CN115861368A - Multi-class contour tracking method, device and medium for multi-class images - Google Patents
Multi-class contour tracking method, device and medium for multi-class images Download PDFInfo
- Publication number
- CN115861368A CN115861368A CN202211403758.0A CN202211403758A CN115861368A CN 115861368 A CN115861368 A CN 115861368A CN 202211403758 A CN202211403758 A CN 202211403758A CN 115861368 A CN115861368 A CN 115861368A
- Authority
- CN
- China
- Prior art keywords
- point
- contour
- pixel
- tracking
- starting point
- 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
Abstract
The invention discloses a multi-class contour tracking method, a multi-class contour tracking device and a multi-class contour tracking medium for multi-class images, and relates to the technical field of industrial production contour tracking. The method comprises a segmentation assignment process, a starting point acquisition process, a contour point tracking process and an execution termination process. According to the embodiment of the invention, the information of the inner contour and the outer contour of all different types of target objects in the image can be quickly and accurately obtained only by scanning the images of multiple types once, so that the relations of inclusion, overlapping and the like among different types of connected domains cannot be lost, the accuracy of the contour information can be ensured, meanwhile, the algorithm time and space complexity is low, and the method is suitable for low-computing-power equipment and is convenient to transplant to the existing low-power-consumption industrial equipment.
Description
Technical Field
The invention relates to the technical field of industrial production contour tracking, in particular to a multi-class contour tracking method, a multi-class contour tracking device and a multi-class contour tracking medium for multi-class images.
Background
Contour tracing, sometimes also referred to as boundary tracing, is one of the basic techniques of image processing. The method aims to obtain the contour information of a target object in an image, and geometric characteristics of the target object, such as angle, area, perimeter, curvature, center, eccentricity and projection, can be more accurately calculated according to the contour information. Contour tracing algorithms are the basis of image processing such as image compression, object shape representation, object recognition and contour-based region analysis, and have wide industrial applications.
With the continuous development of scientific technology, the size of an image processed by the image processing is larger and larger, the types of target objects in the image are extracted more and more, and the precision requirement is higher and higher. Image segmentation techniques based on deep learning in recent years have made it easier and more elaborate to segment images into different categories, with the wider application of image processing. There are several important features of industrial visual inspection application scenarios: the real-time performance requirement is high, the running stability requirement is high, the accuracy requirement is high, the equipment power consumption requirement is low, and the cost is sensitive. The real-time requirement here includes not only the requirement for the running speed of the algorithm itself, but also the requirement for low latency of data transmission, such as the transmission of image data needs to be completed within 1 ms. Due to the installation space limitation of an industrial production line, long-time uninterrupted work and cost sensitivity, the performance of the algorithm is a great challenge.
When the existing multi-class image tracking algorithm is applied to an industrial application scene, if parallel computing is used, delay in program operation is inevitably generated in information transmission of nodes, and the detection requirements of real-time performance and low delay in industrial application cannot be well met. Meanwhile, in the industrial application, the cost input is considered, the performance of the equipment is often greatly limited, and in most cases, the condition of parallel computing is not provided. However, the tracking mode of the existing algorithm has obvious defects on a serial machine: the computational complexity is redundant, and the memory overhead is large. Therefore, the contour tracing algorithm is applied to an industrial production line, and the existing algorithm, especially the contour tracing algorithm aiming at various images, is difficult to adapt to the requirements of online real-time and accurate detection.
Disclosure of Invention
The technical problem to be solved by the invention is to provide a multi-class contour tracking method, device and medium for multi-class images, which can quickly and accurately track complex contour information by only one-time scanning of the multi-class images, can distinguish contour types, can quickly and accurately track the contours of the multi-class images, and is easy to apply to low-cost equipment in industrial production.
In a first aspect, the present invention provides a multi-class contour tracking method for multi-class images, including: a segmentation assignment process, a starting point acquisition process, a contour point tracking process and a termination execution process;
the segmentation assignment process comprises: the method comprises the steps of carrying out image segmentation on an image to obtain a plurality of types of images, and dividing pixel points of the plurality of types of images into background pixels and at least one nth type pixel point, wherein n is an integer, and a circle of pixels on a frame of the plurality of types of images are constantly background pixels; assigning the background pixels as first set values, and assigning the nth type of pixel points as n +1 set values respectively;
the starting point obtaining process comprises the following steps: performing pixel-by-pixel scanning on the assigned multi-class images in the sequence from top to bottom and from left to right; when the outer contour starting point or the inner contour starting point is found by comparing the values of the left and right adjacent pixel points, the scanning is interrupted, and the outer contour starting point or the inner contour starting point is found to be used as a tracking starting point for contour tracking;
the contour point tracking process includes: detecting pixel points in the appointed direction around the starting point according to the appointed sequence, taking the found first same-class pixel point e as a tracking ending point of the outline, if the pixel point e meeting the condition cannot be found, indicating that the tracking starting point is an isolated point, restoring scanning from the pixel point on the right side after the operation of setting the pixel value of the tracking starting point as a negative class value is executed, and returning to the starting point acquisition process;
after finding the tracking end point, detecting 8 pixel points around the current contour point b one by one from the tracking start point based on the chain code according to a predefined sequence, and starting from the pixel point at the position of 0, sequentially detecting the pixel points in sequence until the similar pixel points of the b point are detected, thereby representing that the next contour point is found; then, taking the next contour point as the current contour point b to continuously and sequentially detect pixel points, setting the pixel values of all contour points as negative category values, and modifying the flag variable of the contour point with the pixel point on the right side of the contour point detected as a set value so that the flag variable cannot become a tracking starting point of other contours; performing circular execution, namely when the tracking starting point is tracked from the tracking ending point, stopping tracking of the current contour and obtaining a complete contour, then starting to recover scanning from the pixel point on the right side of the contour starting point, returning to the starting point obtaining process, and finding the tracking starting points of other contours by the same method and tracking the contours;
the termination of the execution process includes: and when the last pixel point of the multi-class images is scanned, ending the scanning to obtain the contour tracking result.
