WO2020098177A1 - 一种基于线性回归拟合的天线下倾角测量方法 - Google Patents

一种基于线性回归拟合的天线下倾角测量方法 Download PDF

Info

Publication number
WO2020098177A1
WO2020098177A1 PCT/CN2019/076720 CN2019076720W WO2020098177A1 WO 2020098177 A1 WO2020098177 A1 WO 2020098177A1 CN 2019076720 W CN2019076720 W CN 2019076720W WO 2020098177 A1 WO2020098177 A1 WO 2020098177A1
Authority
WO
WIPO (PCT)
Prior art keywords
antenna
image
linear regression
mask
measuring
Prior art date
Application number
PCT/CN2019/076720
Other languages
English (en)
French (fr)
Inventor
邓文博
翟懿奎
柯琪锐
伍月婷
甘俊英
徐颖
王天雷
吴细
陈丽燕
Original Assignee
五邑大学
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 五邑大学 filed Critical 五邑大学
Priority to EP19870065.0A priority Critical patent/EP3683541A4/en
Priority to US16/975,599 priority patent/US20200410710A1/en
Publication of WO2020098177A1 publication Critical patent/WO2020098177A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/10—Segmentation; Edge detection
    • G06T7/11—Region-based segmentation
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01C—MEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
    • G01C1/00—Measuring angles
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
    • G01R29/00—Arrangements for measuring or indicating electric quantities not covered by groups G01R19/00 - G01R27/00
    • G01R29/08—Measuring electromagnetic field characteristics
    • G01R29/10—Radiation diagrams of antennas
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/10—Segmentation; Edge detection
    • G06T7/13—Edge detection
    • 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
    • G06T7/70—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
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/045—Combinations of networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning 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
    • G06T2207/20—Special algorithmic details
    • G06T2207/20081—Training; Learning
    • 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
    • G06T2207/20084—Artificial neural networks [ANN]

