Classifier integration-based concrete mixer truck mixing tank steering detection and tracking method

A technology for concrete mixer trucks and mixing buckets, applied in the field of image processing, can solve problems such as low precision, unreliable detection and tracking, and low efficiency of real-time monitoring, and achieve the effect of improving tracking accuracy

Pending Publication Date: 2020-11-13
NANJING UNIV OF TECH
0 Cites 0 Cited by

AI-Extracted Technical Summary

Problems solved by technology

It is used to solve the problems of unreliable detection and tracking of the steering of the mixing drum of the concrete mixer truck by the traditional framework in the industry, low precision, and low efficiency of real-time monitoring. Detection and contour detection, track the feature center points extracted by th...
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Method used

Step 3: finally carry out image preprocessing operation, image preprocessing process can be according to commonly used image processing method, such as greyscale, methods such as binarization processing; Or use morpholo...
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Abstract

The invention designs a classifier integration-based concrete mixer truck mixing tank steering detection and tracking method. The invention relates to the fields of image processing, target detectionand target tracking. According to the method, an operation video stream of a stirring barrel is detected and tracked, according to inherent characteristics of the stirring barrel, three detection algorithms of straight line detection, angular point detection and contour detection are used for detecting characteristics of the stirring barrel, and the characteristics and characteristic center pointsof the stirring barrel are marked; displacement of a central point in a sequence image is used to judge the running direction of the stirring barrel, the tracking accuracy is difficult to guarantee;based on this, a prediction method is added. A relatively reliable tracking result is provided by calculating a predicted value, each detection algorithm can be regarded as a base classifier in tracking, an integration idea is cited, and the base classifier integrates the tracking result of the base classifier through a simple majority voting method or a weighted voting method, so that the tracking efficiency and precision of a traditional method are better improved.

Application Domain

Character and pattern recognition

Technology Topic

TruckVideo streaming +6

Image

  • Classifier integration-based concrete mixer truck mixing tank steering detection and tracking method
  • Classifier integration-based concrete mixer truck mixing tank steering detection and tracking method
  • Classifier integration-based concrete mixer truck mixing tank steering detection and tracking method

Examples

  • Experimental program(1)