Further, in the starting point obtaining process, comparing values of two left and right adjacent pixel points to find an outer contour starting point or an inner contour starting point, specifically including:
judging two continuous pixel points p i,j And p i,j-1 Or p i,j And p i,j+1 Whether the value of (c) satisfies the following condition:
condition 1: t is a unit of i,j ≠T i,j-1 And f is i,j >0；
Condition 2: t is i,j ≠T i,j+1 And f is i,j >0；
Condition 3: t is i,j ≠T i,j+1 And p is i,j Is not flag-point;
wherein, T i,j For the current pixel point p i,j Is a non-negative value and does not change; f. of i,j Is p i,j A pixel value of (a); the flag-point indicates that the value of the flag variable is not an initial value;
when only the condition 1 is satisfied, the pixel point p i,j Is the starting point of the outer contour; when condition 2 or condition 3 is satisfied, the pixel point p i,j Is the starting point of the inner contour; when the condition 1 and the condition 2 are simultaneously satisfied, the pixel point p is processed i,j Only as the starting point of the outer contour.
Further, in the initial point obtaining process, when an inner contour initial point or an outer contour initial point is found, if the contour initial point is the outer contour initial point, the p-code of the pixel point is set to be 7, if the contour initial point is the inner contour initial point, the p-code of the pixel point is set to be 3, and the p-code is a current chain code and represents a direction from a previous contour point to a current contour point.
Further, in the contour point tracking process, pixel points on the right, right lower, left lower and left of the starting point are detected in sequence, and the found first same type pixel point e is used as a tracking termination point of the contour; when tracing to the starting point from a contour point, judging whether the contour point is the tracing ending point, if not, continuing the tracing of the current contour, if so, terminating the tracing of the current contour.
In a second aspect, the present invention provides a multi-class contour tracking apparatus for multi-class images, comprising: the device comprises a segmentation assignment module, a starting point acquisition module, a contour point tracking module and a termination execution module;
the segmentation assignment module is used for performing image segmentation on an image to obtain a plurality of types of images, and dividing pixel points of the plurality of types of images into background pixels and at least one nth type pixel point, wherein n is a real number, and a circle of pixels of a frame of the plurality of types of images are constant as the background pixels; assigning the background pixels as first set values, and assigning the nth type of pixel points as n +1 set values respectively;
the starting point acquisition module is used for scanning the assigned multi-class images pixel by pixel in the sequence from top to bottom and from left to right; when the outer contour starting point or the inner contour starting point is found by comparing the values of the left and right adjacent pixel points, the scanning is interrupted, and the outer contour starting point or the inner contour starting point is found to be used as a tracking starting point for contour tracking;
the contour point tracking module is used for detecting pixel points in the appointed direction around the initial point according to the appointed sequence, taking the found first same-kind pixel point e as the tracking end point of the contour, if the pixel point e meeting the condition cannot be found, the tracking initial point is an isolated point, and after the operation of setting the pixel value of the tracking initial point as a negative category value is executed, scanning is resumed from the pixel point on the right side;
after finding the tracking end point, detecting 8 pixel points around the current contour point b one by one from the tracking start point based on the chain code according to a predefined sequence, and starting from the pixel point at the position of 0, sequentially detecting the pixel points in sequence until the similar pixel points of the b point are detected, thereby representing that the next contour point is found; then, taking the next contour point as the current contour point b to continuously and sequentially detect pixel points, setting the pixel values of all contour points as negative category values, and modifying the flag variable of the contour point with the pixel point on the right side of the contour point detected as a set value so that the flag variable cannot become a tracking starting point of other contours; performing circular execution, namely when a tracking starting point is tracked from a tracking ending point, stopping tracking the current contour and obtaining a complete contour, then starting to resume scanning from a pixel point on the right side of the contour starting point, returning to a starting point acquisition module, and finding tracking starting points of other contours by using the same method and tracking the contour;
and the termination execution module is used for terminating scanning when the last pixel point of the multi-class images is scanned to obtain the contour tracking result.
Further, in the starting point obtaining module, comparing values of two left and right adjacent pixel points to find an outer contour starting point or an inner contour starting point, specifically including:
judging two continuous pixel points p i,j And p i,j-1 Or p i,j And p i,j+1 Whether the value of (c) satisfies the following condition:
condition 1: t is i,j ≠T i,j-1 And f is i,j >0；
Condition 2: t is i,j ≠T i,j+1 And f is i,j >0；
Condition 3: t is i,j ≠T i,j+1 And p is i,j Is not flag-point;
wherein, T i,j For the current pixel point p i,j Is a non-negative value and does not change; f. of i,j Is p i,j The pixel value of (a); the flag-point indicates that the value of the flag variable is not an initial value;
when only the condition 1 is satisfied, the pixel point p i,j Is the starting point of the outer contour; when condition 2 or condition 3 is satisfied, the pixel point p i,j Is the starting point of the inner contour; when the condition 1 and the condition 2 are simultaneously satisfied, the pixel point p is processed i,j Only as the starting point of the outer contour.
Further, in the starting point obtaining module, when an inner contour starting point or an outer contour starting point is found, if the contour starting point is the outer contour starting point, the p-code of the pixel point is set to be 7, if the contour starting point is the inner contour starting point, the p-code of the pixel point is set to be 3, and the p-code is a current chain code and represents a direction from a previous contour point to a current contour point.
Further, in the contour point tracking module, pixel points on the right, right lower, left lower and left of the starting point are sequentially detected, and the found first same-kind pixel point e is used as a tracking termination point of the contour; when tracing to the starting point from a contour point, judging whether the contour point is the tracing ending point, if not, continuing the tracing of the current contour, if so, terminating the tracing of the current contour.
In a third aspect, the invention provides a computer-readable storage medium having stored thereon a computer program which, when executed by a processor, performs the method of the first aspect.
The technical scheme provided by the embodiment of the invention at least has the following technical effects or advantages:
1. the information of the inner contour and the outer contour of all target objects of different categories in the image can be quickly and accurately obtained only by scanning the images of multiple categories once, so that the relations of inclusion, overlapping and the like among different categories of connected domains cannot be lost, the accuracy of the contour information can be guaranteed, meanwhile, the algorithm has low time and space complexity, is suitable for low-computing-power equipment, and is convenient to transplant to the existing low-power-consumption industrial equipment;
2. when one contour point is tracked, the pixel value of the point is set as a negative category value, so that the tracked contour points are distinguished, and the same contour is prevented from being repeatedly tracked;
3. the marking method through the flag variable guarantees that the algorithm does not lose any contour, does not repeatedly track the same contour, and can distinguish different categories.
