WO2021208515A1 - 基于图像与xrf矿物反演的岩性识别系统及方法 - Google Patents

基于图像与xrf矿物反演的岩性识别系统及方法 Download PDF

Info

Publication number
WO2021208515A1
WO2021208515A1 PCT/CN2020/141564 CN2020141564W WO2021208515A1 WO 2021208515 A1 WO2021208515 A1 WO 2021208515A1 CN 2020141564 W CN2020141564 W CN 2020141564W WO 2021208515 A1 WO2021208515 A1 WO 2021208515A1
Authority
WO
WIPO (PCT)
Prior art keywords
information
image
mineral
lithology
surrounding rock
Prior art date
Application number
PCT/CN2020/141564
Other languages
English (en)
French (fr)
Inventor
许振浩
林鹏
许广璐
石恒
余腾飞
王朝阳
华一磊
Original Assignee
山东大学
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 山东大学 filed Critical 山东大学
Priority to US17/777,700 priority Critical patent/US11796493B2/en
Publication of WO2021208515A1 publication Critical patent/WO2021208515A1/zh

Links

Images

Classifications

    • GPHYSICS
    • G01MEASURING; TESTING
    • G01NINVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N23/00Investigating or analysing materials by the use of wave or particle radiation, e.g. X-rays or neutrons, not covered by groups G01N3/00 – G01N17/00, G01N21/00 or G01N22/00
    • G01N23/22Investigating or analysing materials by the use of wave or particle radiation, e.g. X-rays or neutrons, not covered by groups G01N3/00 – G01N17/00, G01N21/00 or G01N22/00 by measuring secondary emission from the material
    • G01N23/223Investigating or analysing materials by the use of wave or particle radiation, e.g. X-rays or neutrons, not covered by groups G01N3/00 – G01N17/00, G01N21/00 or G01N22/00 by measuring secondary emission from the material by irradiating the sample with X-rays or gamma-rays and by measuring X-ray fluorescence
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06NCOMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00Computing arrangements based on biological models
    • G06N3/02Neural networks
    • G06N3/04Architecture, e.g. interconnection topology
    • G06N3/044Recurrent networks, e.g. Hopfield networks
    • EFIXED CONSTRUCTIONS
    • E21EARTH OR ROCK DRILLING; MINING
    • E21DSHAFTS; TUNNELS; GALLERIES; LARGE UNDERGROUND CHAMBERS
    • E21D9/00Tunnels or galleries, with or without linings; Methods or apparatus for making thereof; Layout of tunnels or galleries
    • GPHYSICS
    • G01MEASURING; TESTING
    • G01NINVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N2223/00Investigating materials by wave or particle radiation
    • G01N2223/40Imaging
    • G01N2223/406Imaging fluoroscopic image
    • GPHYSICS
    • G01MEASURING; TESTING
    • G01NINVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N2223/00Investigating materials by wave or particle radiation
    • G01N2223/40Imaging
    • G01N2223/42Imaging image digitised, -enhanced in an image processor
    • GPHYSICS
    • G01MEASURING; TESTING
    • G01NINVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N2223/00Investigating materials by wave or particle radiation
    • G01N2223/60Specific applications or type of materials
    • G01N2223/616Specific applications or type of materials earth materials
    • GPHYSICS
    • G01MEASURING; TESTING
    • G01NINVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N2223/00Investigating materials by wave or particle radiation
    • G01N2223/60Specific applications or type of materials
    • G01N2223/631Specific applications or type of materials large structures, walls
    • GPHYSICS
    • G01MEASURING; TESTING
    • G01VGEOPHYSICS; GRAVITATIONAL MEASUREMENTS; DETECTING MASSES OR OBJECTS; TAGS
    • G01V5/00Prospecting or detecting by the use of ionising radiation, e.g. of natural or induced radioactivity
    • G01V5/04Prospecting or detecting by the use of ionising radiation, e.g. of natural or induced radioactivity specially adapted for well-logging
    • G01V5/08Prospecting or detecting by the use of ionising radiation, e.g. of natural or induced radioactivity specially adapted for well-logging using primary nuclear radiation sources or X-rays
    • G01V5/12Prospecting or detecting by the use of ionising radiation, e.g. of natural or induced radioactivity specially adapted for well-logging using primary nuclear radiation sources or X-rays using gamma or X-ray sources

Definitions

  • the invention belongs to the field of intelligent identification of tunnel surrounding rock lithology, and in particular relates to a lithology identification system and method based on image and XRF mineral inversion.
  • the first aspect of the present invention provides a lithology recognition system based on image and XRF mineral inversion.
  • Intelligent identification of lithology not only avoids subjective misjudgment, but also realizes the automation and intelligence of lithology identification, which greatly reduces the time spent and improves work efficiency.
  • XRF X Ray Fluorescence
  • a lithology recognition system based on image and XRF mineral inversion including an automatic driving vehicle, the automatic driving vehicle is provided with a probe, an image acquisition device and an on-board processor;
  • the probe is mounted on the end of the robotic arm, and the robotic arm is installed on an autonomous vehicle; the probe is an X-ray fluorescence spectrometer, which is used to detect the element information of the surrounding rock in the area to be measured;
  • the image acquisition device is arranged around the autonomous vehicle for collecting surrounding rock images and making these surrounding rock images form a closed ring;
  • the on-board processor is used to: invert the received element information of the surrounding rock in the area to be measured into mineral information based on the Barth-Nigri standard mineral calculation method;
  • the second aspect of the present invention provides a recognition method of a lithology recognition system based on image and XRF mineral inversion.
  • the intelligent identification of surrounding rock lithology not only avoids subjective misjudgment, but also realizes the automation and intelligence of lithology identification, which greatly reduces the time spent and improves work efficiency.
  • a recognition method for lithology recognition system based on image and XRF mineral inversion including:
  • Control the robotic arm to drive the probe to the area to be measured, and use the probe to detect the element information of the surrounding rock in the area to be measured;
  • the received element information of the surrounding rock in the area to be measured is inverted into mineral information; the surrounding rock image of the same measured area is converted into a one-dimensional vector format The image information is then spliced with the mineral information in a one-dimensional format, and the spliced information is judged based on a preset neural network to identify rock lithology.
  • the self-driving vehicle of the present invention moves to a preset position, and based on the rock image information and the mineral information retrieved from the element, the preset neural network is used to intelligently identify the surrounding rock lithology, which not only avoids subjective misjudgment, but also realizes The automation and intelligence of lithology identification have greatly shortened the time spent and improved work efficiency.
  • the present invention replaces the traditional method of manually identifying lithology, avoids the adverse effects of unskilled personnel and subjective factors on the results, and improves the accuracy of lithology recognition.
  • Fig. 1 is a working principle diagram of a lithology recognition system based on image and XRF mineral inversion according to an embodiment of the present invention
  • Figure 2 is a schematic diagram of a mechanical arm and a probe according to an embodiment of the present invention
  • Fig. 3 is a schematic diagram of the probe structure according to an embodiment of the present invention.
  • azimuth or position relationship is based on the azimuth or position relationship shown in the drawings, and is only a relationship term determined to facilitate the description of the structural relationship of each component or element of the present invention. It does not specifically refer to any component or element in the present invention and cannot be understood as a reference Limitations of the invention.
  • Flatten Refers to the flattening of the output of the convolutional network, taking the three-dimensional image into one dimension, and then inputting it into the fully connected network.
  • the lithology recognition system based on image and XRF mineral inversion of this embodiment includes an autonomous vehicle 1, which is provided with a probe 6, an image acquisition device 4, and an on-board processor 7. ;
  • the probe 6 is mounted on the end of the robotic arm 5, and the robotic arm 5 is mounted on the autonomous vehicle 1; the probe is an X-ray fluorescence spectrometer, which is used to detect the element information of the surrounding rock in the area to be measured;
  • the image acquisition device 4 is arranged around the autonomous vehicle 1 for collecting surrounding rock images and making these surrounding rock images form a closed ring;
  • the on-board processor 7 is used for: inverting the received element information of the surrounding rock in the area to be measured into mineral information based on the Barth-Nigri standard mineral calculation method;
  • Receive the surrounding rock image and the corresponding probe inclination convert the surrounding rock image of the same measured area into image information in one-dimensional vector format, and then splice it with the mineral information in one-dimensional format itself, and pair the spliced information based on a preset neural network Perform discrimination to identify rock lithology, store the result of rock lithology recognition and the inclination of the probe.
  • the Barth-Nigri standard mineral calculation method is also known as the "cation standard mineral method".
  • the image acquisition device 4 adopts a camera to realize the omnidirectional photography of the surrounding rock of the tunnel, collect the image information of the surrounding rock, indicate the angle, and store the collected image information to the on-board processor 7.
  • the self-driving vehicle 1 when the self-driving vehicle 1 is traveling in the tunnel, the self-driving vehicle 1 moves to a preset position, and the self-driving vehicle 1 is also provided with a laser receiver 3, which is used to receive the laser transmitter 2. Measure and record the displacement distance and location of the emitted laser to realize the positioning of the autonomous vehicle 1; the laser transmitter 2 is placed at the tunnel entrance and is used to emit lasers into the tunnel.
  • the GPS positioning module is installed on the self-driving vehicle 1 to realize the automatic positioning function of the self-driving vehicle 1.
  • the robotic arm 5 is connected to a control system, and the control system is used to control the robotic arm 5 to drive the probe 6 to move around the autonomous vehicle 1 and reach the surrounding rock surface.
  • the robot arm includes 4 robot arm joints, and the robot arm 5 is controlled to rotate and move through the robot arm joints.
  • the probe 6 is provided with a pressure sensor 8 and an inclination angle sensor 9.
  • the pressure sensor 8 is used to measure the pressure information between the probe's pre-surrounding rock surface and feed it back to the control system, and the inclination sensor 9 is used to detect the inclination angle of the probe and Feedback to the control system.
  • the control system includes a controller and a memory.
  • the controller can be a programmable logic device or a CPU, etc.; the memory can be ROM, RAM, or an external storage device, such as a U disk.
  • the controller is used to receive the information transmitted by the pressure sensor and the inclination sensor and control the movement of the mechanical arm 5.
  • the on-board processor 7 includes:
  • the data storage unit is used to sequentially store the surrounding rock images according to the position information of the autonomous vehicle; and sequentially store the element information of the corresponding surrounding rock according to the position information of the autonomous vehicle and the inclination information of the probe;
  • Lithology identification part which is used to identify rock lithology.
  • the data processing unit is also used to perform preprocessing such as clipping and occlusion on the surrounding rock image.
  • the lithology recognition unit is used to convert the surrounding rock image of the same measured area into one-dimensional vector image information through convolution, pooling and flattening based on the B-P neural network.
  • the lithology recognition unit is also used to join the image information of the one-dimensional vector with the one-dimensional mineral information based on the BP neural network, and then transfer the joined rock information to the fully connected layer to distinguish it. Intelligent recognition of rock lithology.
  • B-P neural networks restricted Boltzmann machines, support vector machines, etc. can also complete tasks.
  • a preset neural network is used to intelligently identify surrounding rock lithology, which not only avoids subjective misjudgment, but also realizes the automation and intelligence of lithology recognition, which greatly Shorten the time spent and improve work efficiency.
  • Control the robotic arm to drive the probe to the area to be measured, and use the probe to detect the element information of the surrounding rock in the area to be measured;
  • the received element information of the surrounding rock in the area to be measured is inverted into mineral information; the surrounding rock image of the same measured area is converted into a one-dimensional vector format The image information is then spliced with the mineral information in a one-dimensional format.
  • the spliced information is judged based on a preset neural network to identify rock lithology, and the result of rock lithology recognition and the inclination of the probe are stored.
  • the surrounding rock image of the same measured area is converted into one-dimensional vector image information through convolution, pooling and flattening, and then it is one-dimensional with itself.
  • the mineral information is spliced together, and then the spliced rock information is transferred to the fully connected layer to distinguish it and intelligently identify the rock lithology.
  • the identification process of the lithology identification system based on image and XRF mineral inversion is as follows:
  • a laser transmitter is placed in the center of the tunnel entrance, and the autonomous vehicle will automatically navigate and drive through its own camera, and receive the emitted laser through a laser receiver placed at the rear of the vehicle for automatic positioning;
  • the manipulator is extended, and the probe is carried to the inner surface of the surrounding rock at a certain angle.
  • the probe is close to the rock surface to measure the rock element information, and the rock element information of several places in the area can be measured, which is numbered and saved according to the different inclination angle of the probe;
  • the on-board processor preprocesses the image to obtain effective information, and automatically inverts the elements into minerals according to the Barth-Negri standard mineral calculation method, selects the processed image information and mineral information, and combines the information in the same area Corresponding to save;
  • the on-board processor uses the image information and mineral information of the same area to intelligently identify the lithology of the area, which will be processed by convolution, pooling, and flattening into image information in one-dimensional vector format and mineral information in one-dimensional format. Splicing together, through the method of machine learning, based on the BP neural network to distinguish the spliced information to identify the rock lithology, and mark the result coordinates and inclination angles, and save them;
  • the autonomous vehicle moves to the next location, collects the surrounding rock information of the next area and performs lithology identification.

Landscapes

  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Health & Medical Sciences (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • General Health & Medical Sciences (AREA)
  • Pathology (AREA)
  • Immunology (AREA)
  • Biochemistry (AREA)
  • Analytical Chemistry (AREA)
  • Chemical & Material Sciences (AREA)
  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Biophysics (AREA)
  • General Engineering & Computer Science (AREA)
  • Computing Systems (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • Molecular Biology (AREA)
  • Evolutionary Computation (AREA)
  • Data Mining & Analysis (AREA)
  • Computational Linguistics (AREA)
  • Biomedical Technology (AREA)
  • Artificial Intelligence (AREA)
  • Image Analysis (AREA)
  • Analysing Materials By The Use Of Radiation (AREA)

Abstract

一种基于图像与XRF矿物反演的岩性识别系统及方法,其不依靠人工识别,耗费时间短,不容易误判,可自动识别岩性,工作效率高。识别系统包括自动驾驶车辆(1),探头(6),图像采集装置(4),车载处理器(7)。探头(6)为用于检测围岩的元素信息的X射线荧光光谱仪,车载处理器(7)将接收的待测区域围岩的元素信息反演成为矿物信息。识别方法接收围岩图像及对应倾角,将图像信息转化为一维向量格式,再与一维格式的矿物信息拼接,基于拼接后的信息识别岩石岩性。

