
an area intersection value calculating module 13, configured to calculate an area intersection degree of the frame prediction box and the frame detection box to obtain an area intersection value;
a distance similarity value calculation module 14, configured to calculate distance similarity between the current frame prediction frame and the current frame detection frame to obtain a distance similarity value;
an area ratio calculating module 15, configured to calculate an area ratio, where a first intersection area of the tracking area corresponding to the current frame prediction frame and the current frame detection frame occupies an area of the current frame detection frame;
a moving angle cosine value calculation module 16, configured to perform lateral movement degree calculation on the current frame prediction frame and the current frame detection frame based on the tracking area to obtain a moving angle cosine value;
and a selecting module 17, configured to select a frame target frame corresponding to the frame prediction frame from the frame detection frame based on the area intersection value, the distance closeness value, the area ratio, and the moving angle cosine value.
For more specific working processes of the modules, reference may be made to corresponding contents disclosed in the foregoing embodiments, and details are not repeated here.
Therefore, the method comprises the steps of obtaining a high-altitude parabolic current frame detection frame in a current frame image, and obtaining a current frame prediction frame of a previous frame target frame in the current frame image; the previous frame target frame is a detection frame of a high-altitude parabola in the previous frame image; calculating the area intersection degree of the frame prediction frame and the frame detection frame to obtain an area intersection value, and calculating the distance similarity degree of the frame prediction frame and the frame detection frame to obtain a distance similarity value; calculating the area ratio of a first intersection area of the tracking area corresponding to the frame prediction frame and the frame detection frame to the area of the frame detection frame, and calculating the transverse movement degree of the frame prediction frame and the frame detection frame based on the tracking area to obtain a movement angle cosine value; and selecting the target frame of the current frame corresponding to the prediction frame of the current frame from the detection frame of the current frame based on the area intersection value, the distance similarity value, the area ratio and the moving angle cosine value. Therefore, the target frame corresponding to the frame prediction frame is selected by using the area intersection value, the distance similarity value, the area ratio and the moving angle cosine value, so that the influence of various interferences can be reduced, and the detection tracking precision is improved.
Further, an electronic device is provided in the embodiments of the present application, and fig. 8 is a block diagram of an electronic device 20 according to an exemplary embodiment, which should not be construed as limiting the scope of the application.
Fig. 8 is a schematic structural diagram of an electronic device 20 according to an embodiment of the present disclosure. The electronic device 20 may specifically include: at least one processor 21, at least one memory 22, a power supply 23, an input output interface 24, a communication interface 25, and a communication bus 26. Wherein the memory 22 is used for storing a computer program, which is loaded and executed by the processor 21 to implement the relevant steps of the high altitude parabolic tracking method disclosed in any of the foregoing embodiments.
In this embodiment, the power supply 23 is configured to provide a working voltage for each hardware device on the electronic device 20; the communication interface 25 can create a data transmission channel between the electronic device 20 and an external device, and the communication protocol followed by the communication interface is any communication protocol that can be applied to the technical solution of the present application, and is not specifically limited herein; the input/output interface 24 is configured to obtain external input data or output data to the outside, and a specific interface type thereof may be selected according to specific application requirements, which is not specifically limited herein.
In addition, the storage 22 is used as a carrier for resource storage, and may be a read-only memory, a random access memory, a magnetic disk or an optical disk, etc., the storage 22 is used as a non-volatile storage that may include a random access memory as an operating memory and a storage purpose for an external memory, and the storage resources on the storage include an operating system 221, a computer program 222, etc., and the storage manner may be a transient storage or a permanent storage.
The operating system 221 is used for managing and controlling each hardware device on the electronic device 20 on the source host and the computer program 222, and the operating system 221 may be Windows, unix, linux, or the like. The computer program 222 may further include a computer program that can be used to perform other specific tasks in addition to the computer program that can be used to perform the high altitude parabolic tracking method performed by the electronic device 20 disclosed in any of the foregoing embodiments.