The foregoing description is only an overview of the technical solutions of the present invention, and the embodiments of the present invention are described below in order to make the technical means of the present invention more clearly understood and to make the above and other objects, features, and advantages of the present invention more clearly understandable.
Drawings
The invention will be further described with reference to the following examples and figures.
FIG. 1 is a schematic diagram of the four-way communication and the eight-way communication in the communication relationship of the embodiment of the present invention;
FIG. 2 is a schematic diagram of pixels of a multi-class image according to an embodiment of the present invention;
FIG. 3 is a schematic diagram of an original image/binary image/multi-class image of a cable defect detection image;
FIG. 4 is a schematic view of a hole communication area and an inner contour according to an embodiment of the present invention;
FIG. 5 is a diagram illustrating a chain code in the prior art;
FIG. 6 is a schematic diagram of 8 different detection sequences for contour tracking according to an embodiment of the present invention;
FIG. 7 is a schematic view of the resulting pixel after the 1-connected domain outer contour of the pixel diagram of FIG. 2 has been completely traced;
FIG. 8 is a schematic flow chart of a method according to one embodiment of the present invention;
fig. 9 is a schematic structural diagram of a device according to a second embodiment of the present invention.
Detailed Description
The technical solution of the present invention will be described clearly and completely with reference to the accompanying drawings and the detailed description. 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.
In the description of the present invention, it should be noted that the terms "center", "upper", "lower", "left", "right", "inner", "outer", and the like indicate orientations or positional relationships based on those shown in the drawings, and are only for convenience of description and simplification of description, and should not be construed as limiting the present invention.
The technical scheme in the embodiment of the invention has the following general idea:
in order to solve the problem that the existing contour tracking algorithm is difficult to realize the online real-time and accurate detection of multiple types of images, the invention provides an efficient and accurate multiple types of contour tracking method for the multiple types of images. The multi-class contour tracking algorithm has low time and space complexity, is suitable for low-calculation-force equipment, and is convenient to transplant to the existing low-power-consumption industrial equipment.
Considering an image with size H multiplied by W, the pixel point is marked as pi ,j And (i, j) is the coordinate of a pixel, representing that the pixel is positioned at the ith row and the jth column in the image, and the position of the leftmost upper corner of the image is (1, 1). f. of i,j Representative pixel point p i,j The pixel value of (2). The communication relation involved in the present invention is two kinds of four communication and eight communication, as shown in fig. 1. Some of the others are defined below.
Definition 1 (multi-class images): each pixel point in the image has its unique attributive category, which is T i,j Is recorded as a pixel point p i,j Class value of 0. Ltoreq.T i,j And (5) less than or equal to 255, and the pixel points with the same category value are called like pixel points. The pixel point with the category value of 0 is called a background pixel, and the rest pixel points are called target object pixels. In an initial multi-class image, the pixel value of each pixel is equal to its class value. Fig. 2 depicts a pixel diagram of a multi-class image, where each square represents a pixel and the pixel with a pixel value of 0 represents the background. The remaining four are represented by four classes, 1,2,3,4 respectively. Contour tracing is to find the contours of all regions with the same number.
The multi-class images are images needing to be directly subjected to contour tracing, and are greatly different from binary images. In order to better display multiple types of images, when multiple types of images need to be displayed, pixels of different types are given different colors as distinctions, and the images are called as multiple-type image schematic diagrams. Fig. 3 shows a schematic diagram of camera artwork of a cable in industrial defect detection and its corresponding binary image and multi-class image. Fig. 3 (a) is an original image of a camera, fig. 3 (b) is a binary image corresponding to fig. 3 (a), and fig. 3 (c) is a schematic diagram of a plurality of types of images obtained by dividing the image of fig. 3 (a), each color representing one type.
Definition 2 (K-connected domain): on the premise of four (eight) communication, all the mutually communicated similar pixel points form a four (eight) communicated K-communication domain, wherein the value of K is the common class value of the pixel points forming the communication domain, and K is more than or equal to 0 and less than or equal to 255. Particularly, if a non-zero pixel point does not have any similar neighbor point, the pixel point is called an isolated point and is regarded as an independent K-connected domain. Taking fig. 2 as an example, if the pixel points with the category value of 1 in the graph are assumed to be in the eight-connected mode, a 1-connected domain is formed; if the four-way mode is used as a premise, two 1-way domains are formed. Therefore, it is specified that the background pixels are always connected in a four-way connection manner to form a background connection domain. For a given eight-connected nonzero K-connected domain S, the outermost pixels in the connected domain S are called outline pixels, and the set of all outline pixels in S forms the outline of S.
Definition 3 (surrounding): for a given two connected domains S 1 And S 2 In which S is 1 Is eight connected K 1 -connected component, S 2 Is four-way through K 2 -connected component, K 1 ≠K 2 ，K 1 Not equal to 0. First satisfies S 2 At least one pixel point and S 1 Has a common edge, and then satisfies the following S 2 Starting from any pixel point in the image, extending to the frame of the whole image by taking the four-connected direction as a path, wherein all the paths pass through S 1 The middle pixel point is called S 1 Surrounding S 2 。
Definition 4 (pore communicating region and inner contour): when an eight-connected nonzero K 1 -connected domain S 1 Surrounding a four-way K 2 -connected domain S 2 When S is present 2 And all others with S 2 Four-connected domain S with common edge 3 ，S 4 ，……，S n Form a hole region H 1 Then, H 1 And all others with H 1 Four-connected domain S with common edge n+1 ，……，S m Form a hole region H 2 By analogy, a hole area H which can not be expanded can be obtained finally t Otherwise, contradict with the surrounding relationship. Connected domain S in which all four are connected i All have class values different from S 1 I ≠ 1. At this time, scale H t Is S 1 One of the holes of (A) communicates with the domain, and S 1 The set of all the pixel points with the common edge of the hole communicating region forms S 1 An inner contour of (a). Fig. 4 shows the entire bore communication area and inner contour of the schematic of fig. 2. The shaded part in the figure represents a hole connected domain, and the pixel point with the negative pixel value represents an inner contour pixel point. Fig. 4 (a) and 4 (b) are two different inner contours of the 1-connectivity domain, and fig. 4 (c) is an inner contour of the 3-connectivity domain.