Description

基于图像与XRF矿物反演的岩性识别系统及方法 技术领域
本发明属于隧道围岩岩性智能识别领域,尤其涉及一种基于图像与XRF矿物反演的岩性识别系统及方法。
背景技术
本部分的陈述仅仅是提供了与本发明相关的背景技术信息,不必然构成在先技术。
近些年,随着中国基础设施建设的快速发展,越来越多的隧道施工难题出现在工程人面前。由于隧道施工在岩土体中进行,工作环境存在很强的复杂性和不确定性,灾害也存在着突发性,如何保障施工的安全,始终是困扰工程人的难题。在长期的工程实践探索中,判别围岩岩性对于隧道施工的重要性逐渐凸显出来。大量工程实例证明,在隧道工程施工时,围岩及断面岩石的岩性对于围岩强度分级、施工方法的选择、灾害的超前预报都起着十分重要的作用。在一定程度上,围岩的岩性甚至决定了隧道施工的成本和后期运行的安全。因此,为了保证隧道的安全及质量,在隧道施工时必须要对围岩及断面进行岩性识别。
发明人发现,现有的围岩识别方法大多依靠人工进行,由于工作人员经验和水平的差异及个人主观性的存在,岩性识别的结果往往相差较大,准确性无法得到保障,不能很好满足工程实际的需求。而且由于隧道环境比较艰苦、恶劣,由工作人员人工进行岩性识别工作强度很大,耗费的时间较长,效率难以保证。
发明内容
为了解决上述问题,本发明的第一个方面提供一种基于图像与XRF矿物反演的岩性识别系统,其基于岩石图像信息和元素反演的矿物信息,利用预设神经网络对围岩岩性进行智能识别,不仅避免了主观误判,还实现了岩性识别的自动化、智能化,极大地缩短了所耗费的时间,提高了工作效率。
其中,XRF:X射线荧光光谱分析(X Ray Fluorescence)。
为了实现上述目的,本发明采用如下技术方案:
一种基于图像与XRF矿物反演的岩性识别系统,包括自动驾驶车辆,所述自动驾驶车辆上设置有探头、图像采集装置和车载处理器;
所述探头搭载在机械臂末端,机械臂安装在自动驾驶车辆上;所述探头为X射线荧光光谱仪,用于检测待测区域围岩的元素信息;
所述图像采集装置环绕自动驾驶车辆一周设置,用于采集围岩图像且使这些围岩图像可形成一个闭合的环;
所述车载处理器用于:基于巴尔特-尼格里标准矿物计算法将接收的待测区域围岩的元素信息反演成为矿物信息;
接收围岩图像及对应探头倾角,将同一被测区域的围岩图像转化为一维向量格式的图像信息,再与本身为一维格式的矿物信息拼接,基于预设神经网络对拼接后的信息进行判别以识别岩石岩性。
为了解决上述问题,本发明的第二个方面提供一种基于图像与XRF矿物反演的岩性识别系统的识别方法,其基于岩石图像信息和元素反演的矿物信息,利用预设神经网络对围岩岩性进行智能识别,不仅避免了主观误判,还实现了岩性识别的自动化、智能化,极大地缩短了所耗费的时间,提高了工作效率。
为了实现上述目的,本发明采用如下技术方案:
一种基于图像与XRF矿物反演的岩性识别系统的识别方法,包括:
控制机械臂带动探头至待测区域,利用探头检测待测区域围岩的元素信息;
利用图像采集装置采集围岩图像且使这些围岩图像可形成一个闭合的环;
在车载处理器中,基于巴尔特-尼格里标准矿物计算法将接收的待测区域围岩的元素信息反演成为矿物信息;将同一被测区域的围岩图像转化为一维向量格式的图像信息,再与本身为一维格式的矿物信息拼接,基于预设神经网络对拼接后的信息进行判别以识别岩石岩性。
本发明的有益效果是:
(1)本发明自动驾驶车辆移动至预设位置,且基于岩石图像信息和元素反演的矿物信息,利用预设神经网络对围岩岩性进行智能识别,不仅避免了主观误判,还实现了岩性识别的自动化、智能化,极大地缩短了所耗费的时间,提高了工作效率。
(2)本发明取代了传统人工辨别岩性的方法,最大限度避免了由于人员不熟练和主观因素对于结果的不良影响,提升了岩性识别的准确率。
(3)本发明全程由自动化设备完成,不需要工作人员深入隧道,极大提高了岩石辨别工作的安全性。
附图说明
构成本发明的一部分的说明书附图用来提供对本发明的进一步理解,本发明的示意性实施例及其说明用于解释本发明,并不构成对本发明的不当限定。
图1是本发明实施例的基于图像与XRF矿物反演的岩性识别系统工作原理图;
图2是本发明实施例的机械臂及探头原理图;
图3是本发明实施例的探头结构示意图。
其中,1、自动驾驶车辆;2、激光发射器;3、激光接收器;4、图像采集装置;5、机械臂;6、探头;7、车载处理器;8、压力传感器;9、倾角传感器。
具体实施方式
下面结合附图与实施例对本发明作进一步说明。
应该指出,以下详细说明都是例示性的,旨在对本发明提供进一步的说明。除非另有指明,本文使用的所有技术和科学术语具有与本发明所属技术领域的普通技术人员通常理解的相同含义。
需要注意的是,这里所使用的术语仅是为了描述具体实施方式,而非意图限制根据本发明的示例性实施方式。如在这里所使用的,除非上下文另外明确指出,否则单数形式也意图包括复数形式,此外,还应当理解的是,当在本说明书中使用术语“包含”和/或“包括”时,其指明存在特征、步骤、操作、器件、组件和/或它们的组合。
在本发明中,术语如“上”、“下”、“左”、“右”、“前”、“后”、“竖直”、“水平”、“侧”、“底”等指示的方位或位置关系为基于附图所示的方位或位置关系,只是为了便于叙述本发明各部件或元件结构关系而确定的关系词,并非特指本发明中任一部件或元件,不能理解为对本发明的限制。
本发明中,术语如“固接”、“相连”、“连接”等应做广义理解,表示可以是固定连接,也可以是一体地连接或可拆卸连接;可以是直接相连,也可以通过中间媒介间接相连。对于本领域的相关科研或技术人员,可以根据具体情况确定上述术语在本发明中的具体含义,不能理解为对本发明的限制。
术语解释:
Flatten:指的是卷积网络的输出拍扁,把三维的拍成一维,之后再输入全连接网络。
随着各种隧道工法不断发展,依靠人工去辨别岩石岩性的方法明显不合时宜,如果无法提高岩性辨别的准确率和效率,可能就无法发挥隧道整体施工的潜力。目前无论是隧道施工还是科研研究,都需要一种能够保证岩石岩性辨别准确率和效率的方法。
如图1所示,本实施例的基于图像与XRF矿物反演的岩性识别系统,包括自动驾驶车辆1,所述自动驾驶车辆1上设置有探头6、图像采集装置4和车载处理器7;
所述探头6搭载在机械臂5末端,机械臂5安装在自动驾驶车辆1上;所述探头为X射线荧光光谱仪,用于检测待测区域围岩的元素信息;
所述图像采集装置4环绕自动驾驶车辆1一周设置,用于采集围岩图像且使这些围岩图像可形成一个闭合的环;
所述车载处理器7用于:基于巴尔特-尼格里标准矿物计算法将接收的待测区域围岩的元素信息反演成为矿物信息;
接收围岩图像及对应探头倾角,将同一被测区域的围岩图像转化为一维向量格式的图像信息,再与本身为一维格式的矿物信息拼接,基于预设神经网络对拼接后的信息进行判别以识别岩石岩性,存储岩石岩性识别结果及探头倾角。
其中,因标准矿物是用矿物中阳离数来表示,巴尔特-尼格里标准矿物计算法又称“阳离子标准矿物法”。尼格里指出,岩石中主要级分是硅酸盐造岩矿物,硅酸盐的基本元素是氧,其它元素绝大多数位于周期表前几列,这就决定了它 们在原子量上差别不是很大。如果把阳离子标准矿物分子的大小规定为该矿物除以阳离子数而得的分子量,就保证了主要阳离子标准矿物的分子量大体相等。
例如:
图像采集装置4采用摄像头来实现,对隧道围岩进行全方位摄像,收集围岩图像信息,标明角度,并将所收集到的图像信息存储至车载处理器7。
其中,当自动驾驶车辆1在隧道内行驶的过程中,自动驾驶车辆1移动到预设位置处,而且自动驾驶车辆1上还设置有激光接收器3,激光接收器3用于接收激光发射器2发出的激光,测算并记录位移距离及所在位置,以实现自动驾驶车辆1的定位;所述激光发射器2安置在隧道洞口,用于向隧道内发射激光。
当自动驾驶车辆1在隧道外地面上行驶时,自动驾驶车辆1上设置GPS定位模块,可以实现自动驾驶车辆1的自动定位功能。
在具体实施中,所述机械臂5与控制系统相连,控制系统用于控制机械臂5带动探头6绕自动驾驶车辆1运动并抵至围岩表面。
具体地,机械臂包含4个机械臂关节,通过机械臂关节控制机械臂5旋转和运动。
所述探头6内部设置有压力传感器8和倾角传感器9,所述压力传感器8用于测量探头预围岩表面之间的压力信息并反馈至控制系统,所述倾角传感器9用于检测探头倾角并反馈至控制系统。
其中,控制系统包括控制器及存储器。控制器可采用可编程逻辑器件或CPU等;存储器可采用ROM、RAM或是外置存储设备,比如U盘等。控制器用于接收压力传感器及倾角传感器所传送来的信息并控制机械臂5运动。
在具体实施中,所述车载处理器7,包括:
数据存储部,其用于根据自动驾驶车辆的位置信息将围岩图像依次存储;及根据自动驾驶车辆的位置信息及探头倾角信息,将相应围岩的元素信息依次存储;
数据处理部,其用于反演矿物信息及将同一被测区域的围岩图像和相应矿物信息放置到一起;
岩性识别部,其用于识别岩石岩性。
所述数据处理部,还用于对围岩图像进行剪裁和遮挡等预处理。
其中,所述岩性识别部用于基于B-P神经网络将同一被测区域的围岩图像经过卷积、池化和flatten处理转化为一维向量的图像信息。
所述岩性识别部还用于基于B-P神经网络将一维向量的图像信息与本身就为一维的矿物信息拼接起来,之后将拼接起来的岩石信息传递至全连接层,对其进行判别,智能识别岩石岩性。除B-P神经网络,受限波尔兹曼机、支持向量机等也可完成任务。
本实施例基于岩石图像信息和元素反演的矿物信息,利用预设神经网络对围岩岩性进行智能识别,不仅避免了主观误判,还实现了岩性识别的自动化、智能化,极大地缩短了所耗费的时间,提高了工作效率。
本实施例的基于图像与XRF矿物反演的岩性识别系统的识别方法,包括:
控制机械臂带动探头至待测区域,利用探头检测待测区域围岩的元素信息;
利用图像采集装置采集围岩图像且使这些围岩图像可形成一个闭合的环;
在车载处理器中,基于巴尔特-尼格里标准矿物计算法将接收的待测区域围岩的元素信息反演成为矿物信息;将同一被测区域的围岩图像转化为一维向量 格式的图像信息,再与本身为一维格式的矿物信息拼接,基于预设神经网络对拼接后的信息进行判别以识别岩石岩性,存储岩石岩性识别结果及探头倾角。
具体地,在识别岩石岩性的过程中,基于B-P神经网络将同一被测区域的围岩图像经过卷积、池化和flatten处理转化为一维向量的图像信息,再与本身就为一维的矿物信息拼接起来,之后将拼接起来的岩石信息传递至全连接层,对其进行判别,智能识别岩石岩性。
具体地,基于图像与XRF矿物反演的岩性识别系统的识别具体过程为:
在隧道洞口中心位置安置好激光发射器,自动驾驶车辆通过本身带有的摄像机自动导航、驾驶,通过置于车尾的激光接收器接收所发射的激光进行自动定位;
自动驾驶车辆到达指定区域后,车身安装的若干个摄像头对隧道围岩进行全方位摄像,收集围岩图像信息,标明角度,并将所收集到的图像信息存储至车载处理器;
机械臂外伸,间隔一定角度携带探头至围岩内表面,探头紧贴岩面测量岩石元素信息,可测得该区域若干处岩石元素信息,按照探头倾斜角度的不同分别编号保存;
车载处理器将图像进行预处理,获取有效信息,并依据巴尔特-尼格里标准矿物计算法自动将元素反演成矿物,对经过处理的图像信息和矿物信息进行挑选,将同一区域的信息对应保存;
车载处理器利用同一区域的图像信息和矿物信息对该区域岩性进行智能识别,其将经过卷积、池化、flatten处理后为一维向量格式的图像信息和本身为一维格式的矿物信息拼接起来,通过机器学习的方法,基于B-P神经网络对拼接 后的信息进行判别,以识别岩石岩性,并将标明结果坐标、倾角,保存起来;
自动驾驶车辆移动至下一位置,采集下一区域围岩信息并进行岩性识别。
以上所述仅为本发明的优选实施例而已,并不用于限制本发明,对于本领域的技术人员来说,本发明可以有各种更改和变化。凡在本发明的精神和原则之内,所作的任何修改、等同替换、改进等,均应包含在本发明的保护范围之内。

Claims (10)

  1. 一种基于图像与XRF矿物反演的岩性识别系统,包括自动驾驶车辆,其特征在于,所述自动驾驶车辆上设置有探头、图像采集装置和车载处理器;
    所述探头搭载在机械臂末端,机械臂安装在自动驾驶车辆上;所述探头为X射线荧光光谱仪,用于检测待测区域围岩的元素信息;
    所述图像采集装置环绕自动驾驶车辆一周设置,用于采集围岩图像且使这些围岩图像可形成一个闭合的环;
    所述车载处理器用于:基于巴尔特-尼格里标准矿物计算法将接收的待测区域围岩的元素信息反演成为矿物信息;
    接收围岩图像及对应探头倾角,将同一被测区域的围岩图像转化为一维向量格式的图像信息,再与本身为一维格式的矿物信息拼接,基于预设神经网络对拼接后的信息进行判别以识别岩石岩性。
  2. 如权利要求1所述的基于图像与XRF矿物反演的岩性识别系统,其特征在于,所述自动驾驶车辆上还设置有激光接收器,激光接收器用于接收激光发射器发出的激光,测算并记录位移距离及所在位置;所述激光发射器安置在隧道洞口,用于向隧道内发射激光。
  3. 如权利要求1所述的基于图像与XRF矿物反演的岩性识别系统,其特征在于,所述机械臂与控制系统相连,控制系统与用于控制机械臂带动探头绕自动驾驶车辆运动并抵至围岩表面。
  4. 如权利要求3所述的基于图像与XRF矿物反演的岩性识别系统,其特征在于,所述探头内部设置有压力传感器和倾角传感器,所述压力传感器用于测量探头预围岩表面之间的压力信息并反馈至控制系统,所述倾角传感器用于检测探头倾角并反馈至控制系统。
  5. 如权利要求1所述的基于图像与XRF矿物反演的岩性识别系统,其特征在于,所述车载处理器,包括:
    数据存储部,其用于根据自动驾驶车辆的位置信息将围岩图像依次存储;及根据自动驾驶车辆的位置信息及探头倾角信息,将相应围岩的元素信息依次存储;
    数据处理部,其用于反演矿物信息及将同一被测区域的围岩图像和相应矿物信息放置到一起;
    岩性识别部,其用于识别岩石岩性。
  6. 如权利要求5所述的基于图像与XRF矿物反演的岩性识别系统,其特征在于,所述数据处理部,还用于对围岩图像进行剪裁和遮挡等预处理。
  7. 如权利要求5所述的基于图像与XRF矿物反演的岩性识别系统,其特征在于,所述岩性识别部用于基于B-P神经网络将同一被测区域的围岩图像经过卷积、池化和flatten处理转化为一维向量的图像信息。
  8. 如权利要求7所述的基于图像与XRF矿物反演的岩性识别系统,其特征在于,所述岩性识别部还用于基于B-P神经网络将一维向量的图像信息与本身就为一维的矿物信息拼接起来,之后将拼接起来的岩石信息传递至全连接层,对其进行判别,智能识别岩石岩性。
  9. 一种如权利要求1-8中任一项所述的基于图像与XRF矿物反演的岩性识别系统的识别方法,其特征在于,包括:
    自动驾驶车辆移动至待测区域;
    控制机械臂带动探头至待测区域,利用探头检测待测区域围岩的元素信息;
    利用图像采集装置采集围岩图像且使这些围岩图像可形成一个闭合的环;
    在车载处理器中,基于巴尔特-尼格里标准矿物计算法将接收的待测区域围岩的元素信息反演成为矿物信息;将同一被测区域的围岩图像转化为一维向量格式的图像信息,再与本身为一维格式的矿物信息拼接,基于预设神经网络对拼接后的信息进行判别以识别岩石岩性。
  10. 如权利要求9所述的识别方法,其特征在于,在识别岩石岩性的过程中,基于B-P神经网络将同一被测区域的围岩图像经过卷积、池化和flatten处理转化为一维向量的图像信息,再与本身就为一维的矿物信息拼接起来,之后将拼接起来的岩石信息传递至全连接层,对其进行判别,智能识别岩石岩性。
PCT/CN2020/141564 2020-04-17 2020-12-30 基于图像与xrf矿物反演的岩性识别系统及方法 WO2021208515A1 (zh)