Example Embodiment

[0008] 1. A detection and tracking method based on the steering of the mixing bucket of a concrete mixer truck integrated by a classifier, is characterized in that, the method comprises the following steps:
[0009] Step 1: Collect the video data set for the operation of the mixer drum of the concrete mixer, collect the video images of the mixer drum in operation, exclude the area outside the mixer drum as much as possible during the video collection process, filter out the data sources for detection and tracking, and then select the interested Region of interest, and perform image preprocessing operations;
[0010] Step 2: feature detection of the mixing tank, according to the data set collected in step 1, perform line detection, corner detection and contour detection on it, label the detected features, and extract the center point of the corresponding feature;
[0011] Step 3: Tracking the features of the mixing barrel, according to the feature center point obtained in step 2, and then by comparing the change of the center point of the feature position of the sequence image, the tracking is performed to obtain the result, and the turning of the mixing barrel is preliminarily judged;
[0012] Step 4: Predict the tracking result and use the classifier integration, give the initial image a predicted value, calculate the predicted value of each frame of image, judge the tracking result of the base classifier according to the predicted value, and then use the voting method to vote on the predicted tracking result. Voting to determine the final turning of the mixing barrel;
[0013] 2. The method for detecting and tracking the steering of the mixing drum of a concrete mixer truck based on the integration of the classifiers according to claim 1, wherein the method for acquiring the video data set of the operation of the mixing drum of the concrete mixer truck described in step 1 comprises the following steps: the following steps;
[0014] Step 1: First, use the existing tools to simulate the operation video of the mixing bucket of the concrete mixer truck, or collect the video of the construction site by yourself to extract the pictures by frame;
[0015] Step 2: Then manually set the area of ​​interest, or use the existing method to automatically select the area of ​​interest. The area of ​​interest requires the maximum feature area that includes the mixing tank, and at the same time excludes the part outside the feature area, so as to avoid the formation of image feature detection process. interference;
[0016] Step 3: Finally, perform the image preprocessing operation. The image preprocessing process can be based on common image processing methods, such as grayscale, binarization, etc.; or use morphological operation methods to make the image smoother and eliminate tiny noise in the image Such as mud spots stuck on the mixing bucket) will affect the detection process; or use deep learning related algorithms to improve the image quality.
[0017] 3. The method for detecting and tracking the steering of the mixing bucket of a concrete mixer truck based on the integration of the classifier according to claim 1, characterized in that the detection of the characteristics of the mixing bucket described in step 2 comprises the following steps:
[0018] Step 1: Extract and label the line segment features of the mixing bucket through the straight line detection algorithm. In order to accurately extract the line segment features of the mixing bucket, it is necessary to limit the angle of the detected line segment to maximize the angle that conforms to the real line segment. In addition, if you encounter line segments with similar positions or ductility trends, you need to fit them. Set the distance between the center points of the two line segments to be less than d, and perform a fitting operation on them. The position coordinates of the center point M (x m , y m ) is calculated as:
[0019]
[0020] where, (x i , y i )(i=1,2,...,N) represents the coordinates of the left endpoint of the detected line segment, (x j , y j ) (j=1, 2, ..., N) represents the coordinates of the right endpoint of the detected line segment, and then calculates the distance of the center point of the line segment to be fitted, and the distance calculation formula is as follows;
[0021]
[0022] where, (x i , y i )(i=1, 2,..., m) indicates the coordinates of the center point of the detected line segment i to be fitted, (x j , y j ) (j=1, 2, . . . , N) represents the detected coordinates of the center point of the line segment j to be fitted, and d represents the distance of the center point of the line segment to be fitted.
[0023] Step 2: Extract and label the corner features of the dense area of ​​the mixing bucket pattern through the corner detection algorithm. In order to avoid the interference of the interference corners, the clustering method is used to cluster the corners into clusters, and the distance from the cluster center is discarded. For far points, if there are many cluster centers, use the method of formula (1) to sum and average the center points;
[0024] Step 3: Use the contour detection algorithm to extract and mark part of the pattern outline of the mixing bucket. The algorithm may detect multiple contours of different sizes, and combine these small outlines that are part of the pattern outline into a pattern feature closest to the mixing bucket. The contour of , the calculation formula of the contour center point is as follows;
[0025]
[0026] where (x a , y a ), (x b , y b ), (x c , y c ) and (x d , y d ) respectively represent the position coordinates of the four corners of the contour, (x m , y m ) (m=1, 2, ..., N) represents the coordinate position of a single contour center point, and finally collects all the contour center point coordinates that meet the conditions. The calculation formula is as follows:
[0027]
[0028] 4. The method for detecting and tracking the steering of the mixing bucket of a concrete mixer truck based on the integration of the classifier according to claim 1, wherein the feature tracking of the mixing bucket described in step 3 comprises the following steps:
[0029] Step 1: Use the center point extracted by the detection algorithm to locate, take its image coordinate point, and compare the coordinates of the feature center point detected on the adjacent images of consecutive frames;
[0030] Step 2: The feature center point comparison takes the x-axis or y-axis as the standard, and calculates the feature center point p of the current frame c (x c , y c ) and the feature center point p of the previous frame c-1 (x c-1 , y c-1 ) coordinates on the x-axis or y-axis difference to track the feature, the formula is as follows:
[0031]
[0032] 5. The method for detecting and tracking the steering of the mixing bucket of a concrete mixer truck based on classifier integration according to claim 1, wherein the step 4 is to predict the tracking result and use the classifier integration to include the following steps:
[0033] Step 1: In the process of detecting and tracking a single feature, if the feature cannot be detected or the tracking result is wrong, the function of the prediction method is to set a prediction value for the tracking result of each frame, and judge the current value according to the prediction value of the previous turn. the turning of the frame;
[0034] The prediction method is calculated as:
[0035]
[0036]Among them, threshold represents the threshold (probability) for judging the rotation of the drum, the default value is 0.5, the initial prediction value of clockwise and counterclockwise is 0.5, p1 represents the clockwise prediction value of the previous frame, and p2 represents the anticlockwise prediction value of the previous frame. , if the tracking result before the prediction of the current frame is 1 (that is, clockwise), p1 increases with the probability of each k, and p2 decreases with the probability of each k, on the contrary, when the tracking result is 0 (that is, counterclockwise), p1 decreases with probability every k, p2 increases with probability every k. Generally, k takes the empirical value of 0.1. The final steering is determined by the results calculated by the prediction method;
[0037] Step 2: Each of the above detection methods can be regarded as a kind of classifier. Because the single classifier is used to judge that the steering of the mixing truck drum does not meet the requirements, the voting method is used to integrate the above three classifiers; the voting method Simple majority voting method or weighted voting method can be used, and the specific calculation method is as follows;
[0038] a) Set the base classifier h i (i=1, 2, 3), C j is the steering judgment mark, h i (x) is h i At steering judgment mark C j , the final result H(x) of the simple majority voting method can be expressed as:
[0039]
[0040] b) Let w i is h i (x) weight, usually w i ≥0, Then the final result H(x) of the weighted voting method can be expressed as:
[0041]
[0042] That is, the token with the most votes is predicted. If there are multiple tokens with the highest votes at the same time, one of them will be randomly selected. Vote on the tracking results through the voting method to obtain a more accurate judgment of the turning of the mixing barrel.
[0043] The invention includes straight line detection algorithm, corner detection algorithm, contour detection algorithm, clustering algorithm, morphological operation and other technologies.
[0044] The above algorithm is only a preferred implementation of the present invention, and is not intended to limit the present invention. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the present invention. within the scope of protection.
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