According to the definition, the contour pixel points form a closed contour in an eight-communication mode, and the background pixel points are not the contour pixel points. The multi-class contour tracking algorithm provided by the invention can quickly and accurately obtain the information of the inner contour and the outer contour of all target objects of different classes in an image after one-time scanning is carried out on one multi-class image. The information of the defects can be acquired more accurately in industrial visual inspection.
The multi-class contour tracking algorithm provided by the invention is mainly divided into three parts. The first part is how to determine the starting point for the tracking of a contour. In the subsequent description of the present invention, the concept of dots is equivalent to that of pixels. The second part is the tracking strategy, i.e. how to track from the current contour point to the next contour point. The last part is to determine the termination of the trackingAnd (5) conditioning, ending the current tracking and obtaining a complete contour. The scanning order of the image is from top to bottom, from left to right, pixel by pixel. When scanning to the pixel p i,j When the starting point of a certain contour is found, the scanning is interrupted, then contour tracking is started, when the tracking meets the termination condition, the current tracking is ended, and the pixel point p is selected i,j+1 The scan is resumed. The algorithm terminates when the scan reaches the lower right corner of the image. Without loss of generality, the invention assumes that a circle of pixels of a frame of a plurality of types of images to be processed are always background pixels.
By comparing the values of the left and right adjacent pixel points, a new contour can be found. In order to avoid repeatedly tracking the same contour, special marking methods are usually used to distinguish the tracked contour points. The invention provides a new marking method which comprises the following steps: when a contour point p x,y After being tracked, the pixel value of the point is set to a negative class value, i.e. f x,y ＝-T x,y . At the same time, give the contour point p x,y Setting an additional flag variable of the bol type with an initial value of 0, and setting a pixel point p on the right side of the contour point x,y+1 When the point is detected in the tracking process, the point is not possible to be the tracking starting point of other contours, so the contour point p is used x,y The flag variable of (1) is set to indicate differentiation. The contour point at which the flag variable has a value of 1 is referred to as "flag-point". This labeling method ensures that the algorithm does not lose any contour, does not repeatedly trace the same contour, and can distinguish between different classes.
In the scanning process, if two continuous pixel points p i,j And p i,j-1 Or p i,j And p i,j+1 Satisfies one of the following conditions, and then the pixel point p i,j Is the starting point for the tracking of a contour. Wherein f is i,j Is p i,j There will be a positive and negative change in the tracking process. T is i,j For the current pixel point p i,j Is a non-negative value and does not change.
Condition 1: t is i,j ≠T i,j-1 And f is i,j >0; then p is i,j Is the starting point of the outer contour.
Condition 2: t is i,j ≠T i,j+1 And f is i,j >0; then p is i,j Is the starting point of the inner contour.
Condition 3: t is i,j ≠T i,j+1 And p is i,j Is not flag-point; then p is i,j Is the starting point of the inner contour.
When the condition 1 and the condition 2 are simultaneously satisfied, the pixel point p is processed i,j Only as a starting point of the outer contour.
In the process of tracking from the current contour point to the next contour point, the tracker detects pixel points one by one according to a predefined sequence. The tracking algorithm of the present invention is based on chain codes, and the direction codes as shown in fig. 5 are chain codes. Since each pixel has only 8 neighbors at the most, it is sufficient to use 3 bits to indicate the direction code of the next contour pixel.
The detection sequence adopted in the embodiment of the present invention is shown in fig. 6, where each square represents a pixel, b is a current contour point, a is a previous contour point, and a current chain code p-code represents a direction from the previous contour point to the current contour point, and is marked as a → b, and its numerical value corresponds to the chain code shown in fig. 5. The number 0 → 6 (0 → 5) in fig. 6 represents the order in which the tracker detects the pixel points. When a new contour point is tracked each time, a corresponding sequence diagram is found according to the difference of p-codes, and the tracker starts from the pixel point at the position of 0 and sequentially detects the pixel points according to the sequence of 0 → 6 (0 → 5) until the similar pixel points of the b point are detected, which represents that the next contour point is found. The p-code is then updated. In addition, it has been proved that the pixel point for identifying the x position is not necessarily the same type of pixel point of the contour point b, otherwise, according to the counterclockwise detection sequence, for the previous contour point a, the current contour point should be the pixel point of the x position, not the b point. Since there are no previous contour points for the contour start point, it is specified that at the start of the outer contour start point and inner contour start point traces, the p-code sets default values of 7 and 3, respectively, to ensure that the contour traces can start properly.
The specific tracking strategy is as follows: whenever a new contour pixel b is detected i,j The pixel of the pointClass value with negative value, i.e. f i,j ＝-T i,j Then, according to the value of the current chain code p-code, the detection is performed according to the corresponding detection sequence in fig. 6 until the next contour point is found, and the p-code is updated. In this process, if pixel point p i,j+1 After detection, the current contour point b i,j The value of the flag variable of (1) is changed to 1. Note in particular that when the value of p-code is 2 or 3, since the current contour point b i,j Right side pixel point p i,j+1 The next contour point is not possible, so the current contour point b needs to be set at this time i,j The value of the flag variable of (1) is also changed to indicate that the pixel on the right side thereof has been detected. And the process is circulated until a complete contour is tracked.
The outline tracking process of the overall multi-class image is described in the following embodiments.