Definitions

  • the invention relates to the field of communication measurement, in particular to a method for measuring antenna downtilt angle based on linear regression fitting.
  • Antenna downtilt is one of the important parameters that determine the signal coverage of the base station. Not only the downtilt angle of each antenna needs to be accurately designed in the early stage of network planning, after the base station is put into operation, with the development of services, user changes and the surrounding signal environment Changes, it is also necessary to make accurate adjustments to the downtilt angle.
  • the inclinometer is generally used to measure the mechanical downtilt angle of the base station antenna.
  • the surveyor When using the inclinometer to measure the mechanical downtilt angle of the antenna, the surveyor must climb the iron tower or the pole to measure close to the antenna, which is very dangerous and troublesome, and also makes the accuracy of the measurement suffer. influences.
  • the GSM-R system With the development of technology, the GSM-R system has emerged. This system is a measurement system tool that can accurately measure the downtilt angle of the antenna without being close to the antenna. Measurement work, and can be connected to the test points of each base station to achieve real-time monitoring of the base station antenna inclination.
  • an object of the embodiments of the present invention is to provide an antenna downtilt measurement method based on linear regression fitting, so as to measure the antenna downtilt angle safely, effectively, quickly, and accurately.
  • An antenna downtilt measurement method based on linear regression fitting includes the following steps: the input original antenna image is subjected to image instance segmentation processing using a deep learning method to obtain a segmented image; masking the segmented image; masking processing Mathematical linear modeling and fitting of the segmented image after the masking; the mathematical linear modeling and fitting of the masked segmented image includes the following steps: extracting pixel value coordinates of the antenna edge contour from the masked segmented image , Intercept the pixel value at the right edge of the front side antenna plane; fit the pixel value coordinates into a straight line by mathematical linear modeling and fitting method and obtain the straight line slope to obtain the antenna downtilt angle.
  • the input antenna image is subjected to image instance segmentation processing using a deep learning method to obtain a segmented image including the following steps: using a convolutional neural network to obtain an antenna candidate frame and an antenna feature map; generating an interest area from the antenna candidate frame and Combined with the antenna feature map to obtain the feature map of the region of interest, pixel correction is performed on the region of interest.
  • the input antenna image is subjected to image instance segmentation processing using a deep learning method.
  • Obtaining a segmented image further includes the following steps: predicting the region of interest, obtaining a regression border mapped by the antenna feature map, and predicting pixels of the region of interest The category of points to get the segmented image.
  • the pixel correction is performed by using a residual network for alignment processing; the pixel correction includes two quantization processes, namely, mapping of the region of interest to the antenna feature map and mapping of the antenna feature map to the original antenna image.
  • the masking process of the divided image includes the following steps: extracting the image coordinates of the outline of the antenna from the divided image; mapping the image coordinates to the pixel coordinate system, and converting to binary coordinates through the Boer operation, and setting Convolution of the mask coordinates generates a new mask; the color generator is used to fill the new mask.
  • mapping the image coordinates to the pixel coordinate system includes converting the coordinate system.
  • the mathematical linear modeling fitting includes using gradient descent least squares to optimize the data samples.
  • the beneficial effects of the embodiments of the present invention are: a method for measuring the downtilt angle of an antenna based on linear regression fitting used in the embodiment of the present invention, which is directly output through processing of a deep learning network to obtain the downtilt angle of the antenna; at the same time, it is divided by mask processing
  • the segmented image afterward makes the straight line of mathematical linear modeling more closely match the true value of the antenna, and makes the downtilt angle of the antenna more accurate; the embodiment of the present invention avoids the danger of climbing measurement and reduces the cost of installing the sensor, which can be more effective and safe , Low-cost, accurate acquisition of antenna downtilt data.