In this embodiment, the input/output interface 24 may specifically include, but is not limited to, a USB interface, a hard disk reading interface, a serial interface, a voice input interface, a fingerprint input interface, and the like.
Further, the embodiment of the application also discloses a computer readable storage medium for storing a computer program; wherein the computer program when executed by a processor implements the high altitude parabolic tracking method as disclosed in the foregoing.
For the specific steps of the method, reference may be made to corresponding contents disclosed in the foregoing embodiments, and details are not repeated here.
The computer-readable storage medium includes a Random Access Memory (RAM), a Memory, a Read-only Memory (ROM), an electrically programmable ROM, an electrically erasable programmable ROM, a register, a hard disk, a magnetic or optical disk, or any other form of storage medium known in the art. Wherein the computer program when executed by a processor implements the aforementioned high altitude parabolic tracking method. For the specific steps of the method, reference may be made to the corresponding contents disclosed in the foregoing embodiments, which are not described herein again.
In the present specification, the embodiments are described in a progressive manner, and each embodiment focuses on differences from other embodiments, and the same or similar parts between the embodiments are referred to each other. The device disclosed by the embodiment corresponds to the high altitude parabolic tracking method disclosed by the embodiment, so that the description is simple, and the relevant points can be referred to the method part for description.
Those of skill would further appreciate that the various illustrative components and algorithm steps described in connection with the embodiments disclosed herein may be implemented as electronic hardware, computer software, or combinations of both, and that the components and steps of the various examples have been described above generally in terms of their functionality in order to clearly illustrate this interchangeability of hardware and software. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the technical solution. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present application.
The steps of an algorithm described in connection with the embodiments disclosed herein may be embodied directly in hardware, in a software module executed by a processor, or in a combination of the two. A software module may reside in Random Access Memory (RAM), memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disk, a removable disk, a CD-ROM, or any other form of storage medium known in the art.
Finally, it should also be noted that, in this document, relational terms such as first and second, and the like are used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Also, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising a … …" does not exclude the presence of another identical element in a process, method, article, or apparatus that comprises the element.
The method, the device, the equipment and the medium for tracking the high altitude parabola provided by the invention are introduced in detail, specific examples are applied in the text to explain the principle and the implementation mode of the invention, and the description of the embodiments is only used for helping to understand the method and the core idea of the invention; meanwhile, for a person skilled in the art, according to the idea of the present invention, there may be variations in the specific embodiments and the application scope, and in summary, the content of the present specification should not be construed as a limitation to the present invention.
Claims (10)
1. A high altitude parabolic tracking method, comprising:
acquiring a frame detection frame of a high-altitude parabola in a frame image, and acquiring a frame prediction frame of a previous frame target frame in the frame image; the previous frame target frame is a detection frame of a high-altitude parabola in the previous frame image;
calculating the area intersection degree of the frame prediction frame and the frame detection frame to obtain an area intersection value, and calculating the distance similarity degree of the frame prediction frame and the frame detection frame to obtain a distance similarity value;
calculating the area ratio of a first intersection area of the tracking area corresponding to the current frame prediction frame and the current frame detection frame to the area of the current frame detection frame, and calculating the transverse movement degree of the current frame prediction frame and the current frame detection frame based on the tracking area to obtain a movement angle cosine value;
and selecting the target frame of the current frame corresponding to the prediction frame of the current frame from the detection frame of the current frame based on the area intersection value, the distance similarity value, the area proportion and the moving angle cosine value.
2. The high-altitude parabolic tracking method according to claim 1, wherein the calculating of the distance closeness degree of the current frame prediction block and the current frame detection block to obtain a distance closeness value comprises:
calculating a target distance between the central point of the frame prediction frame and the central point of the frame detection frame, and calculating a diagonal distance of the frame image;
and calculating the ratio of the target distance to the diagonal distance to calculate the distance similarity degree to obtain a distance similarity value between the current frame prediction frame and the current frame detection frame.
3. The high altitude parabolic tracking method according to claim 1, wherein before calculating an area ratio of a first intersection area of the tracking area corresponding to the current frame prediction frame and the current frame detection frame to an area of the current frame detection frame, the method further comprises:
calculating the height and width of the tracking area corresponding to the frame prediction frame based on the first coordinate of the central point of the frame prediction frame and the width and height of the frame prediction frame;
determining the tracking area based on the first coordinates and a height and a width of the tracking area.
4. The high-altitude parabolic tracking method according to claim 3, wherein the calculating of the degree of lateral movement of the frame prediction frame and the frame detection frame based on the tracking area to obtain a movement angle cosine value comprises:
determining a first vector based on the first coordinate and a second coordinate of the center point of the frame detection frame, determining a second vector based on a middle coordinate close to the boundary of the frame detection frame in the tracking area and the first coordinate, and determining a third vector based on the intersection coordinate between the center point of the frame prediction frame and the center point of the frame detection frame and the first coordinate;
calculating a first angular cosine value based on the second vector and the third vector, and calculating a second angular cosine value based on the first vector and the third vector;
and if the first angle cosine value is not smaller than the second angle cosine value, taking the second angle cosine value as the moving angle cosine value, and if the first angle cosine value is smaller than the second angle cosine value, taking a preset value as the moving angle cosine value.
5. The high altitude parabolic tracking method according to claim 1, wherein the calculating an area intersection degree of the frame prediction block and the frame detection block to obtain an area intersection value comprises:
calculating a second intersection area of the frame prediction frame and the frame detection frame, and calculating a difference area obtained by subtracting the second intersection area from the common total area of the frame prediction frame and the frame detection frame;
and calculating the ratio of the second intersection area to the difference area so as to calculate the area intersection degree to obtain the area intersection value between the frame prediction frame and the frame detection frame.
6. The high altitude parabolic tracking method according to claim 1, wherein the selecting a current frame target frame corresponding to the current frame prediction frame from the current frame detection frame based on the area intersection value, the distance closeness value, the area ratio, and the movement angle cosine value includes:
calculating the matching distance between the frame prediction frame and the frame detection frame based on the area intersection value, the distance approximation value, the area ratio and the moving angle cosine value;
and determining the minimum distance corresponding to the frame prediction frame from the matching distance, and determining the frame detection frame corresponding to the minimum distance as the frame target frame corresponding to the frame prediction frame.
7. The high altitude parabolic tracking method according to any one of claims 1 to 6, wherein the obtaining a current frame prediction frame of a previous frame target frame in the current frame image includes:
determining ID information of the previous frame target frame in the previous frame image;
acquiring a current frame prediction frame of a previous frame target frame in the current frame image, and transmitting the ID information to the current frame prediction frame;
correspondingly, after the frame target frame corresponding to the frame prediction frame is selected from the frame detection frame based on the area intersection value, the distance closeness value, the area ratio and the moving angle cosine value, the method further includes:
and transmitting the ID information corresponding to the frame prediction frame to the corresponding frame target frame.
8. A high altitude parabolic tracking apparatus, comprising:
the frame detection frame acquisition module is used for acquiring a frame detection frame of a high-altitude parabola in the frame image;
the current frame prediction frame acquisition module is used for acquiring a current frame prediction frame of a previous frame target frame in the current frame image; the previous frame target frame is a detection frame of a high-altitude parabola in the previous frame image;
the area intersection value calculation module is used for calculating the area intersection degree of the frame prediction frame and the frame detection frame to obtain an area intersection value;
the distance similarity value calculation module is used for calculating the distance similarity between the current frame prediction frame and the current frame detection frame to obtain a distance similarity value;
the area ratio calculation module is used for calculating the area ratio of a first intersection area of the tracking area corresponding to the current frame prediction frame and the current frame detection frame to the area of the current frame detection frame;
the moving angle cosine value calculation module is used for calculating the transverse moving degree of the frame prediction frame and the frame detection frame based on the tracking area to obtain a moving angle cosine value;
and the selection module is used for selecting the target frame of the current frame corresponding to the prediction frame of the current frame from the detection frame of the current frame based on the area intersection value, the distance similarity value, the area ratio and the cosine value of the movement angle.