Example one
The present embodiment provides a multi-class contour tracking method for multi-class images, as shown in fig. 7, including: a segmentation assignment process, a starting point acquisition process, a contour point tracking process and a termination execution process;
the segmentation assignment process comprises: dividing an image to obtain a plurality of types of images, dividing pixel points of the plurality of types of images into background pixels and at least one nth type of pixel point, wherein n is an integer, and a circle of pixels of a frame of the plurality of types of images are constantly background pixels (a circle of background pixels can be added around the plurality of types of images, or a circle of pixels of the frame is directly set as the background pixels); assigning the background pixels to a first set value, and assigning the nth type of pixels to an n +1 th set value (for example, in fig. 2, there are four types of pixels except for the background pixels, the background pixels are assigned to 0, and the other 4 types of pixels are assigned to 1,2,3, and 4, respectively);
the starting point obtaining process comprises the following steps: scanning the assigned images in a pixel-by-pixel manner according to the sequence from top to bottom and from left to right (the scanning sequence is set for convenience of description of each pixel and the orientation of the pixels around the pixel in the subsequent method, and cannot be understood as the limitation of the invention; the scanning sequence can also be replaced by other sequences, such as from bottom to top, from right to left, or any sequence combination which can ensure that the pixels are scanned one by one in sequence, but the comparison mode, the positioning mode and the sequence of the corresponding points need to be correspondingly adjusted); when the outer contour starting point or the inner contour starting point is found by comparing the values of the left and right adjacent pixel points, the scanning is interrupted, and the outer contour starting point or the inner contour starting point is found to be used as a tracking starting point for contour tracking;
the contour point tracking process includes: detecting pixel points in the appointed direction around the initial point according to the appointed sequence, taking the found first same-kind pixel point e as a tracking end point of the outline, if the pixel point e meeting the condition cannot be found, indicating that the tracking initial point is an isolated point, restoring scanning from the pixel point on the right side after the operation of setting the pixel value of the tracking initial point as a negative category value is executed, and returning to the initial point acquisition process;
after finding the tracking end point, detecting 8 pixel points around the current contour point b one by one from the tracking start point based on the chain code according to a predefined sequence, and starting from the pixel point at the position of 0, sequentially detecting the pixel points in sequence until the similar pixel points of the b point are detected, thereby representing that the next contour point is found; then, taking the next contour point as a current contour point b to continuously and sequentially detect pixel points, setting the pixel values of all contour points as negative category values, and modifying the flag variable of the contour point detected by the pixel point on the right side of the contour point into a set value (if other pixel points are detected as the same type pixel points before the pixel point on the right side, the state of the pixel point on the right side is undetected), so that the next contour point cannot be a tracking starting point of other contours (only cannot be used as the tracking starting point, but can still be used as the contour point); performing circular execution, namely when the tracking starting point is tracked from the tracking ending point, stopping tracking of the current contour and obtaining a complete contour, then starting to recover scanning from the pixel point on the right side of the contour starting point, returning to the starting point obtaining process, and finding the tracking starting points of other contours by the same method and tracking the contours;
the termination of the execution process includes: and when the last pixel point of the multi-class images is scanned, ending the scanning to obtain the contour tracking result.
In a possible implementation manner, in the starting point obtaining process, comparing values of two left and right adjacent pixel points to find an outer contour starting point or an inner contour starting point, specifically including:
judging two continuous pixel points p i,j And p i,j-1 Or p i,j And p i,j+1 Whether the value of (c) satisfies the following condition:
condition 1: t is i,j ≠T i,j-1 And f is i,j >0；
Condition 2: t is i,j ≠T i,j+1 And f is i,j >0；
Condition 3: t is i,j ≠T i,j+1 And p is i,j Is not flag-point;
wherein, T i,j For the current pixel point p i,j Is a non-negative value and does not change; f. of i,j Is p i,j A pixel value of (a); flag-point indicates that the value of the flag variable is not an initial value;
when only the condition 1 is satisfied, the pixel point p i,j Is the starting point of the outer contour; when condition 2 or condition 3 is satisfied, the pixel point p i,j Is the starting point of the inner contour; when the condition 1 and the condition 2 are simultaneously satisfied, the pixel point p is processed i,j Only as the starting point of the outer contour.
In a possible implementation manner, in the starting point obtaining process, when an inner contour starting point or an outer contour starting point is found, if the contour starting point is the outer contour starting point, the p-code of the pixel point is set to be 7, and if the contour starting point is the inner contour starting point, the p-code of the pixel point is set to be 3, where the p-code is a current chain code and indicates a direction from a previous contour point to the current contour point.
In a possible implementation manner, in the contour point tracking process, pixels on the right side, the right lower side, the left lower side and the left side of the starting point are sequentially detected, and the found first same-kind pixel e is used as a tracking termination point of the contour; when tracing to the starting point from a contour point, judging whether the contour point is the tracing ending point, if not, continuing the tracing of the current contour, if so, terminating the tracing of the current contour.
In one embodiment, the starting point p for tracking when the scan finds a contour i,j When the scanning is suspended, and the right, lower left and lower left pixels of the starting point are sequentially detected, i.e. p is sequentially detected according to the sequence of 0,7,6,5 and 4 in the chain code direction in FIG. 5 i,j Surrounding pixel points, the first p found i,j The same type pixel point e of (2) is called the tracking end point of the contour. After recording the tracing end point of the contour, the tracker carries out contour point tracing one by one from the starting point until the tracker traces from the end point e to the starting point p i,j Then, the tracking of the current contour is stopped to obtain a complete contour, and then the pixel point p on the right side is used for tracking i,j+1 A resume scan is started. If the pixel point e meeting the condition can not be found, the tracking starting point is indicated as an isolated point, and f is carried out i,j ＝-T i,j After operation (c), pixel point p is displayed from the right i,j+1 A resume scan is started.
Taking the pixel schematic diagram shown in fig. 2 as an example, the pixel point p can be determined 2,4 Is a tracking starting point of the outer contour, pixel point e 2,5 It is its corresponding tracking termination point. After the first contour is completely traced by applying the contour tracing method of the present invention, the result is shown in fig. 8. In FIG. 8, the contour tracing is from p 2,4 Starting at the beginning and ending at e 2,5 Back to p 2,4 And obtaining a complete outline, namely obtaining the outline of the 1-connected domain. The pixel values of all the traced contour points in fig. 8 change to negative values, wherein the contour point circled by a circle represents the contour point flag-point.
Based on the same inventive concept, the application also provides a device corresponding to the method in the first embodiment, which is detailed in the second embodiment.