  • FIG. 1 is a structural diagram of a deep learning method for performing image instance segmentation processing according to an embodiment of the present invention
  • FIG. 2 is a flow frame diagram of image instance segmentation according to an embodiment of the present invention.
  • FIG. 3 is a schematic diagram of the residual network alignment of the network of interest according to an embodiment of the present invention.
  • FIG. 4 is a diagram of a correspondence relationship between an image coordinate system and a pixel coordinate system according to an embodiment of the invention
  • FIG. 5 is an operation diagram of a mask operation according to an embodiment of the invention.
  • FIG. 6 is a graph of mathematical linear modeling and fitting according to an embodiment of the present invention.
  • a method for measuring the downtilt angle of an antenna based on linear regression fitting includes the following steps: the input original antenna image is subjected to image instance segmentation processing using a deep learning method to obtain a segmented image; Masking the segmented images; performing mathematical linear modeling and fitting on the segmented images after masking; performing mathematical linear modeling and fitting on the segmented images after masking includes the following steps: Extract the pixel value coordinates of the antenna edge contour from the segmented image and intercept the pixel values at the right edge of the antenna plane on the front side; fit the pixel value coordinates into a straight line by mathematical linear modeling and fitting method and obtain the straight line slope to obtain the antenna under the antenna Angle of inclination.
  • the input antenna image is subjected to image instance segmentation processing using a deep learning method to obtain a segmented image including the following steps: using a convolutional neural network to obtain an antenna candidate frame and an antenna feature map ; Generate the region of interest from the antenna candidate frame and combine the antenna feature map to obtain the feature map of the region of interest, and perform pixel correction on the region of interest.
  • the input antenna image is subjected to image instance segmentation processing using a deep learning method.
  • Obtaining a segmented image further includes the following steps: predicting the region of interest, obtaining a regression border mapped by the antenna feature map, and predicting pixels of the region of interest The category of points to get the segmented image.
  • the pixel correction is performed by using a residual network for alignment processing; the pixel correction includes two quantization processes, which are the mapping process of the region of interest to the antenna feature map and the mapping of the antenna feature map to the original antenna image, respectively Process; ensure one-to-one correspondence between input and output at the pixel level.
  • the masking process of the divided image includes the following steps: extracting the image coordinates of the outline of the antenna from the divided image; mapping the image coordinates to the pixel coordinate system, and converting by the Boer operation For the binary coordinates, convolution with the set mask coordinates to generate a new mask; use the color generator to fill the new mask.
  • the mapping of the image coordinates to the pixel coordinate system includes conversion processing of the coordinate system.
  • the pixel coordinate system and the image coordinate system are both on the imaging plane of the antenna image, but their respective origins and measurement units are different.
  • the origin of the image coordinate system is the intersection of the optical axis of the camera and the imaging plane, usually the center point of the imaging plane.
  • the unit of the image coordinate system is mm, and the unit of the pixel coordinate system is pixel.
  • the coordinate conversion formula is as follows:
  • u 0 and v 0 are the horizontal and vertical coordinates of the center point of the image coordinate system; R is the 3X3 orthogonal status quo matrix; T is the three-dimensional translation vector.
  • the divided images need to be masked through a mask branch network.
  • the mask branch network is a convolutional network that takes the positive region selected by the region of interest classifier as input and generates a mask of the positive region.
  • the generated mask corresponds to a low resolution of 28x28 pixels.
  • the generated mask is a soft mask represented by floating point numbers, and has more details than the binary mask.
  • the small size attribute of the mask helps keep the mask branch network lightweight. In the inference process, the predicted mask is enlarged to the size of the frame of the region of interest to give the final mask result.
  • the mathematical linear modeling fitting includes the use of gradient descent least squares to optimize the data samples.
  • the sample points are fitted to an approximate straight line, and the slope of the line obtained by the least squares error is obtained, and then the downtilt value of the base station antenna is accurately obtained.