Priority Applications (1)

Application Number Priority Date Filing Date Title
US17/777,700 US11796493B2 (en) 2020-04-17 2020-12-30 System and method for identifying lithology based on images and XRF mineral inversion

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202010305833.4 2020-04-17
CN202010305833 2020-04-17

Publications (1)

Publication Number Publication Date
WO2021208515A1 true WO2021208515A1 (zh) 2021-10-21

Family

ID=72675910

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2020/141564 WO2021208515A1 (zh) 2020-04-17 2020-12-30 基于图像与xrf矿物反演的岩性识别系统及方法

Country Status (3)

Country Link
US (1) US11796493B2 (zh)
CN (1) CN111751394B (zh)
WO (1) WO2021208515A1 (zh)

Families Citing this family (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111751394B (zh) * 2020-04-17 2021-08-27 山东大学 基于图像与xrf矿物反演的岩性识别方法及系统
CN112348014A (zh) * 2020-11-03 2021-02-09 招商局重庆公路工程检测中心有限公司 基于机器视觉的隧道掌子面围岩级别快速识别方法
CN112801035B (zh) * 2021-02-24 2023-04-07 山东大学 基于知识与数据双驱动的搭载式岩性智能识别方法及系统
CN113297962B (zh) * 2021-05-24 2022-08-19 山东大学 无人机搭载式xrf和图像融合的岩性识别系统及方法
CN116105692B (zh) * 2023-02-08 2024-04-05 成都理工大学 用于围岩分级和变形预测的隧道围岩形貌采集装置及方法

Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104769422A (zh) * 2012-09-07 2015-07-08 卡尔蔡司X射线显微镜公司 组合共焦x射线荧光和x射线计算机断层扫描的系统和方法
CN104335032B (zh) * 2012-05-31 2017-07-04 赛默科技便携式分析仪器有限公司 使用组合的x射线荧光和拉曼光谱法的样品分析
CN107620593A (zh) * 2017-08-08 2018-01-23 中国科学院地质与地球物理研究所 一种油气田现场随钻岩屑扫描系统及方法
WO2018098565A1 (en) * 2016-12-02 2018-06-07 National Research Council Of Canada Optical imaging of mineral species using hyperspectral modulation transfer techniques
CN110031493A (zh) * 2019-04-04 2019-07-19 山东大学 基于图像与光谱技术的岩性智能识别系统与方法
CN110346349A (zh) * 2019-07-30 2019-10-18 辽宁石油化工大学 岩屑检测装置
CN110662961A (zh) * 2017-05-15 2020-01-07 沙特阿拉伯石油公司 分析岩石样本
CN111751394A (zh) * 2020-04-17 2020-10-09 山东大学 基于图像与xrf矿物反演的岩性识别方法及系统

Family Cites Families (11)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US8326537B2 (en) * 2004-01-20 2012-12-04 Westerngeco L.L.C. Survey design using earth observation data
CN102620672B (zh) * 2011-12-01 2015-01-14 中国人民解放军63653部队 图像拼接技术在隧道围岩变形非接触测量中的应用
CN105221142B (zh) * 2014-07-04 2018-05-08 中国石油化工股份有限公司 一种识别页岩地层矿物质量含量的方法
CN105115874B (zh) * 2015-08-18 2018-02-02 中国石油天然气股份有限公司 多源信息融合的多组分三维数字岩心构建方法
CN109424360A (zh) * 2017-08-28 2019-03-05 中国石油天然气股份有限公司 确定地层中矿物含量的方法及装置
CN112525093B (zh) * 2018-09-19 2022-07-22 成都理工大学 一种基于双护盾tbm工艺的建立隧道三维模型的系统
CN109710968B (zh) * 2018-11-14 2022-03-01 中国石油天然气股份有限公司 一种基岩潜山裂缝预测方法及装置
CN110043267B (zh) * 2019-04-04 2020-07-31 山东大学 基于岩性与不良地质前兆特征识别的tbm搭载式超前地质预报系统及方法
CN110031491B (zh) * 2019-04-04 2020-05-26 山东大学 车载式岩性与不良地质前兆特征识别系统及方法
CN110686616A (zh) * 2019-10-23 2020-01-14 中南大学 一种tbm隧道围岩形貌采集方法及装置
CN110749609A (zh) * 2019-11-20 2020-02-04 北京九恒质信能源技术有限公司 地下岩石的弹性力学数据确定方法、装置、系统和介质

Patent Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104335032B (zh) * 2012-05-31 2017-07-04 赛默科技便携式分析仪器有限公司 使用组合的x射线荧光和拉曼光谱法的样品分析
CN104769422A (zh) * 2012-09-07 2015-07-08 卡尔蔡司X射线显微镜公司 组合共焦x射线荧光和x射线计算机断层扫描的系统和方法
WO2018098565A1 (en) * 2016-12-02 2018-06-07 National Research Council Of Canada Optical imaging of mineral species using hyperspectral modulation transfer techniques
CN110662961A (zh) * 2017-05-15 2020-01-07 沙特阿拉伯石油公司 分析岩石样本
CN107620593A (zh) * 2017-08-08 2018-01-23 中国科学院地质与地球物理研究所 一种油气田现场随钻岩屑扫描系统及方法
CN110031493A (zh) * 2019-04-04 2019-07-19 山东大学 基于图像与光谱技术的岩性智能识别系统与方法
CN110346349A (zh) * 2019-07-30 2019-10-18 辽宁石油化工大学 岩屑检测装置
CN111751394A (zh) * 2020-04-17 2020-10-09 山东大学 基于图像与xrf矿物反演的岩性识别方法及系统

Also Published As

Publication number Publication date
US20230003674A1 (en) 2023-01-05
CN111751394B (zh) 2021-08-27
CN111751394A (zh) 2020-10-09
US11796493B2 (en) 2023-10-24

Similar Documents

Publication Publication Date Title
WO2021208515A1 (zh) 基于图像与xrf矿物反演的岩性识别系统及方法
WO2021068848A1 (zh) 隧道结构病害多尺度检测与智能诊断系统及方法
WO2021135157A1 (zh) 岩体结构探测及危石探测系统及方法
CN109544679B (zh) 管道内壁的三维重建方法
CN104238566B (zh) 电子电路用基于图像识别的巡线机器人控制系统
WO2021042668A1 (zh) Tbm搭载式隧道围岩结构虚拟再现系统与方法
CN108189040B (zh) 一种污水管线检测机器人系统
CN111178148B (zh) 一种基于无人机视觉系统的地面目标地理坐标定位方法
KR102123983B1 (ko) 시설물 안전 진단용 스마트 드론 및 이를 이용한 시설물 안전 진단 방법
EP3680648A1 (en) Port machinery inspection device and inspection method
CN214520204U (zh) 一种基于深度相机和激光雷达的港区智能巡检机器人
CN114113118B (zh) 一种地铁隧道衬砌裂缝渗漏水病害快速检测装置及其检测方法
WO2021043308A1 (zh) 基于图像识别与分析的隧道内围岩石英含量预测系统与方法
CN115793649A (zh) 一种电缆沟自动巡检装置及巡检方法
CN115743540A (zh) 基于多源信息融合的隧道全息巡检无人机装置及方法
CN212377561U (zh) 一种管道检测机器人
CN112001945B (zh) 一种适用于生产线作业的多机器人监控方法
CN113432529B (zh) 一种基于无人机摄像的震损结构层间残余变形检测方法
CN112925326B (zh) 一种基于激光雷达和深度相机数据融合的agv避障方法
CN112113605A (zh) 基于激光slam与视觉的电缆损坏检测方法及装置
CN210879689U (zh) 一种适用于地铁车辆列检工作的智能机器人
KR101541064B1 (ko) 무선 용접 시스템
US12013485B2 (en) Multi-scale inspection and intelligent diagnosis system and method for tunnel structural defects
CN115032206B (zh) 一种桥梁机器人智能检测方法
KR102204832B1 (ko) 농업용 작물의 생육정보 구축시스템

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: 20931277

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 20931277

Country of ref document: EP

Kind code of ref document: A1

32PN Ep: public notification in the ep bulletin as address of the adressee cannot be established

Free format text: NOTING OF LOSS OF RIGHTS PURSUANT TO RULE 112(1) EPC (EPO FORM 1205A DATED 30.06.2023)

122 Ep: pct application non-entry in european phase

Ref document number: 20931277

Country of ref document: EP

Kind code of ref document: A1