Example two
In the present embodiment, there is provided a multi-class contour tracing apparatus for a multi-class image, as shown in fig. 9, including: the device comprises a segmentation assignment module, a starting point acquisition module, a contour point tracking module and a termination execution module;
the segmentation assignment module is used for performing image segmentation on an image to obtain a plurality of types of images, and dividing pixel points of the plurality of types of images into background pixels and at least one nth type pixel point, wherein n is a real number, and a circle of pixels of a frame of the plurality of types of images are constant as the background pixels; assigning the background pixels as first set values, and assigning the nth type of pixel points as n +1 set values respectively;
the starting point acquisition module is used for scanning the assigned multi-class images pixel by pixel in the sequence from top to bottom and from left to right; when the outer contour starting point or the inner contour starting point is found by comparing the values of the left and right adjacent pixel points, the scanning is interrupted, and the outer contour starting point or the inner contour starting point is found to be used as a tracking starting point for contour tracking;
the contour point tracking module is used for detecting pixel points in the appointed direction around the initial point according to the appointed sequence, taking the found first same-kind pixel point e as the tracking end point of the contour, if the pixel point e meeting the condition cannot be found, the tracking initial point is an isolated point, and after the operation of setting the pixel value of the tracking initial point as a negative category value is executed, scanning is resumed from the pixel point on the right side;
after finding the tracking end point, detecting 8 pixel points around the current contour point b one by one from the tracking start point based on the chain code according to a predefined sequence, and starting from the pixel point at the position of 0, sequentially detecting the pixel points in sequence until the similar pixel points of the b point are detected, thereby representing that the next contour point is found; then, taking the next contour point as the current contour point b to continuously and sequentially detect pixel points, setting the pixel values of all contour points as negative category values, and modifying the flag variable of the contour point with the pixel point on the right side of the contour point detected as a set value so that the flag variable cannot become a tracking starting point of other contours; performing circular execution, namely when the tracking starting point is tracked from the tracking ending point, stopping tracking of the current contour and obtaining a complete contour, then starting to recover scanning from the pixel point on the right side of the contour starting point, returning to the starting point acquisition module, and finding the tracking starting points of other contours by the same method and tracking the contours;
and the termination execution module is used for terminating scanning when the last pixel point of the multi-class images is scanned to obtain the contour tracking result.
In a possible implementation manner, in the starting point obtaining module, comparing values of two left and right adjacent pixel points to find an outer contour starting point or an inner contour starting point, specifically including:
judging two continuous pixel points p i,j And p i,j-1 Or p i,j And p i,j+1 Whether the value of (c) satisfies the following condition:
condition 1: t is i,j ≠T i,j-1 And f is i,j >0；
Condition 2: t is i,j ≠T i,j+1 And f is i,j >0；
Condition 3: t is i,j ≠T i,j+1 And p is i,j Is not flag-point;
wherein, T i,j For the current pixel point p i,j Is a non-negative value and does not change; f. of i,j Is p i,j The pixel value of (a); the flag-point indicates that the value of the flag variable is not an initial value;
when only the condition 1 is satisfied, the pixel point p i,j Is the starting point of the outer contour; when condition 2 or condition 3 is satisfied, the pixel point p i,j Is the starting point of the inner contour; when the condition 1 and the condition 2 are simultaneously satisfied, the pixel point p is processed i,j Only as the starting point of the outer contour.
In a possible implementation manner, when the inner contour starting point or the outer contour starting point is found in the starting point obtaining module, if the contour starting point is the outer contour starting point, the p-code of the pixel point is set to be 7, and if the contour starting point is the inner contour starting point, the p-code of the pixel point is set to be 3, where the p-code is a current chain code and indicates a direction from a previous contour point to the current contour point.
In a possible implementation manner, in the contour point tracking module, pixels on the right side, the right lower side, the left lower side and the left side of the starting point are sequentially detected, and the found first same-kind pixel e is used as a tracking termination point of the contour; when tracing to the starting point from a contour point, judging whether the contour point is the tracing ending point, if not, continuing the tracing of the current contour, if so, terminating the tracing of the current contour.
Since the apparatus described in the second embodiment of the present invention is an apparatus used for implementing the method of the first embodiment of the present invention, based on the method described in the first embodiment of the present invention, a person skilled in the art can understand the specific structure and the deformation of the apparatus, and thus the details are not described herein. All the devices adopted in the method of the first embodiment of the present invention belong to the protection scope of the present invention.
Based on the same inventive concept, the application provides a storage medium corresponding to the third embodiment.
EXAMPLE III
The present embodiment provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, any one of the first embodiment can be implemented.
Since the computer-readable storage medium described in this embodiment is a computer-readable storage medium used for implementing the method in the first embodiment of the present application, a person skilled in the art can understand a specific implementation manner of the computer-readable storage medium and various modifications thereof based on the method described in the first embodiment of the present application, and therefore, how to implement the method in the embodiment of the present application by using the computer-readable storage medium is not described in detail herein. Computer-readable storage media that can be used by those skilled in the art to implement the methods of the embodiments of the present application are all within the scope of the present application.
The invention focuses on the field of visual detection of products manufactured intelligently by industry, and the field has the particularity that: the real-time performance requirement is high, the operation stability requirement is high, the accuracy requirement is high, the time delay is low, the equipment power consumption requirement is low, and the cost is sensitive. The invention provides a novel multi-class contour tracking algorithm, which can directly track the contours of multi-class images and can accurately obtain the contours of all classes only by scanning the images once.
The multi-class contour tracking algorithm is also applicable to processing binary images, even is superior to some traditional single-class tracking algorithms in efficiency, and the efficiency is remarkably improved when the multi-class images are processed. The algorithm can reach the running speed of a single digit millisecond level when processing small-size multi-class images, the running speed can be kept below fifty milliseconds when processing large-size multi-class images, and the algorithm can be expected to meet the real-time requirement in industrial application after optimization on the premise that codes are not optimized. In addition, the multi-class algorithm has low time and space complexity, is suitable for low-computation-power equipment, is convenient to transplant to the existing low-power-consumption industrial equipment, and well meets the requirement of industrial visual detection application on the performance of the algorithm due to the characteristics.
According to the embodiment of the invention, the information of the inner contour and the outer contour of all different types of target objects in the image can be quickly and accurately obtained only by scanning the images of multiple types once, so that the relations of inclusion, overlapping and the like among different types of connected domains cannot be lost, the accuracy of the contour information can be ensured, meanwhile, the algorithm time and space complexity is low, and the method is suitable for low-computing-power equipment and is convenient to transplant to the existing low-power-consumption industrial equipment; when one contour point is tracked, the pixel value of the point is set as a negative category value, so that the tracked contour points are distinguished, and the same contour is prevented from being repeatedly tracked; the marking method through the flag variable guarantees that the algorithm does not lose any contour, does not repeatedly track the same contour, and can distinguish different categories.
As will be appreciated by one skilled in the art, embodiments of the present invention may be provided as a method, system, or computer program product. Accordingly, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
The present invention is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
Although specific embodiments of the invention have been described above, it will be understood by those skilled in the art that the specific embodiments described are illustrative only and are not limiting upon the scope of the invention, and that equivalent modifications and variations can be made by those skilled in the art without departing from the spirit of the invention, which is to be limited only by the appended claims.
Claims (9)
1. A multi-class contour tracking method for multi-class images is characterized by comprising the following steps: a segmentation assignment process, a starting point acquisition process, a contour point tracking process and a termination execution process;
the segmentation assignment process comprises: the method comprises the steps of carrying out image segmentation on an image to obtain a plurality of types of images, and dividing pixel points of the plurality of types of images into background pixels and at least one nth type pixel point, wherein n is an integer, and a circle of pixels on a frame of the plurality of types of images are constantly background pixels; assigning the background pixels as first set values, and assigning the nth type of pixel points as n +1 set values respectively;
the starting point obtaining process comprises the following steps: performing pixel-by-pixel scanning on the assigned multi-class images in the sequence from top to bottom and from left to right; when the outer contour starting point or the inner contour starting point is found by comparing the values of the left and right adjacent pixel points, the scanning is interrupted, and the outer contour starting point or the inner contour starting point is found to be used as a tracking starting point for contour tracking;
the contour point tracking process includes: detecting pixel points in the appointed direction around the initial point according to the appointed sequence, taking the found first same-kind pixel point e as a tracking end point of the outline, if the pixel point e meeting the condition cannot be found, indicating that the tracking initial point is an isolated point, restoring scanning from the pixel point on the right side after the operation of setting the pixel value of the tracking initial point as a negative category value is executed, and returning to the initial point acquisition process;
after finding the tracking end point, detecting 8 pixel points around the current contour point b one by one from the tracking start point based on the chain code according to a predefined sequence, and starting from the pixel point at the position of 0, sequentially detecting the pixel points in sequence until the similar pixel points of the b point are detected, thereby representing that the next contour point is found; then, taking the next contour point as the current contour point b to continuously and sequentially detect pixel points, setting the pixel values of all contour points as negative category values, and modifying the flag variable of the contour point with the pixel point on the right side of the contour point detected as a set value so that the flag variable cannot become a tracking starting point of other contours; performing circular execution, namely when the tracking starting point is tracked from the tracking ending point, stopping tracking of the current contour and obtaining a complete contour, then starting to recover scanning from the pixel point on the right side of the contour starting point, returning to the starting point obtaining process, and finding the tracking starting points of other contours by the same method and tracking the contours;
the termination of the execution process includes: and when the last pixel point of the multi-class images is scanned, ending the scanning to obtain the contour tracking result.
2. The method of claim 1, wherein: in the starting point obtaining process, comparing values of two adjacent pixel points on the left and right to find an outer contour starting point or an inner contour starting point, specifically comprising:
judging two continuous pixel points p i,j And p i,j-1 Or p i,j And p i,j+1 Whether the value of (c) satisfies the following condition:
condition 1: t is i,j ≠T i,j-1 And f is i,j >0；
Condition 2: t is i,j ≠T i,j+1 And f is i,j >0；
Condition 3: t is i,j ≠T i,j+1 And p is i,j Is not flag-point;
wherein, T i,j For the current pixel point p i,j Is a non-negative value and does not change; f. of i,j Is p i,j A pixel value of (a); the flag-point indicates that the value of the flag variable is not an initial value;
when only the condition 1 is satisfied, the pixel point p i,j Is the starting point of the outer contour; when condition 2 or condition 3 is satisfied, the pixel point p i,j Is the starting point of the inner contour; when the condition 1 and the condition 2 are simultaneously satisfied, the pixel point p is processed i,j Only as the starting point of the outer contour.
3. The method according to claim 1 or 2, characterized in that: in the initial point obtaining process, when an inner contour initial point or an outer contour initial point is found, if the contour initial point is the outer contour initial point, the p-code of the pixel point is set to be 7, if the contour initial point is the inner contour initial point, the p-code of the pixel point is set to be 3, and the p-code is a current chain code and represents the direction from a previous contour point to the current contour point.
4. The method of claim 1, wherein: in the process of tracing the contour point, sequentially detecting pixel points on the right, right lower, left lower and left of the starting point, and taking the found first same type pixel point e as a tracing termination point of the contour; when tracing to the starting point from a contour point, judging whether the contour point is the tracing ending point, if not, continuing the tracing of the current contour, if so, terminating the tracing of the current contour.
5. A multi-class contour tracking apparatus for a multi-class image, comprising: the device comprises a segmentation assignment module, a starting point acquisition module, a contour point tracking module and a termination execution module;
the segmentation assignment module is used for performing image segmentation on an image to obtain a plurality of types of images, and dividing pixel points of the plurality of types of images into background pixels and at least one nth type pixel point, wherein n is a real number, and a circle of pixels of a frame of the plurality of types of images are constant as the background pixels; assigning the background pixels as first set values, and assigning the nth type of pixel points as n +1 set values respectively;
the starting point acquisition module is used for scanning the multiple classes of images subjected to value assignment pixel by pixel in the sequence from top to bottom and from left to right; when the outer contour starting point or the inner contour starting point is found by comparing the values of the left and right adjacent pixel points, the scanning is interrupted, and the outer contour starting point or the inner contour starting point is found to be used as a tracking starting point for contour tracking;
the contour point tracking module is used for detecting pixel points in the appointed direction around the initial point according to the appointed sequence, taking the found first same-kind pixel point e as the tracking end point of the contour, if the pixel point e meeting the condition cannot be found, the tracking initial point is an isolated point, and after the operation of setting the pixel value of the tracking initial point as a negative category value is executed, scanning is resumed from the pixel point on the right side;
after finding the tracking termination point, detecting 8 pixel points around the current contour point b one by one from the tracking starting point based on the chain code according to a predefined sequence, starting from the pixel point at the position of 0, sequentially detecting the pixel points in sequence until the similar pixel points of the b point are detected, and then representing that the next contour point is found; then, taking the next contour point as the current contour point b to continuously and sequentially detect pixel points, setting the pixel values of all contour points as negative category values, and modifying the flag variable of the contour point with the pixel point on the right side of the contour point detected as a set value so that the flag variable cannot become a tracking starting point of other contours; performing circular execution, namely when a tracking starting point is tracked from a tracking ending point, stopping tracking the current contour and obtaining a complete contour, then starting to resume scanning from a pixel point on the right side of the contour starting point, returning to a starting point acquisition module, and finding tracking starting points of other contours by using the same method and tracking the contour;
and the termination execution module is used for terminating scanning when the last pixel point of the multi-class images is scanned to obtain the contour tracking result.