Abstract

本发明公开了一种基于线性回归拟合的天线下倾角测量方法，包括以下步骤：对输入的原始天线图像利用深度学习方法进行图像实例分割处理，得到分割图像；对分割图像进行掩膜处理；对掩膜处理后的分割图像进行数学线性建模拟合；所述对掩膜处理后的分割图像进行数学线性建模拟合包括以下步骤：从掩膜处理后的分割图像中提取天线边缘轮廓的像素值坐标，截取位于天线右端边缘的像素值；通过数学线性建模拟合的方法将像素值坐标拟合成一条直线进而得到天线下倾角角度。通过对天线图像进行深度学习网络结合掩膜和线性拟合的处理得到天线下倾角角度，建立一种方便、安全、有效、准确的天线测量方法。

Description

一种基于线性回归拟合的天线下倾角测量方法 技术领域
本发明涉及通信测量领域，特别是一种基于线性回归拟合的天线下倾角测量方法。
背景技术
在通讯领域里，经常要对天线下倾角进行调整。天线下倾角是决定基站信号覆盖范围的重要参数之一，不但在网络规划的初期需要准确设计每个天线的下倾角，在基站投入运行以后，随着业务的发展，用户的变化以及周围信号环境的变化，还需要对下倾角做出准确调整。
目前对基站天线机械下倾角的测量普遍采用坡度计，使用坡度计测量天线机械下倾角时，测量者必须爬上铁塔或者抱杆贴近天线进行测量，相当危险和麻烦，也使得测量的准确性受到影响。而随着技术发展，出现了GSM-R系统，该系统是一种测量人员可以不用贴近天线就可以准确测量出天线下倾角的测量系工具，能够实现不登塔作业即可完成基站天线倾角的测量工作，并可对各基站测试点进行联网，实现对基站天线倾角的实时监测。但安装传感器，不仅耗时，成本较高，且新旧塔间、基站塔层数及数量等都存在差异性，因而该方法实用性不高，运行周期长，实现较为困难。因此设计出简单操作，性能可靠的角度测量方法就很有必要。
发明内容
为解决上述问题，本发明实施例的目的在于提供一种基于线性回归拟合的天线下倾角测量方法，以便安全、有效、快速、准确地对天线下倾角进行测量。
本发明实施例解决其问题所采用的技术方案是：
一种基于线性回归拟合的天线下倾角测量方法，包括以下步骤：对输入的原始天线图像利用深度学习方法进行图像实例分割处理，得到分割图像；对分割图像进行掩膜处理；对掩膜处理后的分割图像进行数学线性建模拟合；所述对掩膜处理后的分割图像进行数学线性建模拟合包括以下步骤：从掩膜处理后的分割图像中提取天线边缘轮廓的像素值坐标，截取位于正侧面天线平面右端边缘的像素值；通过数学线性建模拟合的方法将像素值坐标拟合成一条直线并获取直线斜率进而得到天线下倾角角度。
进一步，所述对输入的天线图像利用深度学习方法进行图像实例分割处理，得到分割图像包括以下步骤：利用卷积神经网络获得天线候选框和天线特征图；从天线候选框中生成感兴趣区域并结合天线特征图进而得到感兴趣区域的特征图，对感兴趣区域进行像素校正。
进一步，所述对输入的天线图像利用深度学习方法进行图像实例分割处理，得到分割图像还包括以下步骤：对感兴趣区域进行预测，得到天线特征图映射的回归边框，并预测感兴趣区域的像素点的类别，得到分割图像。
进一步，所述像素校正是通过利用残差网络进行对齐处理；所 述像素校正包括两个量化过程，分别是感兴趣区域到天线特征图的映射和天线特征图到原始天线图像的映射。
进一步，所述对分割图像进行掩膜处理包括以下步骤：从分割图像中提取该天线的轮廓的图像坐标；映射图像坐标到像素坐标系，并通过波尔运算转换为二值坐标，与设置的掩膜坐标卷积生成新掩膜；采用颜色生成器将新掩膜进行填充处理。
进一步，所述映射图像坐标到像素坐标系包括对坐标系进行转换处理。
优选地，生成新掩膜的操作公式为：I(i，j)＝5*I(i，j)-[I(i-1，j)+I(i+1，j)+I(i，j-1)+I(i，j+1)]；其中I(i，j)是图像中心元素。
进一步，所述数学线性建模拟合包括采用梯度下降最小二乘法实现对数据样本的优化。
优选地，拟合的直线的模型为：f(x)＝w Tx+b；其中WT为权重矩阵的转置，b为偏移量；天线下倾角角度计算公式为：θ＝arctan(|k|)；其中，k为通过梯度下降最小二乘法线性拟合的直线的斜率。
本发明实施例的有益效果是：本发明实施例采用的一种基于线性回归拟合的天线下倾角测量方法，经过深度学习网络的处理直接输出得到天线下倾角角度；同时经过掩膜处理实例分割后的分割图像使得数学线性建模的直线更加贴合天线的真实值，使天线下倾角角度更加准确；本发明实施例避免了攀爬测量的危险和减少安装传感器的成本， 能更加有效、安全、低成本、准确地获得天线下倾角数据。
附图说明
下面结合附图和实例对本发明作进一步说明。
图1是本发明实施例的进行图像实例分割处理的深度学习方法的结构图；