9. An electronic device comprising a processor and a memory; wherein the processor, when executing the computer program stored in the memory, implements a high altitude parabolic tracking method as claimed in any one of claims 1 to 7.
10. A computer-readable storage medium for storing a computer program; wherein the computer program, when executed by a processor, implements the high altitude parabolic tracking method as claimed in any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211704001.5A CN115965654A (en) | 2022-12-29 | 2022-12-29 | High-altitude parabolic tracking method, device, equipment and medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211704001.5A CN115965654A (en) | 2022-12-29 | 2022-12-29 | High-altitude parabolic tracking method, device, equipment and medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115965654A true CN115965654A (en) | 2023-04-14 |
Family
ID=87353904
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211704001.5A Pending CN115965654A (en) | 2022-12-29 | 2022-12-29 | High-altitude parabolic tracking method, device, equipment and medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115965654A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116935074A (en) * | 2023-07-25 | 2023-10-24 | 苏州驾驶宝智能科技有限公司 | Multi-target tracking method and device based on adaptive association of depth affinity network |
-
2022
- 2022-12-29 CN CN202211704001.5A patent/CN115965654A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116935074A (en) * | 2023-07-25 | 2023-10-24 | 苏州驾驶宝智能科技有限公司 | Multi-target tracking method and device based on adaptive association of depth affinity network |
| CN116935074B (en) * | 2023-07-25 | 2024-03-26 | 苏州驾驶宝智能科技有限公司 | Multi-target tracking method and device based on adaptive association of depth affinity network |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US9947077B2 (en) | Video object tracking in traffic monitoring | |
| CN111709975B (en) | Multi-target tracking method, device, electronic equipment and storage medium | |
| CN113177968A (en) | Target tracking method and device, electronic equipment and storage medium | |
| CN109658433B (en) | Image background modeling and foreground extracting method and device and electronic equipment | |
| CN111105495A (en) | Laser radar mapping method and system fusing visual semantic information | |
| CN114677554A (en) | Statistical filtering infrared small target detection tracking method based on YOLOv5 and Deepsort | |
| CN111354022B (en) | Target Tracking Method and System Based on Kernel Correlation Filtering | |
| CN111476814B (en) | Target tracking method, device, equipment and storage medium | |
| CN115797736A (en) | Method, device, equipment and medium for training target detection model and target detection | |
| CN115965654A (en) | High-altitude parabolic tracking method, device, equipment and medium | |
| CN115719436A (en) | Model training method, target detection method, device, equipment and storage medium | |
| CN113269811A (en) | Data fusion method and device and electronic equipment | |
| CN112487861A (en) | Lane line recognition method and device, computing equipment and computer storage medium | |
| CN115049954A (en) | Target identification method, device, electronic equipment and medium | |
| CN113379789B (en) | Moving target tracking method in complex environment | |
| CN111578938B (en) | Target object positioning method and device | |
| CN111428567A (en) | Pedestrian tracking system and method based on affine multi-task regression | |
| CN116434181A (en) | Ground point detection method, device, electronic equipment and medium | |
| CN111274336B (en) | Target track processing method and device, storage medium and electronic device | |
| CN112630736B (en) | Parameter determination method, device, equipment and storage medium of road side radar | |
| CN115410153A (en) | Door opening and closing state judging method and device, electronic terminal and storage medium | |
| CN113052871A (en) | Target detection and automatic tracking algorithm based on intelligent selection strategy | |
| CN112102356B (en) | Target tracking method, device, terminal equipment and storage medium | |
| CN114694138B (en) | Road surface detection method, device and equipment applied to intelligent driving | |
| Xiao et al. | On the use of contextual information for robust colour-based particle filter tracking |
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 |