6. The apparatus of claim 5, wherein: in the starting point obtaining module, comparing values of two adjacent pixels on the left and right to find an outer contour starting point or an inner contour starting point, specifically including:
judging two continuous pixel points p i,j And p i,j-1 Or p i,j And p i,j+1 Whether the value of (c) satisfies the following condition:
condition 1: t is a unit of i,j ≠T i,j-1 And f is i,j >0；
Condition 2: t is i,j ≠T i,j+1 And f is i,j >0；
Condition 3: t is i,j ≠T i,j+1 And p is i,j Is not flag-point;
wherein, T i,j Is the current pixelPoint p i,j Is a non-negative value and does not change; f. of i,j Is p i,j A pixel value of (a); the flag-point indicates that the value of the flag variable is not an initial value;
when only the condition 1 is satisfied, the pixel point p i,j Is the starting point of the outer contour; when condition 2 or condition 3 is satisfied, the pixel point p i,j Is the starting point of the inner contour; when the condition 1 and the condition 2 are simultaneously satisfied, the pixel point p is processed i,j Only as the starting point of the outer contour.
7. The apparatus of claim 5 or 6, wherein: in the initial point obtaining module, when an inner contour initial point or an outer contour initial point is found, if the contour initial point is the outer contour initial point, the p-code of the pixel point is set to be 7, if the contour initial point is the inner contour initial point, the p-code of the pixel point is set to be 3, and the p-code is a current chain code and represents the direction from a previous contour point to the current contour point.
8. The apparatus of claim 5, wherein in the contour point tracking module, pixels to the right, at the bottom left, and at the left of the starting point are detected in sequence, and the found first homogeneous pixel e is taken as the tracking end point of the contour; when tracing to the starting point from a contour point, judging whether the contour point is the tracing ending point, if not, continuing the tracing of the current contour, if so, terminating the tracing of the current contour.
9. A computer-readable storage medium, on which a computer program is stored which, when being executed by a processor, carries out the method according to any one of claims 1 to 4.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211403758.0A CN115861368A (en) | 2022-11-10 | 2022-11-10 | Multi-class contour tracking method, device and medium for multi-class images |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211403758.0A CN115861368A (en) | 2022-11-10 | 2022-11-10 | Multi-class contour tracking method, device and medium for multi-class images |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115861368A true CN115861368A (en) | 2023-03-28 |
Family
ID=85662991
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211403758.0A Pending CN115861368A (en) | 2022-11-10 | 2022-11-10 | Multi-class contour tracking method, device and medium for multi-class images |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115861368A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117274364A (en) * | 2023-11-17 | 2023-12-22 | 深圳中科精工科技有限公司 | OpenCV-based area calculation method, equipment and storage medium |
-
2022
- 2022-11-10 CN CN202211403758.0A patent/CN115861368A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117274364A (en) * | 2023-11-17 | 2023-12-22 | 深圳中科精工科技有限公司 | OpenCV-based area calculation method, equipment and storage medium |
| CN117274364B (en) * | 2023-11-17 | 2024-03-22 | 深圳中科精工科技有限公司 | OpenCV-based area calculation method, equipment and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Romero-Ramirez et al. | Speeded up detection of squared fiducial markers | |
| Chen et al. | Augmented ship tracking under occlusion conditions from maritime surveillance videos | |
| CN106097361B (en) | Defect area detection method and device | |
| US10410354B1 (en) | Method and apparatus for multi-model primitive fitting based on deep geometric boundary and instance aware segmentation | |
| CN111209978B (en) | Three-dimensional visual repositioning method and device, computing equipment and storage medium | |
| JP7316731B2 (en) | Systems and methods for detecting and classifying patterns in images in vision systems | |
| CN114240939B (en) | Method, system, equipment and medium for detecting appearance defects of mainboard components | |
| CN110909712B (en) | Moving object detection method and device, electronic equipment and storage medium | |
| CN112336342A (en) | Hand key point detection method and device and terminal equipment | |
| CN115861368A (en) | Multi-class contour tracking method, device and medium for multi-class images | |
| CN107527368A (en) | Three-dimensional attitude localization method and device based on Quick Response Code | |
| CN111950523A (en) | Ship detection optimization method and device based on aerial photography, electronic equipment and medium | |
| CN115330757A (en) | Circuit board welding spot defect detection method and system | |
| CN111680680B (en) | Target code positioning method and device, electronic equipment and storage medium | |
| Singh et al. | Performance analysis of object detection algorithms for robotic welding applications in planar environment | |
| Wang et al. | Assembly defect detection of atomizers based on machine vision | |
| CN111429437A (en) | Image non-reference definition quality detection method for target detection | |
| CN108268813B (en) | Lane departure early warning method and device and electronic equipment | |
| Mingzhu et al. | A new method of circle’s center and radius detection in image processing | |
| CN111819567A (en) | Method and apparatus for matching images using semantic features | |
| Somani et al. | Scene perception and recognition in industrial environments for human-robot interaction | |
| CN113793371A (en) | Target segmentation tracking method and device, electronic equipment and storage medium | |
| US11881016B2 (en) | Method and system for processing an image and performing instance segmentation using affinity graphs | |
| CN110705479A (en) | Model training method, target recognition method, device, equipment and medium | |
| Noroozi et al. | Towards Optimal Defect Detection in Assembled Printed Circuit Boards Under Adverse Conditions |
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 |