图2是本发明实施例的图像实例分割的流程框架图；
图3是本发明实施例的残差网络对感兴趣网络对齐的示意图；
图4是本发明实施例的图像坐标系与像素坐标系的对应关系图；
图5是本发明实施例的掩膜运算的运算图；
图6是本发明实施例的数学线性建模拟合的坐标图。
具体实施方式
在本发明的一个实施例中，公开了一种基于线性回归拟合的天线下倾角测量方法，包括以下步骤：对输入的原始天线图像利用深度学习方法进行图像实例分割处理，得到分割图像；对分割图像进行掩膜处理；对掩膜处理后的分割图像进行数学线性建模拟合；所述对掩膜处理后的分割图像进行数学线性建模拟合包括以下步骤：从掩膜处理后的分割图像中提取天线边缘轮廓的像素值坐标，截取位于正侧面天线平面右端边缘的像素值；通过数学线性建模拟合的方法将像素值坐标拟合成一条直线并获取直线斜率进而得到天线下倾角角度。
参照图1和图2，在一个实施例中，所述对输入的天线图像利用深度学习方法进行图像实例分割处理，得到分割图像包括以下步骤： 利用卷积神经网络获得天线候选框和天线特征图；从天线候选框中生成感兴趣区域并结合天线特征图进而得到感兴趣区域的特征图，对感兴趣区域进行像素校正。
进一步，所述对输入的天线图像利用深度学习方法进行图像实例分割处理，得到分割图像还包括以下步骤：对感兴趣区域进行预测，得到天线特征图映射的回归边框，并预测感兴趣区域的像素点的类别，得到分割图像。
参照图3，所述像素校正是通过利用残差网络进行对齐处理；所述像素校正包括两个量化过程，分别是感兴趣区域到天线特征图的映射过程和天线特征图到原始天线图像的映射过程；保证输入和输出之间在像素级别上的一一对应。
参照图5，在一个实施例中，所述对分割图像进行掩膜处理包括以下步骤：从分割图像中提取该天线的轮廓的图像坐标；映射图像坐标到像素坐标系，并通过波尔运算转换为二值坐标，与设置的掩膜坐标卷积生成新掩膜；采用颜色生成器将新掩膜进行填充处理。
优选地，生成新掩膜的操作公式为：I(i，j)＝5*I(i，j)-[I(i-1，j)+I(i+1，j)+I(i，j-1)+I(i，j+1)]；其中I(i，j)是图像中心元素。
参照图4，在一个实施例中，所述映射图像坐标到像素坐标系包括对坐标系进行转换处理。像素坐标系和图像坐标系都在天线图像的成像平面上，只是各自的原点和度量单位不一样。图像坐标系的原点 为相机光轴与成像平面的交点，通常情况下是成像平面的中心点。图像坐标系的单位是mm，而像素坐标系的单位是pixel。这二者之间的转换如下：其中dx和dy表示每一列和每一行分别代表多少mm，即1pixel＝dx mm。其中坐标转换公式如下：
Figure PCTCN2019076720-appb-000001
其中，u 0、v 0分别是图像坐标系的中心点的横坐标和纵坐标；R为3X3正交现状矩阵；T为三维平移向量。
对分割图像需要经过掩码分支网络进行掩膜处理。掩码分支网络是一个卷积网络，取感兴趣区域分类器选择的正区域作为输入，并生成正区域的掩码。生成的掩码是对应28x28像素低分辨率的。生成的掩码是由浮点数表示的软掩码，相对二进制掩码具有更多的细节。掩码的小尺寸属性有助于保持掩码分支网络的轻量性。在推断过程中，将预测的掩码放大为感兴趣区域的边框的尺寸以给出最终的掩码结果。
参照图6，所述数学线性建模拟合包括采用梯度下降最小二乘法实现对数据样本的优化。优选地，拟合的直线的模型为：f(x)＝w Tx+b； 其中w T为权重矩阵的转置，b为偏移量；天线下倾角角度计算公式为：θ＝arctan(|k|)；其中，k为通过梯度下降最小二乘法线性拟合的直线的斜率。
在一个实施例中，计算过程如下：用yi代表第i个点的真实值；f(xi)代表通过模型函数f后的预测值；得到欧式距离的表达式：distance＝(yi-f(xi))2，该式从损失函数的角度看，为平方误差，即
Figure PCTCN2019076720-appb-000002
得到拟合的目标函数为：
Figure PCTCN2019076720-appb-000003
将J(θ)经向量运算得：
Figure PCTCN2019076720-appb-000004
对θ进行偏导运算：
Figure PCTCN2019076720-appb-000005
通过令偏导等于零，将样本点拟合到近似的一条直线上，获取通过最小二乘误差得到的直线斜率，进而准确得到基站天线的下倾角数值。由如下反正切公式可知：θ＝arctan(|k|)；其中θ为天线下倾角，k为通过梯度下降最小二乘法线性拟合直线的斜率。
以上所述，只是本发明的较佳实施例而已，本发明并不局限于上述实施方式，只要其以相同的手段达到本发明的技术效果，都应属于本发明的保护范围。

Claims (9)

  1. 一种基于线性回归拟合的天线下倾角测量方法，其特征在于，包括以下步骤：
    对输入的原始天线图像利用深度学习方法进行图像实例分割处理，得到分割图像；
    对分割图像进行掩膜处理；
    对掩膜处理后的分割图像进行数学线性建模拟合；
    所述对掩膜处理后的分割图像进行数学线性建模拟合包括以下步骤：从掩膜处理后的分割图像中提取天线边缘轮廓的像素值坐标，截取位于正侧面天线平面右端边缘的像素值；通过数学线性建模拟合的方法将像素值坐标拟合成一条直线并获取直线斜率进而得到天线下倾角角度。
  2. 根据权利要求1所述的一种基于线性回归拟合的天线下倾角测量方法，其特征在于，所述对输入的天线图像利用深度学习方法进行图像实例分割处理，得到分割图像包括以下步骤：
    利用卷积神经网络获得天线候选框和天线特征图；
    从天线候选框中生成感兴趣区域并结合天线特征图进而得到感兴趣区域的特征图，对感兴趣区域进行像素校正。
  3. 根据权利要求2所述的一种基于线性回归拟合的天线下倾角测量方法，其特征在于，所述对输入的天线图像利用深度学习方法进行图像实例分割处理，得到分割图像还包括以下步骤：对感兴趣 区域进行预测，得到天线特征图映射的回归边框，并预测感兴趣区域的像素点的类别，得到分割图像。
  4. 根据权利要求2或3任一项所述的一种基于线性回归拟合的天线下倾角测量方法，其特征在于，所述像素校正是通过利用残差网络进行对齐处理；所述像素校正包括两个量化过程，分别是感兴趣区域到天线特征图的映射和天线特征图到原始天线图像的映射。
  5. 根据权利要求1所述的一种基于线性回归拟合的天线下倾角测量方法，其特征在于，所述对分割图像进行掩膜处理包括以下步骤：从分割图像中提取该天线的轮廓的图像坐标；映射图像坐标到像素坐标系，并通过波尔运算转换为二值坐标，与设置的掩膜坐标卷积生成新掩膜；采用颜色生成器将新掩膜进行填充处理。
  6. 根据权利要求5所述的一种基于线性回归拟合的天线下倾角测量方法，其特征在于，所述映射图像坐标到像素坐标系包括对坐标系进行转换处理。
  7. 根据权利要求5所述的一种基于线性回归拟合的天线下倾角测量方法，其特征在于，生成新掩膜的操作公式为：I(i，j)＝5*I(i，j)-[I(i-1，j)+I(i+1，j)+I(i，j-1)+I(i，j+1)]；其中I(i，j)是图像中心元素。
  8. 根据权利要求1所述的一种基于线性回归拟合的天线下倾角测量方法，其特征在于，所述数学线性建模拟合包括采用梯度下降最小二乘法实现对数据样本的优化。
  9. 根据权利要求8所述的一种基于线性回归拟合的天线下倾角测量方法，其特征在于，拟合的直线的模型为：f(x)＝w Tx+b；其中WT为权重矩阵的转置，b为偏移量；天线下倾角角度计算公式为：θ＝arctan(|k|)；其中，k为通过梯度下降最小二乘法线性拟合的直线的斜率。
PCT/CN2019/076720 2018-11-06 2019-03-01 一种基于线性回归拟合的天线下倾角测量方法 WO2020098177A1 (zh)

Priority Applications (2)

Application Number Priority Date Filing Date Title
EP19870065.0A EP3683541A4 (en) 2018-11-06 2019-03-01 METHOD OF MEASURING THE DOWN TILT ANGLE OF AN ANTENNA BASED ON A LINEAR REGRESSION ADJUSTMENT
US16/975,599 US20200410710A1 (en) 2018-11-06 2019-03-01 Method for measuring antenna downtilt based on linear regression fitting

Applications Claiming Priority (3)

Application Number Priority Date Filing Date Title
CN201811321973 2018-11-06
CN201811363450.1A CN109458980B (zh) 2018-11-06 2018-11-15 一种基于线性回归拟合的天线下倾角测量方法
CN201811363450.1 2018-11-15

Publications (1)

Publication Number Publication Date
WO2020098177A1 true WO2020098177A1 (zh) 2020-05-22

Family

ID=65610628

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2019/076720 WO2020098177A1 (zh) 2018-11-06 2019-03-01 一种基于线性回归拟合的天线下倾角测量方法

Country Status (4)

Country Link
US (1) US20200410710A1 (zh)
EP (1) EP3683541A4 (zh)
CN (1) CN109458980B (zh)
WO (1) WO2020098177A1 (zh)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112880622A (zh) * 2021-02-04 2021-06-01 上海航天控制技术研究所 一种应用倾角仪标定柔性喷管摆角传感器的方法
CN113343987A (zh) * 2021-06-30 2021-09-03 北京奇艺世纪科技有限公司 文本检测处理方法、装置、电子设备及存储介质

Families Citing this family (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109579774B (zh) * 2018-11-06 2021-04-23 五邑大学 一种基于深度实例分割网络的天线下倾角测量方法
US20200301021A1 (en) * 2019-03-19 2020-09-24 Adva Optical Networking Se Method and apparatus for automatic detection of antenna site conditions
CN110263390B (zh) * 2019-05-24 2023-04-25 五邑大学 基于无人机视觉测量的天线下倾角自动调整方法及系统
CN110415239B (zh) * 2019-08-01 2022-12-16 腾讯科技（深圳）有限公司 图像处理方法、装置、设备、医疗电子设备以及介质
US11074707B2 (en) * 2019-08-13 2021-07-27 Wuyi University Method and system of antenna measurement for mobile communication base station
CN110660096B (zh) * 2019-10-08 2023-05-23 珠海格力电器股份有限公司 曲线一致性检测方法及存储介质
CN112070721B (zh) * 2020-08-13 2024-01-12 五邑大学 基于实例分割网络的天线参数测量方法、装置及存储介质
CN114931112B (zh) * 2022-04-08 2024-01-26 南京农业大学 基于智能巡检机器人的母猪体尺检测系统

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US4758840A (en) * 1986-03-11 1988-07-19 Centre National D'etudes Spatiales Means for calibrating the elevation and azimuth angles of the scan axis of an antenna
CN105761249A (zh) * 2016-02-01 2016-07-13 南京工程学院 一种基于图像计算天线机械下倾角的方法
CN107121125A (zh) * 2017-06-12 2017-09-01 哈尔滨工业大学 一种通讯基站天线位姿自动检测装置与方法
CN107830846A (zh) * 2017-09-30 2018-03-23 杭州艾航科技有限公司 一种利用无人机和卷积神经网络测量通信塔天线角度方法

Family Cites Families (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6400831B2 (en) * 1998-04-02 2002-06-04 Microsoft Corporation Semantic video object segmentation and tracking
CN102077108B (zh) * 2008-04-28 2015-02-25 康奈尔大学 分子mri中的磁敏度精确量化
US20110150317A1 (en) * 2009-12-17 2011-06-23 Electronics And Telecommunications Research Institute System and method for automatically measuring antenna characteristics
US8907261B1 (en) * 2011-04-28 2014-12-09 Steve Wishstar Electromagnetic wave detection
AT511191B1 (de) * 2011-07-01 2012-10-15 Thomas Dr Neubauer Verfahren und vorrichtung zur bestimmung und speicherung von position und ausrichtung von antennenstrukturen
CN108647663B (zh) * 2018-05-17 2021-08-06 西安电子科技大学 基于深度学习和多层次图结构模型的人体姿态估计方法
US20220004740A1 (en) * 2018-09-26 2022-01-06 Sitesee Pty Ltd Apparatus and Method For Three-Dimensional Object Recognition

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US4758840A (en) * 1986-03-11 1988-07-19 Centre National D'etudes Spatiales Means for calibrating the elevation and azimuth angles of the scan axis of an antenna
CN105761249A (zh) * 2016-02-01 2016-07-13 南京工程学院 一种基于图像计算天线机械下倾角的方法
CN107121125A (zh) * 2017-06-12 2017-09-01 哈尔滨工业大学 一种通讯基站天线位姿自动检测装置与方法
CN107830846A (zh) * 2017-09-30 2018-03-23 杭州艾航科技有限公司 一种利用无人机和卷积神经网络测量通信塔天线角度方法

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
See also references of EP3683541A4 *

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112880622A (zh) * 2021-02-04 2021-06-01 上海航天控制技术研究所 一种应用倾角仪标定柔性喷管摆角传感器的方法
CN113343987A (zh) * 2021-06-30 2021-09-03 北京奇艺世纪科技有限公司 文本检测处理方法、装置、电子设备及存储介质
CN113343987B (zh) * 2021-06-30 2023-08-22 北京奇艺世纪科技有限公司 文本检测处理方法、装置、电子设备及存储介质

Also Published As

Publication number Publication date
CN109458980B (zh) 2021-01-26
US20200410710A1 (en) 2020-12-31
CN109458980A (zh) 2019-03-12
EP3683541A1 (en) 2020-07-22
EP3683541A4 (en) 2021-03-17

Similar Documents

Publication Publication Date Title
WO2020098177A1 (zh) 一种基于线性回归拟合的天线下倾角测量方法
CN112766274B (zh) 一种基于Mask RCNN算法的水尺图像水位自动读数方法及系统
CN107830846B (zh) 一种利用无人机和卷积神经网络测量通信塔天线角度方法
CN108510551B (zh) 一种远距离大视场条件下相机参数的标定方法及系统
CN106971408B (zh) 一种基于时空转换思想的摄像机标定方法
CN110826549A (zh) 基于计算机视觉的巡检机器人仪表图像识别方法及系统
CN104200086A (zh) 宽基线可见光相机位姿估计方法
CN110363758B (zh) 一种光学遥感卫星成像质量确定方法及系统
CN113554667B (zh) 一种基于图像识别的三维位移检测方法及装置
CN116597013B (zh) 一种基于不同经纬度区域的卫星图像几何定标方法
CN112634373A (zh) 基于零膨胀陶瓷标定板的视觉测量系统动态校正方法
CN107784633B (zh) 适用于平面测量的无人机航拍图像校准方法
CN109974618A (zh) 多传感器视觉测量系统的全局标定方法
CN115578315A (zh) 一种基于无人机图像的桥梁应变近景摄影测量方法
CN109308714A (zh) 基于分类惩罚的摄像头和激光雷达信息配准方法
CN110068313B (zh) 一种基于投影变换的数字天顶仪定向方法
CN105701776A (zh) 一种用于自动光学检测的镜头畸变矫正方法及系统
CN111142137B (zh) 一种点源靶标像控点的定位方法和装置
CN112435303B (zh) 振镜系统校正表构建方法、构建系统及振镜系统校正方法
CN110146032A (zh) 基于光场分布的合成孔径相机标定方法
CN116091488B (zh) 一种发动机摇摆试验的位移测试方法及位移测试系统
CN108917722A (zh) 植被覆盖度计算方法及装置
CN116152325A (zh) 一种基于单目视频的道路交通高边坡稳定性监测方法
CN114509018A (zh) 一种全场实时桥梁挠度测量方法
CN113945159B (zh) 一种基于轮廓匹配的螺栓直径测量方法

Legal Events

Date Code Title Description
ENP Entry into the national phase

Ref document number: 2019870065

Country of ref document: EP

Effective date: 20200416

NENP Non-entry into the national phase

Ref country code: DE