PUM

no PUM

Description & Claims & Application Information

We can also present the details of the Description, Claims and Application information to help users get a comprehensive understanding of the technical details of the patent, such as background art, summary of invention, brief description of drawings, description of embodiments, and other original content. On the other hand, users can also determine the specific scope of protection of the technology through the list of claims; as well as understand the changes in the life cycle of the technology with the presentation of the patent timeline. Login to view more.
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

Similar technology patents

Target state prediction method in dynamic environment

ActiveCN107248166ANarrow your searchImprove tracking accuracyImage enhancementImage analysisKalman filterState prediction
Owner:WUHAN INSTITUTE OF TECHNOLOGY

Solar energy photo-thermal high-temperature utilization system

InactiveCN101576321ASimple and reasonable structureImprove tracking accuracySolar heating energySolar heat collector controllersEngineeringSunlight
Owner:李捷

Device utilizing underground superficial layer soil temperature for tracking solar declination angles and adjusting method

InactiveCN105159328ASlow responseImprove tracking accuracyControl using feedbackEngineeringInternal pressure
Owner:ANHUI JOYFULL INFORMATION SCI & TECH

Path tracking anti-saturation robust control method for under-actuated water surface ship

ActiveCN112462773AImprove tracking accuracySustainable transportationAdaptive controlAerospace engineeringSaturation phenomenon
Owner:HARBIN ENG UNIV

Target positioning method, device and system

ActiveCN110133637AFast and precise positioningImprove tracking accuracyBeacon systems using electromagnetic wavesPosition fixationRadar measurementEnvironmental geology
Owner:CHANGCHUN INST OF OPTICS FINE MECHANICS & PHYSICS CHINESE ACAD OF SCI

Classification and recommendation of technical efficacy words

  • Improve tracking accuracy
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2023 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap