WO2023134193A1 - 一种基于红外技术的高速线材盘卷形状与位置检测方法 - Google Patents

一种基于红外技术的高速线材盘卷形状与位置检测方法 Download PDF

Info

Publication number
WO2023134193A1
WO2023134193A1 PCT/CN2022/119357 CN2022119357W WO2023134193A1 WO 2023134193 A1 WO2023134193 A1 WO 2023134193A1 CN 2022119357 W CN2022119357 W CN 2022119357W WO 2023134193 A1 WO2023134193 A1 WO 2023134193A1
Authority
WO
WIPO (PCT)
Prior art keywords
curve
wire coil
point
concave
image
Prior art date
Application number
PCT/CN2022/119357
Other languages
English (en)
French (fr)
Inventor
陈磊
夏奇
王哲
Original Assignee
江苏省沙钢钢铁研究院有限公司
江苏沙钢高科信息技术有限公司
江苏沙钢集团有限公司
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by 江苏省沙钢钢铁研究院有限公司, 江苏沙钢高科信息技术有限公司, 江苏沙钢集团有限公司 filed Critical 江苏省沙钢钢铁研究院有限公司
Publication of WO2023134193A1 publication Critical patent/WO2023134193A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01B—MEASURING LENGTH, THICKNESS OR SIMILAR LINEAR DIMENSIONS; MEASURING ANGLES; MEASURING AREAS; MEASURING IRREGULARITIES OF SURFACES OR CONTOURS
    • G01B11/00—Measuring arrangements characterised by the use of optical techniques
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01B—MEASURING LENGTH, THICKNESS OR SIMILAR LINEAR DIMENSIONS; MEASURING ANGLES; MEASURING AREAS; MEASURING IRREGULARITIES OF SURFACES OR CONTOURS
    • G01B11/00—Measuring arrangements characterised by the use of optical techniques
    • G01B11/24—Measuring arrangements characterised by the use of optical techniques for measuring contours or curvatures
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/0002—Inspection of images, e.g. flaw detection
    • G06T7/0004—Industrial image inspection
    • G06T7/0008—Industrial image inspection checking presence/absence
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/10—Segmentation; Edge detection
    • G06T7/12—Edge-based segmentation
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/10—Segmentation; Edge detection
    • G06T7/136—Segmentation; Edge detection involving thresholding
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/60—Analysis of geometric attributes
    • G06T7/62—Analysis of geometric attributes of area, perimeter, diameter or volume
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/60—Analysis of geometric attributes
    • G06T7/64—Analysis of geometric attributes of convexity or concavity
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/70—Determining position or orientation of objects or cameras
    • G06T7/73—Determining position or orientation of objects or cameras using feature-based methods
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/10—Image acquisition modality
    • G06T2207/10048—Infrared image
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/20—Special algorithmic details
    • G06T2207/20024—Filtering details
    • G06T2207/20032—Median filtering

Definitions

  • the invention relates to the technical field of image recognition, in particular to a high-speed wire coil shape and position detection method based on infrared technology.
  • the laying machine when the wire passes through the high-speed spinning pipe, it is subjected to the normal pressure of the pipe wall, the sliding friction force, the thrust of the finishing mill and the pinch roller, and the centrifugal force of itself.
  • the shape of the silk tube is gradually bent and deformed, gradually bent from the linear motion, and reaches the required curvature at the exit of the spinning tube, forming a helical coil, and the loop is uniformly and smoothly spit out. Since there are many factors affecting the normal operation of the laying machine on site, the following failures are prone to occur in the laying machine: (1)
  • the spinning circle is not round.
  • Tail flick When the tail of the wire is extruded by the laying machine, the coils are disordered, the coils are not round, and the spacing of the coils is not equal. The amplitude of the tail flick becomes more and more serious with the increase of the rolling speed. In severe cases, the tail becomes a bullet and flies out. (3) The spinning coil swings left and right.
  • the invention with the application number CN202110569887.6 discloses a steel coil end surface image segmentation method, device and electronic equipment
  • the invention with the application number CN202110871963.9 discloses an artificial intelligence-based automatic detection method for steel coil unwinding and Device
  • the invention with application number CN202110932645.9 discloses an artificial intelligence-based aluminum coil length prediction method and system
  • the invention with application number CN202011615504.6 discloses a circle detection and fitting method and device , electronic equipment and storage media, the former obtains a large number of steel coil data sets and multiple scans to obtain the edge points of the steel coil for analysis and calculation, while the latter three files use complicated algorithms to improve the recognition accuracy.
  • the method used above not only calculates the data Huge and takes a long time to identify.
  • the present invention provides a high-speed wire coil shape and position detection method capable of systematically judging the abnormality of the wire and reducing labor costs and error rates.
  • the present invention adopts the following technical solutions:
  • the embodiment of the present invention proposes a high-speed wire coil shape and position detection method based on infrared technology, which is characterized in that the detection method includes the following steps:
  • step S2 using a standard median filter method to perform noise reduction processing on the grayscale image in step S1;
  • the concave point refers to the pixel point whose curvature mutation value on the edge contour line is greater than the maximum allowable curvature change threshold ;
  • step S4 according to the concave points obtained in step S3, the edge contour is divided into several curve segments, paired according to the position of the curve segments to generate several curve segment combinations, and the two curve segments in each curve segment combination belong to the same Layer wire coils;
  • the detection method also includes the following steps:
  • step S3 the process of obtaining the edge profile of the coil by using the two-way threshold segmentation method includes the following sub-steps:
  • step S3 the process of obtaining the pits includes the following sub-steps:
  • step S36 rotate the contour edge curve on the left side clockwise along the midpoint of the visual field by 1°, judge again whether the concave point to be estimated calculated in step S35 satisfies any one of the above inequalities after the rotation, and if so, count and add 1, otherwise the count is 0;
  • step S37 repeating step S36 until a complete rotation of 360° is completed, accumulating the number of times the concave point to be estimated satisfies the above inequalities, and if it exceeds the set times threshold T e , record the concave point to be estimated as a concave point;
  • step S38 repeat step S35 to step S37, until all the concave points on the left contour edge curve are analyzed, arrange in order to obtain the concave point sequence L c1 , L c2 ,...,L cn of the left contour edge curve, and then The same method is used to analyze and obtain all concave points on the right contour edge curve, and arrange them sequentially to obtain the concave point sequence R c1 , R c2 , . . . , R cn of the right contour edge curve.
  • step S4 the process of pairing includes the following steps:
  • Step S44 is repeated until all left curve segments are analyzed and matched.
  • the embodiment of the present invention proposes a high-speed wire coil shape and position detector based on infrared technology.
  • the device includes: an image acquisition system, a data transmission system, and a data processing system;
  • the image acquisition system includes an infrared thermal imaging camera installed directly above the wire coil conveying roller table, the infrared thermal imaging camera has a field of view wider than the wire coil conveying roller table width, and is used to obtain the coils on the coil conveying roller table raw material image of ;
  • the data processing system includes an industrial control computer and a display;
  • the industrial control computer is used to calculate and analyze the original material image collected by the image acquisition system by using the detection method described in any one of claims 1-5, to obtain the current original material The shape and position of all coils in the image;
  • the display is used to synchronously display the original material image and the analysis results output by the industrial control computer;
  • the data transmission system includes optical fiber transceivers respectively arranged in the field control box and the indoor control box, and optical fibers and network cables for connecting the image acquisition system and the data processing system.
  • the data processing system includes an alarm module, and the alarm module is connected to an industrial control computer through an I/O interface card;
  • the industrial control computer judges the shape and position of all the coils in the current original material image, and if there is a sudden change in the equivalent diameter or central position, an alarm signal is sent to the alarm module, which makes the alarm module issue an audible and visual alarm.
  • the digital image processing technology proposed by the present invention captures the image of the coiled coil shape and position formed by the laying machine, analyzes and extracts the outline and concave points, and obtains the shape and position of all coils through the method of curve matching. Remind the production line operators in time for sudden changes and other situations, so as to detect the production failure of the laying machine in time, reduce the impact of the problem of the laying machine on the quality of the wire product, and reduce the occurrence of manual misjudgments, and at the same time reduce labor costs.
  • FIG. 1 is a schematic diagram of the flow structure of a high-speed wire coil shape and position detection method using infrared technology according to an embodiment of the present invention.
  • Fig. 2 is a structural schematic diagram of a high-speed wire coil shape and position detection method using infrared technology according to an embodiment of the present invention.
  • Fig. 3 is a schematic diagram of bidirectional threshold segmentation according to an embodiment of the present invention.
  • Fig. 4 is a schematic diagram of a curve segment matching method according to an embodiment of the present invention.
  • FIG. 1 is a schematic diagram of the flow structure of a high-speed wire coil shape and position detection method using infrared technology according to an embodiment of the present invention.
  • Fig. 2 is a structural schematic diagram of a high-speed wire coil shape and position detection method using infrared technology according to an embodiment of the present invention.
  • This embodiment proposes a high-speed wire coil shape and position detection method based on infrared technology.
  • the detection method includes the following steps:
  • step S2 using a standard median filter method to perform noise reduction processing on the grayscale image in step S1.
  • the concave point refers to the pixel point whose curvature mutation value on the edge contour line is greater than the maximum allowable curvature change threshold .
  • the edge contour is divided into several curve segments, paired according to the position of the curve segments to generate several curve segment combinations, and the two curve segments in each curve segment combination belong to the same Layer wire coils.
  • the detection method of the present embodiment also includes the following steps:
  • the original material image on the wire coil transportation roller table is obtained, and the original material image is converted into a corresponding grayscale image.
  • the grayscale image acquired above is denoised using the standard median filter method.
  • this embodiment adopts an infrared thermal imaging camera, and a lens is installed on each camera.
  • the infrared thermal imaging camera is installed directly above the middle of the wire coil conveying roller table, the vertical distance from the conveying roller table is 2.0m, the acquisition frequency of the imaging equipment is 3 frames/s, and the field of view width of the infrared thermal imaging camera needs to be larger than that of the wire coil Coil transport roller width, through this infrared thermal imaging camera to obtain the original material image on the wire coil transport roller.
  • a set of air-cooled protective cover is installed on the outside of the camera and lens.
  • the air-cooled protective cover has an IP56 protection level, and the internal interlayer is filled with compressed air. While the cooling protection equipment is operating normally, a wind curtain is formed at the front of the lens. A hole is provided at the center of the lower end surface of the protective cover to facilitate image collection by the camera and the lens.
  • Fig. 3 is a schematic diagram of bidirectional threshold segmentation according to an embodiment of the present invention.
  • the process of obtaining the edge profile of the coil by using the two-way threshold segmentation method includes the following sub-steps:
  • the acquisition process of pits in step S3 includes the following sub-steps:
  • step S36 rotate the contour edge curve on the left side clockwise along the midpoint of the visual field by 1°, judge again whether the concave point to be estimated calculated in step S35 satisfies any one of the above inequalities after the rotation, and if so, count and add 1, otherwise the count is 0.
  • step S37 repeating step S36 until the complete rotation of 360°, accumulating the number of times the pit to be estimated satisfies the above inequalities, and if it exceeds the set number threshold T e , record the pit to be estimated as a pit.
  • step S38 repeat step S35 to step S37, until all the concave points on the left contour edge curve are analyzed, arrange in order to obtain the concave point sequence L c1 , L c2 ,...,L cn of the left contour edge curve, and then The same method is used to analyze and obtain all concave points on the right contour edge curve, and arrange them sequentially to obtain the concave point sequence R c1 , R c2 , . . . , R cn of the right contour edge curve.
  • the picture is scanned line by line.
  • the scan is performed from the left side of the field of view to the right side of the field of view. If the pixel gray value of a certain point A is higher than the threshold T gray , then A is considered The left boundary point of the outer contour of the line; scan from the right side of the field of view to the left side of the field of view, if the pixel gray value of a certain point B is higher than the threshold T gray , then B is considered to be the right side boundary point of the line's outer contour.
  • the left and right outer contours of the wire coil picture can be obtained.
  • step 3 look for concave points through the above-mentioned scanned pictures.
  • Each layer of wire coil can be approximated as an ellipse, so the intersection of two elliptical arcs shows a concave curve and an obvious abrupt change in curvature, which is called a concave point.
  • step 3 processing two curves composed of discrete pixel points are obtained.
  • the left curve be (L 0 ,L 1 ,...,L n-1 ), and the right curve be (R 0 ,R 1 ,..., R n-1 ), the left and right processing logics are similar, taking the left curve as an example, assuming that the current detection point is L i (xi , y i ), its predecessor and successor points are L pre (x pre , y pre ) and L next (x next , y next ), respectively, then the concave point on the curve must be the local extremum point of x or y, that is, at least satisfy the following four inequalities one:
  • the curve on the left is rotated 1° clockwise along the midpoint of the visual field, and it is judged again whether the corresponding point of the original concave point still satisfies at least one of the above four inequalities after rotation, and if so, add 1 to the count. After one round (360°) of rotation is completed, the number of times that each concave point satisfies at least one of the above four inequalities is accumulated, and if it exceeds the set threshold T e , it is recorded as a concave point.
  • Fig. 4 is a schematic diagram of a curve segment matching method according to an embodiment of the present invention.
  • the process of pairing includes the following steps:
  • Step S44 is repeated until all left curve segments are analyzed and matched.
  • the least squares fitting algorithm is used to fit the equivalent ellipse after the paired curve segments are combined, and the center point position and major axis length of the equivalent ellipse are obtained as the center position and diameter of the wire coil of this layer, and further Get the shape and position of all coils in the current picture. If there is a sudden change in equivalent diameter or center position, an alarm will be triggered to inform the operator to check and adjust the laying machine configuration.
  • An embodiment of the present application provides a high-speed wire coil shape and position detector based on infrared technology.
  • the device includes: an image acquisition system, a data transmission system, and a data processing system.
  • the image acquisition system in this embodiment includes an infrared thermal imaging camera installed directly above the coil conveying roller table. Raw material image of the coil.
  • the data processing system includes an industrial control computer and a display; the industrial control computer is used to calculate and analyze the original material image collected by the image acquisition system by using the detection method described in any one of claims 1-5, and obtain all discs in the current original material image.
  • the shape and position of the roll; the display is used to synchronously display the original material image and the analysis results output by the industrial control computer.
  • the specific implementation process can refer to the method embodiment provided in the first aspect above, and will not be described in detail here.
  • the data transmission system includes optical fiber transceivers respectively arranged in the field control box and the indoor control box, and optical fibers and network cables for connecting the image acquisition system and the data processing system.
  • the data processing system in this embodiment includes an alarm module, which is connected to an industrial control computer through an I/O interface card.
  • the industrial control computer judges the shape and position of all coils in the current original material image, and if there is a sudden change in equivalent diameter or central position, an alarm signal is sent to the alarm module, which makes the alarm module issue an audible and visual alarm.

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Theoretical Computer Science (AREA)
  • Geometry (AREA)
  • Quality & Reliability (AREA)
  • Image Analysis (AREA)

Abstract

本发明公开了一种基于红外技术的高速线材盘卷形状与位置检测方法，通过获取线材盘卷运输辊道上的原始物料图像并转换为相应的灰度图像；采用标准中值滤波法对灰度图像进行降噪处理；再采用双向阈值分割法获取盘卷的边缘轮廓与凹点；采用最小二乘拟合算法拟合得到相应的等效椭圆，并且将计算得到的等效椭圆的中心点位置和长轴长度作为该曲线段组合对应的线材盘卷层的中心位置和直径，推算得到当前原始物料图像中所有盘卷的形状与位置情况。本发明提出的通过分析提取轮廓、凹点，并通过曲线配对的方法获取所有盘卷的形状与位置情况，对突变等情况及时提醒产线操作工进行相应处理，降低了产品质量问题。

Description

一种基于红外技术的高速线材盘卷形状与位置检测方法 技术领域
本发明涉及图像识别技术领域，具体而言涉及一种基于红外技术的高速线材盘卷形状与位置检测方法。
背景技术
目前吐丝机工作时，线材通过高速旋转的吐丝管时，受到吐丝管管壁的正压力、滑动摩擦力、精轧机和夹送辊的推力、自身的离心力的作用下，随着吐丝管的形状逐渐弯曲变形，由直线运动逐渐弯曲，并在吐丝管出口达到所要求的曲率，形成螺旋线圈,均匀平稳的成圈吐出。由于现场影响吐丝机正常工作的因素较多，吐丝机容易发生如下故障：(1)吐圈不圆。吐丝机开始吐圈后，部分圈形不好或吐丝的圆度不够，导致空冷轨道和集卷站经常发生堆钢、卡钢影响生产正常进行。(2)甩尾。线材尾部经吐丝机吐出时出现的线圈乱、成圈不圆，线圈排列间距不等，甩尾幅度随轧制速度提高而越趋严重，严重时尾部成子弹头飞出。(3)吐丝线圈左右摆动。在吐丝的过程中，出现吐出线圈左右摆动的现象，在一定程度上增加集卷难度，并严重影响盘卷包装质量。(4)吐丝圈出现大小圈。吐丝状态不稳定出现吐圈大小圈交替。(5)吐丝时向单面倾斜。吐丝时向一个方向倾斜，造成线圈单面摩擦,线圈变形。
申请号为CN202110569887.6的发明中公开了一种钢卷端面图像分割方法、装置及电子设备，申请号为CN202110871963.9的发明中公开了一种基于人工智能的钢卷松卷自动检测方法和装置，申请号为CN202110932645.9的发明中公开了一种一种基于人工智能的铝卷长度预测方法及系统，申请号为CN202011615504.6的发明中公开了一种圆检测及拟合方法、装置、电子设备和存储介质，前者通过获取大量的钢卷数据集以及多次扫描获取钢卷边缘点进行分析计算，而后3个文件则通过繁复算法以提高识别准确度，上述使用的方法不仅计算数据庞大而且识别时间较长。
当前现场判断吐丝机运行情况依然依靠操作工主观经验判断，并没有形成一套客观的判断依据，不仅存在人工成本较高，并且专业技能人员较少，而且由于人工进行判断将存在及时性较差、准确度较低的问题，导致吐丝机问题对线材产品质量造成的影响。
发明内容
本发明针对现有技术中的不足，提供一种能够系统化判断线材异常情况，降低人工成本与出错率的高速线材盘卷形状与位置检测方法。
为实现上述目的，本发明采用以下技术方案：
第一方面，本发明实施例提出了一种基于红外技术的高速线材盘卷形状与位置检测方法，其特征在于，所述检测方法包括以下步骤：
S1，获取线材盘卷运输辊道上的原始物料图像，将原始物料图像转换为相应的灰度图像；
S2，采用标准中值滤波法对步骤S1中的灰度图像进行降噪处理；
S3，对降噪处理后的灰度图像逐行扫描，采用双向阈值分割法获取盘卷的边缘轮廓与凹点；凹点是指边缘轮廓线上曲率突变值大于最大允许曲率变化阈值的像素点；
S4，根据步骤S3中得到的凹点，将边缘轮廓分割成若干段曲线，按照曲线段的位置进行两两配对，生成若干个曲线段组合，每个曲线段组合中的两个曲线段属于同一层线材盘卷；
S5，针对每个曲线段组合，采用最小二乘拟合算法拟合得到相应的等效椭圆，并且将计算得到的等效椭圆的中心点位置和长轴长度作为该曲线段组合对应的线材盘卷层的中心位置和直径，推算得到当前原始物料图像中所有盘卷的形状与位置情况。
进一步地，所述检测方法还包括以下步骤：
S6，对当前原始物料图像中所有盘卷的形状与位置情况进行判断，若出现等效直径或中心位置突变，触发警报。
进一步地，步骤S3中，采用双向阈值分割法获取盘卷的边缘轮廓的过程包括以下子步骤：
S31，从视野左侧向视野右侧进行逐行扫描，获取所有左侧边界点；其中，当灰度图像中其中一行的其中一个像素点的像素灰度值高于阈值T gray，则认为该像素点是该行外轮廓的左侧边界点；
S32，反向从视野右侧向视野左侧进行扫描，获取所有右侧边界点；其中，当灰度图像中其中一行的其中一个像素点的像素灰度值高于阈值T gray，则认为该像素点是该行外轮廓的右侧边界点；
S33，组合所有左侧边界点，构成左侧轮廓边缘曲线；组合所有右侧边界点，构成右侧轮廓边缘曲线。
进一步地，步骤S3中，所述凹点的获取过程包括以下子步骤：
S34，设左侧轮廓边缘曲线为(L 0,L 1,...，L n-1)，右侧轮廓边缘曲线为(R 0,R 1,...，R n-1)；
S35，从左侧轮廓边缘曲线上选取其中一个检测点L i(x i,y i)，设当前检测点L i(x i,y i)的前继点和后继点分别为L pre(x pre,y pre)和L next(x next,y next)，根据下述不等式组计算得到x i或y i的局部极值点，作为左侧轮廓边缘曲线上的其中一个待估凹点：
Figure PCTCN2022119357-appb-000001
S36，将左侧轮廓边缘曲线沿着视野中点顺时针旋转1°，再次判断旋转后对步骤S35中计算得到的待估凹点是否满足上述不等式组中的任意一个，若满足，则计数加1，反之则计数为0；
S37，重复步骤S36，直至完成旋转一周360°，累计该待估凹点满足上述不等式组的次数，若超过设定次数阈值T e，则将该待估凹点记为凹点；
S38，重复步骤S35至步骤S37，直至分析得到左侧轮廓边缘曲线上所有的凹点，按序排列得到左侧轮廓边缘曲线的凹点序列L c1,L c2,...,L cn，再采用同样的方法分析得到右侧轮廓边缘曲线上所有的凹点，按序排列得到右侧轮廓边缘曲线的凹点序列R c1,R c2,...,R cn。
进一步地，步骤S4中，所述配对的过程包括以下步骤：
S41，按序排列得到左侧轮廓边缘曲线的凹点序列L c1,L c2,...,L cn，按序排列得到右侧轮廓边缘曲线的凹点序列R c1,R c2,...,R cn，其中，左侧连续凹点连接成的曲线段为L c1L c2，L c2L c3，…，L c(n-1)L cn，右侧连续凹点连接成的曲线段为R c1R c2，R c2R c3，…，R c(n-1)R cn；
S42，计算每个左侧曲线段的中点坐标和每个右侧曲线段的中点坐标；
S43，选择其中一个左侧曲线段L ciL c(i+1)的中点坐标(x li,y li)，依次求出右侧曲线段的中点坐标中与中点坐标(x li,y li)在y轴方向上差值最小的中点坐标(x ri,y ri)，若|y li-y ri|小于设定阈值T，则认为左侧曲线段L ciL c(i+1)与右侧曲线段R cjR c(j+1)为一组，若未找到满足条件的右侧曲线段，则该左侧曲线段不计入统计范围；
S44，重复步骤S44，直至分析匹配完成所有左侧曲线段。
第二方面，本发明实施例提出了一种基于红外技术的高速线材盘卷形状与位置的检测仪，所述装置包括：图像采集系统、数据传输系统和数据处理系统；
所述图像采集系统包括安装在线材盘卷运输辊道正上方的红外热成像相机，该红外热成像相机的视野宽度大于线材盘卷运输辊道宽度，用于获取盘卷运输辊道上的盘卷的原始物料图像；
所述数据处理系统包括工业控制计算机和显示器；所述工业控制计算机用于采用权利要求1-5任一项中所述检测方法对图像采集系统采集的原始物料图像进行计算分析，得到当前原始物料图像中所有盘卷的形状与位置情况；所述显示器用于同步显示原始物料图像和工业控制计算机输出的分析结果；
所述数据传输系统包括分别设置在现场控制箱和室内控制箱内的光纤收发器和用于连接图像采集系统和数据处理系统的光纤及网线。
进一步地，所述数据处理系统包括报警模组，报警模组通过I/O接口卡与工业控制计算机连接；
所述工业控制计算机对当前原始物料图像中所有盘卷的形状与位置情况进行判断，若出 现等效直径或中心位置突变，发送警报信号至报警模组，使报警模组发出声光警报。
本发明的有益效果是：
本发明提出的通过数字图像处理技术，对吐丝机形成的盘卷线圈形状及位置进行图像抓取，分析提取轮廓、凹点，并通过曲线配对的方法获取所有盘卷的形状与位置情况，对突变等情况及时提醒产线操作工进行相应处理，从而及时发现吐丝机生产故障，降低了由于吐丝机问题对线材产品质量造成的影响，并减少人工误判等情况的出现，同时降低了人工成本。
附图说明
图1是本发明实施例的红外技术的高速线材盘卷形状与位置检测方法的流程结构示意图。
图2是本发明实施例的红外技术的高速线材盘卷形状与位置检测方法的结构示意图。
图3是本发明实施例的双向阈值分割示意图。
图4是本发明实施例的曲线段匹配方法示意图。
具体实施方式
现在结合附图对本发明作进一步详细的说明。
需要注意的是，发明中所引用的如“上”、“下”、“左”、“右”、“前”、“后”等的用语，亦仅为便于叙述的明了，而非用以限定本发明可实施的范围，其相对关系的改变或调整，在无实质变更技术内容下，当亦视为本发明可实施的范畴。
实施例一
图1是本发明实施例的红外技术的高速线材盘卷形状与位置检测方法的流程结构示意图。图2是本发明实施例的红外技术的高速线材盘卷形状与位置检测方法的结构示意图。本实施例提出了一种基于红外技术的高速线材盘卷形状与位置检测方法，该检测方法包括以下步骤：
S1，获取线材盘卷运输辊道上的原始物料图像，将原始物料图像转换为相应的灰度图像。
S2，采用标准中值滤波法对步骤S1中的灰度图像进行降噪处理。
S3，对降噪处理后的灰度图像逐行扫描，采用双向阈值分割法获取盘卷的边缘轮廓与凹点；凹点是指边缘轮廓线上曲率突变值大于最大允许曲率变化阈值的像素点。
S4，根据步骤S3中得到的凹点，将边缘轮廓分割成若干段曲线，按照曲线段的位置进行两两配对，生成若干个曲线段组合，每个曲线段组合中的两个曲线段属于同一层线材盘卷。
S5，针对每个曲线段组合，采用最小二乘拟合算法拟合得到相应的等效椭圆，并且将计算得到的等效椭圆的中心点位置和长轴长度作为该曲线段组合对应的线材盘卷层的中心位置和直径，推算得到当前原始物料图像中所有盘卷的形状与位置情况。
优选的，本实施例的检测方法还包括以下步骤：
S6，对当前原始物料图像中所有盘卷的形状与位置情况进行判断，若出现等效直径或中心位置突变，触发警报。
一、原始图像处理
首先获取线材盘卷运输辊道上的原始物料图像，将原始物料图像转换为相应的灰度图像。
然后，采用标准中值滤波法对上面获取的灰度图像进行降噪处理。
例如，本实施例采用一个红外热成像相机，每台相机上安装有一个镜头。红外热成像相机安装在线材盘卷运输辊道中间正上方，与运输辊道的垂直距离为2.0m，成像设备的采集频率为3帧/s，另外红外热成像相机的视野宽度需大于线材盘卷运输辊道宽度，通过此红外热成像相机获取线材盘卷运输辊道上的原始物料图像。
此外，由于成像设备安装在靠近吐丝机的位置，环境温度较高，因此在相机及镜头外侧均装有一套风冷防护罩，风冷防护罩具备IP56防护等级，内部夹层通入压缩空气，在降温保护设备正常运行的同时，在镜头前端形成一道风帘。在防护罩下端面的中心处开设孔位，便于相机及镜头采集图像。
二、轮廓、凹点提取
图3是本发明实施例的双向阈值分割示意图。在步骤S3中，采用双向阈值分割法获取盘卷的边缘轮廓的过程包括以下子步骤：
S31，从视野左侧向视野右侧进行逐行扫描，获取所有左侧边界点；其中，当灰度图像中其中一行的其中一个像素点的像素灰度值高于阈值T gray，则认为该像素点是该行外轮廓的左侧边界点。
S32，反向从视野右侧向视野左侧进行扫描，获取所有右侧边界点；其中，当灰度图像中其中一行的其中一个像素点的像素灰度值高于阈值T gray，则认为该像素点是该行外轮廓的右侧边界点。
S33，组合所有左侧边界点，构成左侧轮廓边缘曲线；组合所有右侧边界点，构成右侧轮廓边缘曲线。
优选的，在步骤S3中凹点的获取过程包括以下子步骤：
S34，设左侧轮廓边缘曲线为(L 0,L 1,...，L n-1)，右侧轮廓边缘曲线为(R 0,R 1,...，R n-1)。
S35，从左侧轮廓边缘曲线上选取其中一个检测点L i(x i,y i)，设当前检测点L i(x i,y i)的前继点和后继点分别为L pre(x pre,y pre)和L next(x next,y next)，根据下述不等式组计算得到x i或y i的局部极值点，作为左侧轮廓边缘曲线上的其中一个待估凹点：
Figure PCTCN2022119357-appb-000002
S36，将左侧轮廓边缘曲线沿着视野中点顺时针旋转1°，再次判断旋转后对步骤S35中计算得到的待估凹点是否满足上述不等式组中的任意一个，若满足，则计数加1，反之则计数为0。
S37，重复步骤S36，直至完成旋转一周360°，累计该待估凹点满足上述不等式组的次数，若超过设定次数阈值T e，则将该待估凹点记为凹点。
S38，重复步骤S35至步骤S37，直至分析得到左侧轮廓边缘曲线上所有的凹点，按序排列得到左侧轮廓边缘曲线的凹点序列L c1,L c2,...,L cn，再采用同样的方法分析得到右侧轮廓边缘曲线上所有的凹点，按序排列得到右侧轮廓边缘曲线的凹点序列R c1,R c2,...,R cn。
例如，本实施例中对图片进行逐行扫描，对于图片中的某一行，从视野左侧向视野右侧进行扫描，若某一点A的像素灰度值高于阈值T gray，则认为A是该行外轮廓的左侧边界点；从视野右侧向视野左侧进行扫描，若某一点B的像素灰度值高于阈值T gray，则认为B是该行外轮廓的右侧边界点。
逐行扫描完成后，即可得到本次线材盘卷图片的左右两条外侧轮廓。
然后通过上述扫描过的图片寻找凹点，每层线材盘卷都可近似为一个椭圆形，因此两段椭圆弧相交处表现为曲线凹陷和明显的曲率突变，称为凹点。经过步骤三处理后得到的是两段由离散像素点集合成的曲线，设左侧曲线为(L 0,L 1,...，L n-1)，右侧曲线为(R 0,R 1,...，R n-1)，左侧和右侧处理逻辑类似，以左侧曲线为例，假设当前检测点为L i(x i,y i)，其前继点和后继点分别为L pre(x pre,y pre)和L next(x next,y next)，则曲线上的凹点必须是x或y的局部极值点，即需至少满足下列4个不等式组中的一个：
Figure PCTCN2022119357-appb-000003
左侧曲线沿着视野中点顺时针旋转1°，再次判断旋转后原凹点的对应点是否依然至少满足上述4个不等式组中的一个，若满足，则计数加1。旋转完成一周(360°)后，累计各凹点至少满足上述4个不等式组中一个的次数，若超过设定阈值T e，则记为凹点。
三、配对
图4是本发明实施例的曲线段匹配方法示意图。步骤S4中，所述配对的过程包括以下步骤：
S41，按序排列得到左侧轮廓边缘曲线的凹点序列L c1,L c2,...,L cn，按序排列得到右侧轮廓边缘曲线的凹点序列R c1,R c2,...,R cn，其中，左侧连续凹点连接成的曲线段为L c1L c2，L c2L c3，…，L c(n-1)L cn，右侧连续凹点连接成的曲线段为R c1R c2，R c2R c3，…，R c(n-1)R cn。
S42，计算每个左侧曲线段的中点坐标和每个右侧曲线段的中点坐标。
S43，选择其中一个左侧曲线段L ciL c(i+1)的中点坐标(x li,y li)，依次求出右侧曲线段的中点坐标中与中点坐标(x li,y li)在y轴方向上差值最小的中点坐标(x ri,y ri)，若|y li-y ri|小于设定阈值T，则认为左侧曲线段L ciL c(i+1)与右侧曲线段R cjR c(j+1)为一组，若未找到满足条件的右侧曲线段，则该左侧曲线段不计入统计范围。
S44，重复步骤S44，直至分析匹配完成所有左侧曲线段。
本实施例在配对后的曲线段组合，采用最小二乘拟合算法拟合出等效椭圆，得到等效椭圆的中心点位置和长轴长度作为该层线材盘卷的中心位置和直径，进一步得到当前图片中所有盘卷的形状与位置情况，若出现等效直径或中心位置突变，则触发警报，告知操作工对吐丝机配置进行检查调整。
实施例二
本申请实施例提供了一种基于红外技术的高速线材盘卷形状与位置的检测仪，该装置包括：图像采集系统、数据传输系统和数据处理系统。
本实施例的图像采集系统包括安装在线材盘卷运输辊道正上方的红外热成像相机，该红外热成像相机的视野宽度大于线材盘卷运输辊道宽度，用于获取盘卷运输辊道上的盘卷的原始物料图像。
数据处理系统包括工业控制计算机和显示器；工业控制计算机用于采用权利要求1-5任一项中所述检测方法对图像采集系统采集的原始物料图像进行计算分析，得到当前原始物料图像中所有盘卷的形状与位置情况；显示器用于同步显示原始物料图像和工业控制计算机输出的分析结果，具体的实施过程可以参照上述第一方面提供的方法实施例，此处不做详细阐述说明。
数据传输系统包括分别设置在现场控制箱和室内控制箱内的光纤收发器和用于连接图像采集系统和数据处理系统的光纤及网线。
优选的本实施例的数据处理系统包括报警模组，报警模组通过I/O接口卡与工业控制计算机连接。
其中，工业控制计算机对当前原始物料图像中所有盘卷的形状与位置情况进行判断，若出现等效直径或中心位置突变，发送警报信号至报警模组，使报警模组发出声光警报。
以上仅是本发明的优选实施方式，本发明的保护范围并不仅局限于上述实施例，凡属于本发明思路下的技术方案均属于本发明的保护范围。应当指出，对于本技术领域的普通技术人员来说，在不脱离本发明原理前提下的若干改进和润饰，应视为本发明的保护范围。

Claims (7)

  1. 一种基于红外技术的高速线材盘卷形状与位置检测方法，其特征在于，所述检测方法包括以下步骤：
    S1，获取线材盘卷运输辊道上的原始物料图像，将原始物料图像转换为相应的灰度图像；
    S2，采用标准中值滤波法对步骤S1中的灰度图像进行降噪处理；
    S3，对降噪处理后的灰度图像逐行扫描，采用双向阈值分割法获取盘卷的边缘轮廓与凹点；凹点是指边缘轮廓线上曲率突变值大于最大允许曲率变化阈值的像素点；
    S4，根据步骤S3中得到的凹点，将边缘轮廓分割成若干段曲线，按照曲线段的位置进行两两配对，生成若干个曲线段组合，每个曲线段组合中的两个曲线段属于同一层线材盘卷；
    S5，针对每个曲线段组合，采用最小二乘拟合算法拟合得到相应的等效椭圆，并且将计算得到的等效椭圆的中心点位置和长轴长度作为该曲线段组合对应的线材盘卷层的中心位置和直径，推算得到当前原始物料图像中所有盘卷的形状与位置情况。
  2. 根据权利要求1所述的基于红外技术的高速线材盘卷形状与位置检测方法，其特征在于，所述检测方法还包括以下步骤：
    S6，对当前原始物料图像中所有盘卷的形状与位置情况进行判断，若出现等效直径或中心位置突变，触发警报。
  3. 根据权利要求1所述的基于红外技术的高速线材盘卷形状与位置检测方法，其特征在于，步骤S3中，采用双向阈值分割法获取盘卷的边缘轮廓的过程包括以下子步骤：
    S31，从视野左侧向视野右侧进行逐行扫描，获取所有左侧边界点；其中，当灰度图像中其中一行的其中一个像素点的像素灰度值高于阈值T gray，则认为该像素点是该行外轮廓的左侧边界点；
    S32，反向从视野右侧向视野左侧进行扫描，获取所有右侧边界点；其中，当灰度图像中其中一行的其中一个像素点的像素灰度值高于阈值T gray，则认为该像素点是该行外轮廓的右侧边界点；
    S33，组合所有左侧边界点，构成左侧轮廓边缘曲线；组合所有右侧边界点，构成右侧轮廓边缘曲线。
  4. 根据权利要求3所述的基于红外技术的高速线材盘卷形状与位置检测方法，其特征在于，步骤S3中，所述凹点的获取过程包括以下子步骤：
    S34，设左侧轮廓边缘曲线为(L 0,L 1,...，L n-1)，右侧轮廓边缘曲线为(R 0,R 1,...，R n-1)；
    S35，从左侧轮廓边缘曲线上选取其中一个检测点L i(x i,y i)，设当前检测点L i(x i,y i)的前继点和后继点分别为L pre(x pre,y pre)和L next(x next,y next)，根据下述不等式组计算得到x i或y i的局部极值点，作为左侧轮廓边缘曲线上的其中一个待估凹点：
    Figure PCTCN2022119357-appb-100001
    S36，将左侧轮廓边缘曲线沿着视野中点顺时针旋转1°，再次判断旋转后对步骤S35中计算得到的待估凹点是否满足上述不等式组中的任意一个，若满足，则计数加1，反之则计数为0；
    S37，重复步骤S36，直至完成旋转一周360°，累计该待估凹点满足上述不等式组的次数，若超过设定次数阈值T e，则将该待估凹点记为凹点；
    S38，重复步骤S35至步骤S37，直至分析得到左侧轮廓边缘曲线上所有的凹点，按序排列得到左侧轮廓边缘曲线的凹点序列L c1,L c2,...,L cn，再采用同样的方法分析得到右侧轮廓边缘曲线上所有的凹点，按序排列得到右侧轮廓边缘曲线的凹点序列R c1,R c2,...,R cn。
  5. 根据权利要求4所述的基于红外技术的高速线材盘卷形状与位置检测方法，其特征在于，步骤S4中，所述配对的过程包括以下步骤：
    S41，按序排列得到左侧轮廓边缘曲线的凹点序列L c1,L c2,...,L cn，按序排列得到右侧轮廓边缘曲线的凹点序列R c1,R c2,...,R cn，其中，左侧连续凹点连接成的曲线段为L c1L c2，L c2L c3，…，L c(n-1)L cn，右侧连续凹点连接成的曲线段为R c1R c2，R c2R c3，…，R c(n-1)R cn；
    S42，计算每个左侧曲线段的中点坐标和每个右侧曲线段的中点坐标；
    S43，选择其中一个左侧曲线段L ciL c(i+1)的中点坐标(x li,y li)，依次求出右侧曲线段的中点坐标中与中点坐标(x li,y li)在y轴方向上差值最小的中点坐标(x ri,y ri)，若|y li-y ri|小于设定阈值T，则认为左侧曲线段L ciL c(i+1)与右侧曲线段R cjR c(j+1)为一组，若未找到满足条件的右侧曲线段，则该左侧曲线段不计入统计范围；
    S44，重复步骤S44，直至分析匹配完成所有左侧曲线段。
  6. 一种基于红外技术的高速线材盘卷形状与位置的检测仪，其特征在于，所述装置包括：图像采集系统、数据传输系统和数据处理系统；
    所述图像采集系统包括安装在线材盘卷运输辊道正上方的红外热成像相机，该红外热成像相机的视野宽度大于线材盘卷运输辊道宽度，用于获取盘卷运输辊道上的盘卷的原始物料图像；
    所述数据处理系统包括工业控制计算机和显示器；所述工业控制计算机用于采用权利要求1-5任一项中所述检测方法对图像采集系统采集的原始物料图像进行计算分析，得到当前原始物料图像中所有盘卷的形状与位置情况；所述显示器用于同步显示原始物料图像和工业控制计算机输出的分析结果；
    所述数据传输系统包括分别设置在现场控制箱和室内控制箱内的光纤收发器和用于连接图像采集系统和数据处理系统的光纤及网线。
  7. 根据权利要求6所述的基于红外技术的高速线材盘卷形状与位置的检测仪，其特征在于，所述数据处理系统包括报警模组，报警模组通过I/O接口卡与工业控制计算机连接；
    所述工业控制计算机对当前原始物料图像中所有盘卷的形状与位置情况进行判断，若出现等效直径或中心位置突变，发送警报信号至报警模组，使报警模组发出声光警报。
PCT/CN2022/119357 2022-01-12 2022-09-16 一种基于红外技术的高速线材盘卷形状与位置检测方法 WO2023134193A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202210031709.2A CN114383506B (zh) 2022-01-12 2022-01-12 一种基于红外技术的高速线材盘卷形状与位置检测方法
CN202210031709.2 2022-01-12

Publications (1)

Publication Number Publication Date
WO2023134193A1 true WO2023134193A1 (zh) 2023-07-20

Family

ID=81202086

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2022/119357 WO2023134193A1 (zh) 2022-01-12 2022-09-16 一种基于红外技术的高速线材盘卷形状与位置检测方法

Country Status (2)

Country Link
CN (1) CN114383506B (zh)
WO (1) WO2023134193A1 (zh)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116630315A (zh) * 2023-07-24 2023-08-22 山东东阿亿福缘阿胶制品有限公司 基于计算机视觉的饮料包装缺陷智能检测方法
CN116793229A (zh) * 2023-08-25 2023-09-22 钛玛科（北京）工业科技有限公司 一种基于3d相机的智能视觉测宽装置
CN117760313A (zh) * 2024-02-22 2024-03-26 钛玛科（北京）工业科技有限公司 一种用于卷材测宽和中心点偏差的扫描式测量方法

Families Citing this family (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114383506B (zh) * 2022-01-12 2024-03-19 江苏沙钢高科信息技术有限公司 一种基于红外技术的高速线材盘卷形状与位置检测方法
CN117197140B (zh) * 2023-11-07 2024-02-20 东莞市恒兴隆实业有限公司 基于机器视觉的不规则金属扣成型检测方法

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH10164436A (ja) * 1996-12-04 1998-06-19 Sony Corp 輪郭抽出装置、輪郭抽出方法、キー信号生成装置及びキー信号生成方法
CN101170641A (zh) * 2007-12-05 2008-04-30 北京航空航天大学 一种基于阈值分割的图像边缘检测方法
CN105631876A (zh) * 2015-12-29 2016-06-01 中国兵器科学研究院宁波分院 一种基于全局二值化的ct图像分辨率自动测试方法
CN113888517A (zh) * 2021-10-13 2022-01-04 广东奥普特科技股份有限公司 一种卷绕机下料视觉检测方法
CN114383506A (zh) * 2022-01-12 2022-04-22 江苏省沙钢钢铁研究院有限公司 一种基于红外技术的高速线材盘卷形状与位置检测方法

Family Cites Families (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106372642B (zh) * 2016-08-31 2023-11-03 北京航空航天大学 基于轮廓曲线分割弧合并与组合的椭圆快速检测方法
CN107507208B (zh) * 2017-07-12 2020-04-10 天津大学 一种基于轮廓上曲率估计的图像特征点提取方法
CN111308987B (zh) * 2020-02-21 2022-12-02 西安理工大学 基于图像处理的开卷机自动上卷控制系统及检测方法
CN112634256B (zh) * 2020-12-30 2024-03-15 杭州三坛医疗科技有限公司 圆检测及拟合方法、装置、电子设备和存储介质
CN113284154B (zh) * 2021-05-25 2022-04-26 武汉钢铁有限公司 钢卷端面图像分割方法、装置及电子设备
CN113554635A (zh) * 2021-07-30 2021-10-26 江苏苏桥焊材有限公司 基于人工智能的钢卷松卷自动检测方法和装置
CN113628195A (zh) * 2021-08-13 2021-11-09 宿迁林讯新材料有限公司 一种基于人工智能的铝卷长度预测方法及系统

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH10164436A (ja) * 1996-12-04 1998-06-19 Sony Corp 輪郭抽出装置、輪郭抽出方法、キー信号生成装置及びキー信号生成方法
CN101170641A (zh) * 2007-12-05 2008-04-30 北京航空航天大学 一种基于阈值分割的图像边缘检测方法
CN105631876A (zh) * 2015-12-29 2016-06-01 中国兵器科学研究院宁波分院 一种基于全局二值化的ct图像分辨率自动测试方法
CN113888517A (zh) * 2021-10-13 2022-01-04 广东奥普特科技股份有限公司 一种卷绕机下料视觉检测方法
CN114383506A (zh) * 2022-01-12 2022-04-22 江苏省沙钢钢铁研究院有限公司 一种基于红外技术的高速线材盘卷形状与位置检测方法

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116630315A (zh) * 2023-07-24 2023-08-22 山东东阿亿福缘阿胶制品有限公司 基于计算机视觉的饮料包装缺陷智能检测方法
CN116630315B (zh) * 2023-07-24 2023-09-29 山东东阿亿福缘阿胶制品有限公司 基于计算机视觉的饮料包装缺陷智能检测方法
CN116793229A (zh) * 2023-08-25 2023-09-22 钛玛科（北京）工业科技有限公司 一种基于3d相机的智能视觉测宽装置
CN117760313A (zh) * 2024-02-22 2024-03-26 钛玛科（北京）工业科技有限公司 一种用于卷材测宽和中心点偏差的扫描式测量方法
CN117760313B (zh) * 2024-02-22 2024-04-30 钛玛科（北京）工业科技有限公司 一种用于卷材测宽和中心点偏差的扫描式测量方法

Also Published As

Publication number Publication date
CN114383506A (zh) 2022-04-22
CN114383506B (zh) 2024-03-19

Similar Documents

Publication Publication Date Title
WO2023134193A1 (zh) 一种基于红外技术的高速线材盘卷形状与位置检测方法
US11221107B2 (en) Method for leakage detection of underground pipeline corridor based on dynamic infrared thermal image processing
CN115239704B (zh) 一种木材表面缺陷的精准检测修复方法
CN113628189A (zh) 一种基于图像识别的快速带钢划痕缺陷检测方法
CN113989716B (zh) 煤矿井下输送带异物目标检测方法、系统、设备及终端
CN107742307A (zh) 基于改进帧差法的输电线路舞动特征提取及参数分析方法
CN115239702B (zh) 一种塑料薄膜缺陷分析处理方法
CN114897894A (zh) 一种筒纱菊花芯缺陷检测方法
CN102002582A (zh) 一种利用图像识别的带钢同步分析监控方法和系统
CN115171022A (zh) 一种施工场景下的安全帽佩戴检测方法及系统
CN104392203B (zh) 一种基于视频分析的点状条纹噪声检测方法
CN118096796B (zh) 基于机器学习的径锻钛棒外观视觉检测方法
CN115147416A (zh) 一种倒绳机乱绳检测方法、装置和计算机设备
CN110813795B (zh) 一种化纤丝锭纸筒破损在线检测装置及方法
CN114655655B (zh) 一种基于UNet网络的传送带跑偏检测方法
CN115841633A (zh) 一种电力塔和电力线关联矫正的电力塔和电力线检测方法
CN117392136A (zh) 基于图像处理的玻璃盖板透光异常检测方法
CN115861318B (zh) 一种棉花加工生产质量检测方法
CN116026861A (zh) 一种玻璃瓶的检测方法及系统
CN115049644B (zh) 一种基于铝管表面瑕疵识别的温度控制方法及系统
CN116256371A (zh) 一种碳纤维卷线质量实时检测方法、装置及存储介质
TW201211348A (en) Yarn detecting device and method thereof
CN115239723A (zh) 基于图形识别的卷板机零件检测方法
CN110533698B (zh) 一种基于视觉检测的基坑施工桩检测控制方法
CN113588655A (zh) 一种mdf纤维线条表面缺陷的检测装置及其工作方法

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 22919828

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE