WO2016106958A1 - 一种基于山脊能量校正的山地中带状地下目标的探测方法 - Google Patents

一种基于山脊能量校正的山地中带状地下目标的探测方法 Download PDF

Info

Publication number
WO2016106958A1
WO2016106958A1 PCT/CN2015/072680 CN2015072680W WO2016106958A1 WO 2016106958 A1 WO2016106958 A1 WO 2016106958A1 CN 2015072680 W CN2015072680 W CN 2015072680W WO 2016106958 A1 WO2016106958 A1 WO 2016106958A1
Authority
WO
WIPO (PCT)
Prior art keywords
ridge
sampling
point
latitude
longitude
Prior art date
Application number
PCT/CN2015/072680
Other languages
English (en)
French (fr)
Inventor
张天序
鲁岑
王岳环
杨卫东
桑农
马文绚
郝龙伟
Original Assignee
华中科技大学
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 华中科技大学 filed Critical 华中科技大学
Priority to US15/114,651 priority Critical patent/US9625611B2/en
Publication of WO2016106958A1 publication Critical patent/WO2016106958A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01V—GEOPHYSICS; GRAVITATIONAL MEASUREMENTS; DETECTING MASSES OR OBJECTS; TAGS
    • G01V9/00—Prospecting or detecting by methods not provided for in groups G01V1/00 - G01V8/00
    • G01V9/005—Prospecting or detecting by methods not provided for in groups G01V1/00 - G01V8/00 by thermal methods, e.g. after generation of heat by chemical reactions
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01C—MEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
    • G01C21/00—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00
    • G01C21/005—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00 with correlation of navigation data from several sources, e.g. map or contour matching
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V20/00—Scenes; Scene-specific elements
    • G06V20/10—Terrestrial scenes
    • G06V20/13—Satellite images
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V20/00—Scenes; Scene-specific elements
    • G06V20/10—Terrestrial scenes
    • G06V20/194—Terrestrial scenes using hyperspectral data, i.e. more or other wavelengths than RGB

Definitions

  • the invention belongs to the intersection field of pattern recognition, remote sensing technology and terrain analysis, and particularly relates to a method for detecting a strip-shaped underground target in a mountain based on ridge energy correction, which is to locate the ridge position by terrain analysis and correct the ridge energy to improve
  • the correct recognition rate of underground targets in the mountain environment also reduces the false alarm rate.
  • the energy absorbed by the soil and rocks absorbs sunlight and generates heat.
  • the infrared radiation is detected by a thermal infrared sensor.
  • the mountain thermal field is mainly divided into a stable part and a variable part, and the variable part is affected by sunlight.
  • the shallow mountain with severe temperature changes, the stable part includes the mountain part below the shallow mountain and the underground target buried in it.
  • the sun heats the variable part of the mountain daily, the heat exchange between the stable part of the mountain and the variable part, and the heat exchange and stable partial heat exchange of the underground target eventually lead to the detectable temperature difference between the mountain temperature and the buried target.
  • the physical basis The physical basis.
  • the temperature and energy of the belt-shaped underground target are different from those of the surrounding mountain medium. After heat transfer and diffusion, the pulse pattern appears in the mountain. However, due to the heat island effect of the ridge, the energy field at the ridge position is also in line with the pulse mode, which interferes with the detection of the banded underground target in the mountain.
  • the present invention provides a ridge line detection algorithm for determining the position of a ridge by detecting a strip-shaped subsurface target in a mountainous environment using a pulse pattern, which is always affected by the ridge heat island effect.
  • the ridge energy is corrected, and the corrected energy image is used to detect and locate the strip-shaped underground target, thereby solving the high false alarm problem caused by the ridge in the mountain environment.
  • the method for detecting a strip target in a mountain based on ridge energy correction mainly comprises the following steps:
  • the step of obtaining terrain digital elevation information including the following sub-steps:
  • the soil and air environment are homogeneous within a certain range.
  • the resolution of the latitude and longitude matrix must be the same as the resolution of the energy (infrared) image, so that only the last acquired terrain value elevation information can correspond to the energy (infrared) image.
  • the position of the ridge is detected by the terrain digital elevation information and finally the position on the corresponding energy (infrared) map is found.
  • the width of the rectangular coverage is calculated by the ranging tool provided by Google Earth, the height is marked as width, the height is in meters, and the range of latitude and longitude is calculated.
  • each time sampling step meters calculate the number of vertical and horizontal sampling points, namely: height/step, width/step, and calculate the longitude of each adjacent two sampling points of the latitude and longitude matrix.
  • the step length long_step and the latitude step length lati_step are used to calculate the latitude and longitude of each sample point in the latitude and longitude matrix.
  • Google Earth provides a programming interface that allows us to use Google Earth to obtain elevation data for each sample point using the latitude and longitude matrix (1.2.2) as input, and generate a terrain digital elevation information matrix output based on the altitude data.
  • the noise may come from inaccurate elevation data or the high and low of the whole block generated by Google Earth during splicing.
  • the high and low of the whole block does not affect the detection of the ridgeline. Therefore, for the individual inaccurate noise, we use the mean filtering method to denoise the original terrain digital elevation information. That is, the mean value of the elevation information in the neighborhood of a certain local range, such as a k*k, is taken as an output. This eliminates the effects of randomly distributed noise because: the true elevation value of the i-th sample point in a neighborhood is h i , its noise error is ⁇ h i , and the final observation is h i + ⁇ h i .
  • the process of taking the mean of a pixel in a neighborhood as an output is as follows:
  • the noise is randomly distributed, the average noise of a plurality of sampling points is 0, so that an elevation map that approximates the real situation can be obtained.
  • the ridge line detection step includes the following sub-steps:
  • the ridge line in the horizontal direction is the ridge line in the x direction: traverse along the x direction, and the elevation value of each sampling point and the elevation value of the sampling point within a certain range in the y direction (for example, within 5 sampling points, ie, within 50 meters)
  • a certain range in the y direction for example, within 5 sampling points, ie, within 50 meters
  • the sampling point is a maximum value in the y direction
  • the point is considered to be a candidate point of the ridge line in the horizontal direction, and the certain range may be set in advance.
  • the ridge line in the vertical direction that is, the ridge line in the y direction: traverse along the y direction, and the elevation value of each sampling point and the elevation value of the sampling point within a certain range of the x direction (for example, within 5 sampling points, ie, within 50 meters)
  • a certain range of the x direction for example, within 5 sampling points, ie, within 50 meters
  • the sampling point is a maximum value in the x direction
  • the point is considered to be a candidate point of the ridge line in the vertical direction, and the certain range may be set in advance.
  • Normal ridge line points should have a certain continuity and should not be isolated, but the candidate points of the ridgeline extracted according to steps (3.1) and (3.2) may be discontinuous isolated points, so we should treat each A ridgeline candidate point is continuously distinguished.
  • the criterion is as follows: if the total number of ridge line candidate points in the t*t neighborhood of the candidate ridge line point is greater than th_num, the ridge line candidate point is the final ridge line point; otherwise, the isolated ridge line candidate point is determined to be non- Ridge line points, where t is the default value. This will result in the final continuous ridgeline marking.
  • the ridge position energy correction step includes the following sub-steps:
  • the mountain effect is common in the ridge position: mainly refers to the thermal efficiency of the uplift block.
  • the mountain creates the climate around it.
  • the larger the surface area of the uplift the greater the impact of the mountain on itself and the surrounding environment.
  • the solar radiation is absorbed and converted into thermal energy, which is much higher than the temperature of the free atmosphere at the same altitude.
  • the energy of the ridge position is corrected.
  • the specific method of correction is: the energy of the sampling point of the ridge line position is replaced by the average value of the energy of the sampling points on both sides of the ridge line position.
  • the strip underground target detection step includes the following sub-steps:
  • the infrared image of the mountainous region where the band-shaped underground target may exist is traversed, and the position where the pulse mode exists is detected, thereby realizing the positioning of the belt-shaped underground target while detecting the position of the false alarm.
  • the following parameters need to be set before the traversal detection starts:
  • the sampling segment size and the interval between the central segment of the contrast segment It is assumed that the length and width of the sampling segment of the mountain surface with the banded underground target are l and w pixels respectively; then, the sampling segment of the mountain surface on both sides of the banded underground target is also adopted. The same size.
  • the distance between the center of the sampling section of the mountain surface and the center of the sampling section above the strip underground target is s pixels, where l, w, s are preset values.
  • Each search for a set of mountainous terrain images with a banded underground target below and a mountainous surface image contrasted on both sides translates the distance d to both sides and continues a new set of searches. The search stops automatically until the boundary of the image is reached.
  • Impulse threshold the absolute value of the difference between the gray mean value of the mountain ground surface image sampling segment with the banded underground target and the gray mean value of the mountain ground surface image sampling segment on both sides of each segment is greater than the pulse threshold th
  • the pulse at this point is counted as a valid pulse; if any of the two differences is less than the pulse threshold, the pulse is considered invalid because the signal is weak.
  • the sampling segment is scrolled pixel by pixel from point P0(x0, y0) to point P1 (x1, y1).
  • the sampling segment is scrolled pixel by pixel from point P0(x0, y0) to point P1 (x1, y1).
  • shifting r pixels to the left and right searching for the position where the difference between the middle segment and the left and right sides is the largest, and testing whether the pulse at the position is a valid pulse, and if it is a valid pulse, the effective pulse
  • the number is increased by one.
  • P0 is calculated, and P1 is translated to the two sides by the coordinates P0', P1' after each extended distance, and is counted between P0' and P1' according to the same sampling method between P0 and P1, and the number of effective pulses is counted.
  • the number of effective pulses accounts for the maximum number of two-end coordinates P0, and the position determined by P1 is the strip-shaped underground target position. The position where the remaining pulses appear in the result is the false alarm position.
  • the technical effect of the invention is embodied in: after research and experiment, it is found that the false alarm of the detection of the belt-shaped underground target in the mountain environment mostly appears in the ridge position, and the simulation of the mountain temperature field also realizes the existence of the ridge effect. Therefore, we propose to reduce the position of the ridge and correct the energy of the ridge to reduce the detection method of the ridge-based energy-corrected mountainous underground targets in the false alarm.
  • the experimental results show that the method can significantly reduce the detection of the banded underground targets in the mountains.
  • the false alarm appearing in the ridge position makes the detection result more accurate, and the method is simple to implement, the calculation amount is small, and the required parameters less.
  • FIG. 1 is a schematic flow chart of a method for detecting a strip-shaped underground target in a mountain based on ridge energy correction according to the present invention
  • FIG. 2 is a schematic diagram of coverage of terrain digital elevation information in an embodiment of the present invention.
  • FIG. 3 is a topographical digital elevation information map obtained in an embodiment of the present invention.
  • FIG. 4 is a diagram showing a result of denoising preprocessing of a terrain digital elevation information map in an embodiment of the present invention
  • FIG. 5 is a flow chart of an initial detection algorithm of a ridge line in an embodiment of the present invention.
  • Figure 6 is a diagram showing the initial detection result of the ridge line in the embodiment of the present invention.
  • FIG. 7 is a flow chart of a continuous ridge line extraction algorithm in an embodiment of the present invention.
  • Figure 8 is a diagram showing the results of continuous ridge line extraction in the embodiment of the present invention.
  • FIG. 10 is a flow chart of a ridge energy correction algorithm in an embodiment of the present invention.
  • 11 is an original infrared image obtained by simulation in an embodiment of the present invention.
  • Figure 12 is a result of marking of a ridge position on a simulated infrared image in an embodiment of the present invention
  • Figure 13 is a diagram showing the results of ridge energy correction in the embodiment of the present invention.
  • FIG. 14 is a diagram showing a false alarm flag before ridge energy correction in an embodiment of the present invention.
  • Figure 15 is a diagram showing the detection of false alarms after ridge energy correction in an embodiment of the present invention.
  • the strip-shaped underground target for illustrating the detection and identification method of the strip-shaped underground target in the mountain environment based on the ridge energy correction is a tunnel in a mountain environment, and the energy map of the region is based on the altitude information and the surface material.
  • Infrared image obtained by simulation of infrared radiation characteristics.
  • the algorithm of the present invention is applied for detection, the same effect can be obtained by using a real thermal infrared image instead of the energy map mentioned in the present invention.
  • the invention firstly proposes the use of infrared imaging technology and multi-information to detect the banded underground objects in the mountains.
  • the standard method aims to solve the problem of detecting the false alarm height of the banded underground target in the mountain by detecting the position of the ridge to correct the energy at the ridge.
  • the detection of the ridge position belongs to the field of terrain analysis, that is, the position of the ridge line is automatically extracted by using the terrain information contained in the terrain elevation data.
  • the method of extracting ridge lines from 3D elevation data can be divided into local algorithms and overall algorithms in principle.
  • the local algorithm analyzes the vertical and horizontal sections of the digital elevation grid, finds the elevation maxima on the section as the candidate points on the ridgeline, and then filters and sorts the candidate points that have been obtained according to certain rules.
  • the ridgeline typical algorithms such as: section analysis; the overall algorithm is to simulate the state of the surface of the natural flow of water, find the watershed.
  • the local algorithm can not estimate the overall variation of the terrain. It is difficult to distinguish the terrain noise when determining the ridgeline.
  • the focus of the present invention is to correct the energy of the position of the ridge by detecting the position of the ridge, the purpose of reducing the false alarm rate of the detected underground target and improving the recognition rate is achieved.
  • the invention proposes a ridge detection method with small calculation amount and fast calculation speed, and on the basis of this, the method of ridge energy correction is expounded, thereby achieving the purpose of accurately detecting and locating the belt-shaped underground target by using the pulse mode.
  • the invention provides a method for detecting a strip-shaped underground target in a mountain environment based on ridge energy correction, as shown in FIG. 1 , comprising the following five main steps: (1) obtaining a step of terrain digital elevation information; and (2) digital elevation Information denoising preprocessing step; (3) ridge line detection step; (4) ridge position energy correction step; (5) banded underground target detection step to specifically describe the execution process of the algorithm:
  • the step of obtaining terrain digital elevation information including the following sub-steps:
  • the above-described tunnel as an illustrative example has a length of 3000 meters in order to cover the entire strip shape.
  • the next goal is also to consider that the complexity of using Google Earth to obtain terrain digital elevation information is proportional to the area of the selected area. We finally determine the size of the area to be detected, and we determine its location after further review.
  • the specific location of the area to be detected that is, the latitude and longitude information of the four vertices P1, P2, P3, and P4 of the area to be detected is as follows:
  • Pt1 116.150049, 40.296833
  • Pt2 116.0292983, 40.356959
  • the energy (infrared) image used in the illustrative examples of the present invention was simulated based on infrared radiation characteristics with a resolution of 10 meters.
  • the resolution of the latitude and longitude matrix must be the same as the resolution of the energy (infrared) image, so the sampling interval step here and the latitude matrix is 10 meters.
  • the latitude of the sampling point in the i-th row and the j-th column of the latitude and longitude matrix is
  • Locate(i,j) (locate(i,j)_longti,locate(i,j)_lati).
  • Google Earth provides a programming interface to arrange the coordinates of each sample point in the latitude and longitude matrix locate in (1.2.2) into a vector, that is, the i+1th line of locate is placed one after the i-th row. Vector as input.
  • Google Earth will automatically read each The latitude and longitude data of a sample point, and return the altitude data corresponding to each sample point. We only need to re-route the returned altitude data vector according to each row of long_num sample points, and the total lati_num line is output as a matrix, which is the terrain digital elevation information.
  • the terrain digital elevation information in this example is shown in Figure 3.
  • the topographic digital elevation information we will obtain will have some noise due to the digital elevation information matrix obtained in (1.3), as indicated by the black rectangle in Figure 2, which may come from inaccurate elevation data. Therefore, for these noises, the mean terrain filtering method is used to denoise the original terrain digital elevation information. That is, traversing the whole graph, the value of each sampling point is replaced by the mean value of the elevation information in a neighborhood within a certain local range, such as an s*s. This eliminates the effects of randomly distributed noise.
  • the ridge line detection step includes the following sub-steps:
  • FIG. 5 The overall flow chart of the ridge line initial detection algorithm is shown in Figure 5, which is divided into horizontal ridges. Line detection (3.1) and vertical ridge line detection (3.2) are two sub-steps. The continuous ridgeline removal is performed after the initial detection of the ridgeline (3.3).
  • a continuous ridge line extraction algorithm is proposed, and the algorithm flow chart is shown in FIG. 7.
  • the candidate points of the ridge line in the initial detection result of the ridge line are judged whether it is a discontinuous isolated point or a certain area (for example, within the neighborhood of t*t), the number of ridge line candidate points is less than a certain threshold th_num, and if so, it is judged
  • the candidate points of the ridge line are non-ridge line points to prevent false alarms.
  • the candidate points of the ridge line are non-ridge line points.
  • the continuous ridgeline extraction results marked on the terrain digital elevation map are shown in Figure 8.
  • the ridge position energy correction step includes the following sub-steps:
  • the mountain effect is common in ridge locations: the temperature at the ridge is higher than the temperature at the mountainside.
  • This paper cites the simulation results of the mountain temperature field for the unbanded underground target, as shown in Figure 9. Show. It can be seen that the temperature of the ridge in the center of the mountain is higher than the temperature at the mountainside.
  • the nearest neighbor interpolation method is used to correct the energy at the ridge line.
  • the algorithm flow chart is shown in Figure 10. It is mainly divided into the following two steps:
  • the energy of the ridge line point is replaced by the energy average corresponding to the non-ridge line points in the four neighborhoods, and finally the ridge energy corrected result is obtained.
  • a simulated infrared image (energy image) without energy correction is shown in FIG.
  • the detected ridgeline position result is marked on the simulated infrared image as shown in FIG.
  • the infrared simulation image of the ridge line position energy correction is shown in Fig. 13. Comparing Fig. 11 with Fig. 13, the bright areas of multiple places are significantly reduced, achieving the purpose of energy correction of the ridge position.
  • the false alarms appearing at these locations during the detection of the strip-shaped underground target in step (5) are reduced.
  • the strip underground target detection step includes the following sub-steps:
  • search direction and traverse the center of the search for the first and last coordinates P0 (10, 27), P1 (283, 171), the search direction is approximately 120 °;
  • the false alarm marks obtained by detecting the infrared simulation map before and after the above correction are shown in Fig. 14 and Fig. 15, and the false alarm rate is reduced from 24.7% to 21%.
  • the position of the ridges marked in the box in Fig. 12 is reduced in Fig. 14 after ridge energy correction in Fig. 3 compared to Fig. 3 in which ridge energy correction is not performed.

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Remote Sensing (AREA)
  • Radar, Positioning & Navigation (AREA)
  • Chemical & Material Sciences (AREA)
  • Chemical Kinetics & Catalysis (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • General Life Sciences & Earth Sciences (AREA)
  • Geophysics (AREA)
  • Theoretical Computer Science (AREA)
  • Multimedia (AREA)
  • Astronomy & Astrophysics (AREA)
  • Automation & Control Theory (AREA)
  • Geophysics And Detection Of Objects (AREA)

Abstract

一种通过检测山地环境中山脊位置并进行能量校正来提高对山地环境中带状地下目标的探测、识别、定位方法。该方法属于模式识别、遥感技术、地形分析的交叉领域。带状地下目标的热场与山体的热场不同能产生能量异常，而山脊的热岛效应也会造成山体能量异常，但该异常本质上与带状地下目标的能量异常模式不同，所以通过消除地形中的山脊所产生的热体效应对带状地下目标表现出的微弱能量异常模式的影响，达到降低山地环境中带状地下目标的探测识别虚警率的效果。本方法包括获取地形数字高程信息步骤，数字高程信息去噪预处理步骤，山脊线检测步骤，山脊位置能量校正步骤和带状地下目标探测步骤。

Description

一种基于山脊能量校正的山地中带状地下目标的探测方法 【技术领域】
本发明属于模式识别、遥感技术、地形分析的交叉领域，具体涉及一种基于山脊能量校正的山地中带状地下目标的探测方法，该方法通过地形分析定位山脊位置并对山脊能量进行校正从而提高山地环境中地下目标的识别正确率同时降低虚警率。
【背景技术】
山地环境中普遍存在着大量的带状目标，如自然环境中存在的地下河流、人工修建的地下石油运输管道、穿山的铁路与公路隧道等。地下河流既是重要水资源又是在山地中进行建筑施工时必须避绕开的，所以如何准确探测、定位出地下河流的位置对我们的可持续发展与现代化进程都有着重大意义。公路隧道与铁路隧道可以穿过山体不仅大大缩短的道路的长度节省了人们在旅途中话费的时间，也降低了修建盘山公路、铁路所需要大量人力、物力，并且对于汽车来说，隧道的安全性远远大于盘山公路的安全性。但是一旦这些山地环境中的人工地下建筑出现故障，就面临着勘察位置难度大的问题。所以能够准确的探测、定位这些地下带状目标的位置对人们的交通、生活的方方面面都有巨大的影响。因此，有必要开展以较低虚警率和较高识别率来探测定位山地环境中带状地下目标的研究。
当然，最普通和直接的对隧道设施探测的方法是接触式人工探测，但是，这种方法十分费时且需要大量人力物力。虽然，红外成像作为带状地下目标探测新技术被提出应用于浅层地下管线的探测，但是还未见国内外探测较深埋深的带状地下目标探测的报道。
土壤和岩石吸收太阳光的能量产生热量发出红外辐射被热红外传感器探测。山体热场主要分为稳定部分与可变部分，可变部分即为受日照影响 温度变化剧烈的浅层山体，稳定部分包括浅层山体以下的山体部分与埋藏在其中的地下目标。太阳每日对山体可变部分循环加热，山体内部稳定部分与可变部分热交换以及地下目标自身产热与稳定部分热交换最终导致山体温度与埋藏目标的可探测的温差，这是探测地下目标的物理基础。
带状地下目标的温度与能量与周围的山体介质有一定不同，经过热传递与扩散最终在山体呈现出脉冲模式。但是由于山脊的热岛效应使得山脊位置的能量场也比较符合脉冲模式，对山地中带状地下目标的探测造成干扰。
【发明内容】
针对利用脉冲模式探测山地环境中带状地下目标方法总是受到山脊热岛效应的而影响伴有较高的虚警率的缺陷，本发明提供了一种通过山脊线检测算法确定山脊位置，并对山脊能量进行校正，再利用校正后的能量图像进行带状地下目标的探测、定位，由此解决了山地环境中山脊造成的高虚警问题。本发明的基于山脊能量校正的山地中带状目标的探测方法主要包括下述步骤：
(1)获取地形数字高程信息步骤，包括下述子步骤：
(1.1)确定数字高程信息覆盖范围的经纬度步骤：
土壤和空气环境在一定范围内是同质的。首先我们就应该确定到底获得多大范围内的地形的数值高程信息，并且确定该探测范围的位置，即经纬度信息。由于下面的步骤中还要确定每一个点的经纬度信息，所以我们最好确定一个标准的矩形区域，而这里只要确定矩形四个顶点的经纬度信息，分别标记为Pt1(longti1,lati1)、Pt2(longti2,lati2)、Pt3(longti3,lati3)、Pt4(longti4,lati4)。
(1.2)计算范围内经纬度矩阵的步骤：
经纬度矩阵的分辨率必须和能量(红外)图像的分辨率相同，只有这样最后获取的地形数值高程信息才可以和能量(红外)图像对应起来，从 而实现通过地形数字高程信息检测山脊位置并最终找到对应能量(红外)图上的位置。
(1.2.1)计算地形范围的宽度、高度步骤：
根据(1.1)中确定的矩形四个顶点的经纬度利用Google Earth提供的测距工具计算出矩形覆盖范围的宽度、高度标记为width、height单位为米，并计算出经纬度的范围
(1.2.2)计算经纬度矩阵的步骤：
按照采样间隔step米,每次距离step米采样一次，分别计算出纵向、横向的采样点的数目，即：height/step，width/step，并计算出经纬度矩阵每相邻两个采样点的经度步长long_step和纬度步长lati_step，从而计算出经纬度矩阵中每一个采样点的经纬度。
(1.3)利用Google Earth计算海拔矩阵的步骤：
Google Earth提供了编程接口，使得我们可以将(1.2.2)中经纬度矩阵作为输入利用Google Earth得到每一个采样点的海拔数据，并根据该海拔数据生成地形数字高程信息矩阵输出。
(2)数字高程信息去噪预处理步骤，包括下述子步骤：
由于(1.3)中获得的数字高程信息矩阵会带有一定的噪声，该噪声可能来自于不准确的高程数据或者是Google Earth在拼接时产生的整块的偏高与偏低。整块的偏高偏低是不影响山脊线的检测的，所以针对个别不准确噪声，我们采用均值滤波方法对原始地形数字高程信息进行去噪预处理。即取一定局部范围内，如一个k*k的邻域内高程信息的均值作为输出。这样就可以消除呈随机分布的噪声的影响，因为：设一个邻域内第i个采样点的真实高程值为hi，他的噪声误差为□hi，最终的观测值为hi+□hi，取一个邻域内像素的均值作为输出的过程如下：
Figure PCTCN2015072680-appb-000001
即由于噪声是随机分布的，所以多个采样点的平均噪声为0，这样就可以得到逼近真实情况的高程图。
(3)山脊线检测步骤，包括下述子步骤：
(3.1)水平方向山脊线的检测步骤：
水平方向的山脊线即x方向的山脊线：沿着x方向遍历，将每一个采样点的高程值与y方向一定范围内(如5个采样点即50米范围内)的采样点的高程值比较，如果该采样点在y方向是极大值，就认为该点为水平方向的山脊线的候选点，所述一定范围可以预先设定。
(3.2)垂直方向山脊线的检测步骤：
垂直方向的山脊线即y方向的山脊线：沿着y方向遍历，将每一个采样点的高程值与x方向一定范围内(如5个采样点即50米范围内)的采样点的高程值比较，如果该采样点在x方向是极大值，就认为该点为垂直方向的山脊线的候选点，所述一定范围可以预先设定。
(3.3)提取连续的山脊线步骤：
正常的山脊线点应该有一定的连续性而不应该是孤立的，但是按照(3.1)、(3.2)步骤提取出的山脊线的候选点有可能是不连续的孤立点，所以我们应该对每一个山脊线候选点进行连续性判别。判别准则如下：如果该候选山脊线点的t*t邻域内山脊线候选点的总数大于th_num，则该山脊线候选点为最终的山脊线点；否则，判断该孤立的山脊线候选点为非山脊线点，其中t为预设值。由此就可以得到最后的连续山脊线标记图。
(4)山脊位置能量校正步骤，包括下述子步骤：
(4.1)山脊位置能量分布特征分析步骤：
山脊位置普遍存在山体效应：主要指隆起地块的热力效力。山体创造其周围的气候，在任给定的海拔高度上，隆起地块的表面积越大，山体对其本身和周围环境的影响也就越大。山体作为突起的热岛，太阳辐射被吸收并转换成长波热能，其温度远高于相同海拔自由大气的温度。
另一方面，从传热学角度分析：热量总是沿着最容易传导的方向进行传播。而山体中的岩石的热导率为1.2～2.1W/(m·℃)，而相比之下山体表面接触的外界空气的热导率为0.024W/(m·℃)，所以，山体内部的热量遇到空气时，由于其导热率远远小于岩石之间的热导率，热量大部分沿着岩石传导并在山脊处汇集，产生了山脊处热量明显偏高的能量分布特征。
(4.2)山脊位置能量校正步骤：
根据(4.1)中阐述的山脊位置能量分布特征，对山脊位置的能量进行校正，校正的具体方法为：山脊线位置采样点的能量用山脊线位置两侧的采样点的能量的均值代替。通过校正山脊位置能量，可以有效的减少探测识别虚警。
(5)带状地下目标探测步骤，包括下述子步骤：
(5.1)设置遍历探测带状地下目标所用参数步骤：
对带状地下目标可能存在的山地区域的红外图像进行遍历，探测脉冲模式存在的位置，从而实现带状地下目标的定位的同时探测出虚警的位置。在遍历探测开始前有以下参数需要进行设置：
采样段大小以及对比段距中心段的间隔：假设下方有带状地下目标的山地地表的采样段的长度与宽度各为l和w个像素；那么，带状地下目标两旁山地地表采样段也采用同样大小。两旁山地地表采样段中心距离带状地下目标上方采样段中心的距离为s个像素，其中l、w、s为预设值。
搜索的方向：通过查阅资料可以大致估计出该地区带状地下目标的走 向，按照估计出的走向度数设置遍历搜索的中心首尾坐标P0(x0,y0),P1(x1,y1)，这两点的连线只要可以通过该区域中点附近的位置即可，因为这样方便向两边平移遍历搜索。
每次延展的距离：每搜索一组假设下方存在带状地下目标的山地地表图像与两边对比的山地地表图像就向两边平移距离d再继续新的一组的搜索。直到到达图像的边界，搜索自动停止。
脉冲阈值：每一段中间假设下方存在带状地下目标的山地地表图像采样段的灰度均值与两边对比的山地地表图像采样段的灰度均值的差的绝对值，均大于该脉冲阈值th的时候，该处的脉冲才算作有效脉冲；若两个差值中任意一个小于脉冲阈值该处的脉冲就是因为信号微弱而被认为无效。
(5.2)遍历探测结果的输出步骤：
由(5.1)中确定的遍历搜索的中心首尾坐标P0(x0,y0),P1(x1,y1)，开始使采样段从点P0(x0,y0)逐像素滚动到点P1(x1,y1)，且每次滚动一个像素后，再左右各平移r个像素，寻找其中中间段与左右两边对比段差值最大的位置，并测试该位置的脉冲是否为有效脉冲，若为有效脉冲，则效脉冲数目增加一。接着计算P0,P1向两边平移每次延展的距离之后的坐标P0’,P1’，在P0’,P1’之间按照P0,P1之间同样的采样方法进行统计，并并计数有效脉冲数目。有效脉冲数目占总数目最大的一组两端坐标P0,P1确定的位置为带状地下目标位置，结果中其余脉冲出现的位置即为虚警位置。
本发明的技术效果体现在：经过研究试验发现山地环境中带状地下目标探测的虚警多出现在山脊位置，经过山体温度场仿真也真实了山脊效应的存在。所以，我们提出通过检测山脊位置并校正山脊位置的能量来降低虚警的基于山脊能量校正的山地中带状地下目标的探测方法，试验结果表明该方法确实能够明显降低山地中带状地下目标探测出现在山脊位置的虚警，使得探测结果更加准确，且该方法实现简单、计算量较小、所需参数 较少。
【附图说明】
图1是本发明基于山脊能量校正的山地中带状地下目标的探测方法的流程示意图；
图2是本发明实施例中地形数字高程信息覆盖范围示意图；
图3是本发明实施例中获得地形数字高程信息图；
图4是本发明实施例中地形数字高程信息图去噪预处理结果图；
图5是本发明实施例中山脊线初检测算法流程图；
图6是本发明实施例中山脊线初检测结果图；
图7是本发明实施例中连续山脊线提取算法流程图；
图8是本发明实施例中连续山脊线提取结果图；
图9是本发明实施例中无带状地下目标的山体温度场仿真结果图；
图10是本发明实施例中山脊能量校正算法流程图；
图11是本发明实施例中仿真得到的原始红外图；
图12是本发明实施例中山脊位置在仿真红外图上的标记结果；
图13是本发明实施例中山脊能量校正结果图；
图14是本发明实施例中山脊能量校正前检测虚警标记图；
图15是本发明实施例中山脊能量校正后检测虚警标记图。
【具体实施方式】
本发明中用于说明基于山脊能量校正的山地环境中的带状地下目标探测识别方法的带状地下目标是一条山地环境中的隧道，其所在地区的能量图是我们根据海拔信息与表面材质的红外辐射特征仿真得到的红外图像。在应用本发明的算法做检测的时候，可以利用真实的热红外图像代替本发明中提到的能量图会得到一样的效果。
本发明首次提出利用红外成像技术与多元信息探测山地中带状地下目 标的方法，旨在解决通过检测山脊的位置对山脊处的能量进行校正从而山地中带状地下目标探测虚警高的难题。
山脊位置的检测属于地形分析的领域，即利用地形高程数据中含有的地形信息自动提取山脊线的位置。从三维高程数据中提取山脊线的方法从原理上可以分为局部算法和整体算法。局部算法是分析组成数字高程网格的纵横断面，找出断面上的高程极大值点作为山脊线上的候选点，然后再根据一定的规则对已经得到的候选点进行筛选、排序得到所需要的山脊线，典型算法如：断面分析法；整体算法是模拟地形表面自然流水的状态，找出分水线。然而局部算法不能够估计地形的整体变化规律，在确定山脊线的时候很难区分地形噪音，这样会在其所提取的山脊线的候选点上有大量的噪音，为后续山脊线的判别带来不便，甚至会产生错误和使得后续算法无法进行。整体算法有较强的抗噪音能力，但是其计算量大，且随数字高程网格的增加成平方关系增长。
由于本发明的重点是通过检测山脊位置对山脊位置的能量进行校正从而达到降低探测带状地下目标的虚警率、提高识别率的目的。本发明提出了一个计算量小、计算速度快的山脊检测方法，并在此基础上阐述了山脊能量校正的方法，从而达到了利用脉冲模式准确探测、定位带状地下目标的目的。
本发明提供了一种基于山脊能量校正的山地环境中的带状地下目标探测方法，如图1所示，包括如下五个主要步骤：(1)获取地形数字高程信息步骤；(2)数字高程信息去噪预处理步骤；(3)山脊线检测步骤；(4)山脊位置能量校正步骤；(5)带状地下目标探测步骤来具体阐述其算法的执行过程：
(1)获取地形数字高程信息步骤，包括下述子步骤：
(1.1)确定数字高程信息覆盖范围的经纬度步骤：
上述作为说明示例的隧道的长度为3000米，为了能够覆盖整个带状地 下目标还要考虑到利用Google Earth获取地形数字高程信息的复杂度与选定的区域的面积成正比，我们最终确定了待探测区域的大小，经过进一步查阅资料我们确定了其位置。
待探测区域具体的位置，即：待探测区域四个顶点P1，P2、P3、P4的经纬度信息如下：
Pt1(116.150049,40.296833)、Pt2(116.0292983,40.356959)、
Pt3(116.194775,40.260787)、Pt4(115.970548,40.311917)；四个顶点所标定的具体方法如图2所示。
(1.2)计算范围内经纬度矩阵的步骤：
本发明说明示例所使用的能量(红外)图像是根据红外辐射特征仿真得到的，其分辨率为10米。而经纬度矩阵的分辨率必须和能量(红外)图像的分辨率相同，所以这里及纬度矩阵的采样间隔step为10米。
(1.2.1)计算待探测地形范围的宽度、高度步骤：
根据(1.1)中确定的矩形四个顶点Pt1，Pt2、Pt3、Pt4的经纬度利用Google Earth提供的测距工具得到矩形覆盖范围的宽度width＝3800米、高度height＝4000米。
(1.2.2)计算经纬度矩阵的步骤：
按照采样间隔,每次距离10米采样一次，分别计算出纵向(纬度方向)采样点的数目：
lati_num＝height/step＝4000/10＝400；
横向(经度方向)的采样点的数目：
long_num＝width/step＝3800/10＝380。
并计算出经度范围
long_region
＝longti2-longti4
＝116.0292983-115.970548
＝0.0587503
和纬度范围
lati_region
＝lati2-lati4
＝40.356959-40.311917
＝0.045042。
经纬度矩阵每相邻两个采样点的经度步长
long_step
＝long_region/long_num
＝0.0587503/380
＝0.0001546
和纬度步长
lati_step
＝lati_region/lati_num
＝0.045042/400
＝0.000112605
那么，经纬度矩阵locate的第i行第j列的采样点的经度为
locate(i,j)_longti＝longti4+long_step*j；
经纬度矩阵第i行第j列的采样点的纬度为
locate(i,j)_lati＝lati4+lati_step*(lati_num-i)
则经纬度矩阵第i行第j列的采样点的经纬度为
locate(i,j)(locate(i,j)_longti，locate(i,j)_lati)。
(1.3)利用Google Earth计算海拔矩阵的步骤：
Google Earth提供了编程接口，将(1.2.2)中经纬度矩阵locate中每一个采样点的坐标按行排列为一个向量，即locate的第i+1行放在第i行的后面一次排列成一个向量作为输入。Google Earth会自动依次读取每 一个采样点的经纬度数据，并返回每一个采样点对应的海拔数据。我们只需要将返回的海拔数据向量重新按照每行long_num个采样点，共lati_num行输出为一个矩阵，该矩阵即为地形数字高程信息。将本实例中的地形数字高程信息展示如图3所示。
(2)数字高程信息去噪预处理步骤，包括下述子步骤：
我们将获得的地形数字高程信息因为(1.3)中获得的数字高程信息矩阵会带有一定的噪声，如图2中黑色矩形框所标注的地方，这些噪声可能来自于不准确的高程数据。所以针对这些噪声，采用均值滤波方法对原始地形数字高程信息进行去噪预处理。即遍历全图，每一个采样点的值由取一定局部范围内，如一个s*s的邻域内高程信息的均值代替。如此可以消除呈随机分布的噪声的影响。
去噪预处理算法的解释：设一个邻域内第i个采样点的真实高程值为hi，他的噪声误差为□hi，最终的观测值为hi+□hi，取一个邻域内像素的均值作为输出的过程如下：
Figure PCTCN2015072680-appb-000002
即由于噪声是随机分布的，所以多个采样点的平均噪声为0，这样就可以得到逼近真实情况的高程图。因为本实例中取n＝9，即s＝3。
数字高程信息去噪预处理得到的结果如图4所示，黑色框所在区域的噪声被消除了。
(3)山脊线检测步骤，包括下述子步骤：
山脊线初检测算法的整体流程图如图5所示，具体分为水平方向山脊 线检测(3.1)和垂直方向的山脊线检测(3.2)两个子步骤。山脊线初检测以后还要进行连续山脊线的去除(3.3)。
(3.1)水平方向山脊线的检测步骤：
沿着水平方向一行行遍历地形数字高程矩阵，将每一个采样点的高程值H(x,y)与y方向一定范围内(如5个采样点即50米范围内)的采样点的高程值比较，如果该采样点在y方向是极大值，即
H(x,y)>H(x,y-5),
H(x,y)>H(x,y-4),
H(x,y)>H(x,y-3),
H(x,y)>H(x,y-2),
H(x,y)>H(x,y-1)，
H(x,y)>H(x,y+1)，
H(x,y)>H(x,y+2)，
H(x,y)>H(x,y+3)，
H(x,y)>H(x,y+4)，
H(x,y)>H(x,y+5)，
八个不等式同时成立，就认为该点为水平方向的山脊线的候选点，并在山脊候选点标记矩阵lable中设该点lable(x,y)＝1，否则设该点lable(x,y)＝0为非山脊线候选点。
(3.2)垂直方向山脊线的检测步骤：
沿着垂直方向一列列遍历，将每一个采样点的高程值与x方向一定范围内(如5个采样点即50米范围内)的采样点的高程值比较，如果该采样点在x方向是极大值，即
H(x,y)>H(x-1,y),
H(x,y)>H(x-2,y)，
H(x,y)>H(x-3,y)，
H(x,y)>H(x-4,y)，
H(x,y)>H(x-5,y),
H(x,y)>H(x+1,y)，
H(x,y)>H(x+2,y)，
H(x,y)>H(x+3,y)，
H(x,y)>H(x+4,y)，
H(x,y)>H(x+5,y)，
八个不等式同时成立，就认为该点为垂直方向的山脊线的候选点，并在山脊候选点标记矩阵lable中设该点lable(x,y)＝1，否则设该点lable(x,y)＝0为非山脊线候选点。
经过(3.1)(3.2)标记在地形数字高程信息图上的山脊线初检测结果如图6所示。
(3.3)提取连续的山脊线步骤：
由于按照(3.1)、(3.2)步骤提取出的山脊线的候选点有出现虚警的情况，所以提出了连续山脊线提取算法，其算法流程图如图7所示。逐一对山脊线初检测结果中的山脊线候选点判断是否是不连续的孤立点或者一定区域内(如：t*t的邻域内)山脊线候选点数目小于一定阈值th_num，如果是，就判断该山脊线候选点为非山脊线点，防止产生虚警。
本实例中设t＝7，th_num＝10，即：如果该候选山脊线点的7*7邻域内山脊线候选点的总数大于10，则认为该山脊线候选点为山脊线点；否则，认为该山脊线候选点为非山脊线点。标记在地形数字高程信息图上的连续山脊线提取结果如图8所示。
(4)山脊位置能量校正步骤，包括下述子步骤：
(4.1)山脊位置能量分布特征分析步骤：
山脊位置普遍存在山体效应：山脊位置的温度相对山腰的温度较高。这里引用针对无带状地下目标的山体温度场仿真结果加以证明，如图9所 示。可以看出山体中心的山脊位置的温度高于两边山腰处的温度。
该现象也可以从传热学角度分析：资料显示，山体中的岩石的热导率为1.2～2.1W/(m·℃)，山体表面接触的外界空气的热导率为0.024W/(m·℃)；热量总是沿着最容易传导的方向进行传播，所以山体内部的热量遇到空气时，由于其导热率远远小于岩石之间的热导率，热量大部分沿着岩石传导并在山脊处汇集，形成了山脊处温度较高的现象。
(4.2)山脊位置能量校正步骤：
采取最邻近插值法对山脊线处的能量进行校正，其算法流程图如图10所示，主要分为以下2步：
1、遍历找出山脊线点；
2、对每一个山脊线点lable(x,y)找出其四邻域内的非山脊线点，即：lable(x-1,y)、lable(x+1,y)、lable(x,y-1)、lable(x,y+1)中值为0的点；
3、求该山脊线点lable(x,y)四邻域内的非山脊线点对应的能量的均值；
4、将该山脊线点的能量用四邻域内非山脊线点对应的能量均值代替，并最终得到山脊能量校正后的结果。未进行能量校正的仿真红外图像(能量图像)如图11所示。将检测到山脊线位置结果标记在该仿真红外图像上如图12所示。山脊线位置能量校正后的红外仿真图像如图13所示。比较图11与图13，多处的偏亮区域有明显的降低，达到了山脊位置能量校正的目的。在步骤(5)中带状地下目标探测识别过程中出现在这些位置的虚警会有所降低。
(5)带状地下目标探测步骤，包括下述子步骤：
(5.1)设置遍历探测带状地下目标所用参数步骤：
采样段长度l＝25与宽度w＝3；
两旁山地地表采样段中心距离隧道上方采样段中心的距离s＝3个像素；
搜索的方向及遍历搜索的中心首尾坐标P0(10,27),P1(283,171)，搜索方向为大致为120°；
每次延展的距离d＝13；
左右平移探测的距离r＝3；
脉冲阈值th＝3。
(5.2)遍历探测结果的输出步骤：
利用(5.1)中设置的探测带状地下目标所用参数，对上面矫正前后的红外仿真图进行探测得到的虚警标记如图14和图15所示，虚警率从24.7％降低到21％。图12中方框标记出来的山脊位置，在山脊能量校正后的图14中对比未进行山脊能量校正的图3虚警有所减少。

Claims (10)

  1. 一种基于山脊能量校正的山地中带状地下目标的探测方法，其特征在于，所述方法包括如下步骤：
    (1)获取地形数字高程信息，包括下述子步骤：
    (1.1)确定数字高程信息覆盖范围的经纬度；
    (1.2)根据上述经纬度计算数字高程信息覆盖范围内经纬度矩阵；
    (1.3)根据上述经纬度矩阵计算经纬度矩阵中每个点的海拔，得到数字高程信息矩阵；
    (2)对步骤(1)获得的数字高程信息矩阵进行去噪预处理；
    (3)对降噪处理后的数字高程信息矩阵进行山脊线检测步骤；包括下述子步骤：
    (3.1)沿水平方向检测山脊线；
    (3.2)沿垂直方向检测山脊线；
    (3.3)根据上述水平方向检测的山脊线和垂直方向检测的山脊线，提取连续的山脊线；
    (4)山脊位置能量校正步骤；根据山脊位置能量分布特征对山脊位置能量进行校正；
    (5)在数字高程信息矩阵中进行带状地下目标探测的步骤；包括下述子步骤：
    (5.1)设置遍历探测带状地下目标所用参数；
    (5.2)根据上述参数对数字高程信息矩阵进行遍历探测，并将探测的带状地下目标位置输出。
  2. 如权利要求1所述的方法，其特征在于，所述步骤(1.2)具体包括：
    (1.2.1)计算地形范围的宽度、高度步骤：
    根据步骤(1.1)中确定的矩形四个顶点的经纬度利用Google Earth提供的测距工具计算出矩形覆盖范围的宽度、高度标记为width、height单位为米，并计算出经纬度的范围；
    (1.2.2)计算经纬度矩阵的步骤：
    按照采样间隔step米,每次距离step米采样一次，分别计算出纵向、横向的采样点的数目，即：height/step，width/step，并计算出经纬度矩阵每相邻两个采样点的经度步长long_step和纬度步长lati_step，从而计算出经纬度矩阵中每一个采样点的经纬度。
  3. 如权利要求1或2所述的方法，其特征在于，所述步骤(1.3)具体包括：
    将(1.2.2)中经纬度矩阵作为输入利用Google Earth得到每一个采样点的海拔数据，并根据该海拔数据生成地形数字高程信息矩阵输出。
  4. 如权利要求1至3任一项所述的方法，其特征在于，所述步骤(2)具体为：采用均值滤波方法对上述数字高程信息进行去噪预处理。
  5. 如权利要求1至3任一项所述的方法，其特征在于，所述步骤(3.1)具体为：
    沿着x方向遍历，将每一个采样点的高程值与y方向预设范围内的采样点的高程值比较，如果该采样点在y方向是极大值，则该点为水平方向的山脊线的候选点。
  6. 如权利要求1至3任一项所述的方法，其特征在于，所述步骤(3.2)具体为：
    沿着y方向遍历，将每一个采样点的高程值与x方向预设范围内的采样点的高程值比较，如果该采样点在x方向是极大值，则该点为垂直方向的山脊线的候选点。
  7. 如权利要求1至3任一项所述的方法，其特征在于，所述步骤(3.3)具体为：
    对步骤(3.1)、(3.2)提取出的每一个山脊线候选点进行连续性判别，判别准则如下：如果该候选山脊线点的t*t邻域内不同的山脊线候选点，则该山脊线候选点为最终的山脊线点；否则，判断该孤立的山脊线候选点为非山脊线点，最终得到连续山脊线标记图，其中t为预设值。
  8. 如权利要求1至3任一项所述的方法，其特征在于，所述步骤(4)具体为：对山脊位置的能量进行校正，校正的具体方法为：山脊线位置采样点的能量用山脊线位置两侧的采样点的能量的均值代替。
  9. 如权利要求1至3任一项所述的方法，其特征在于，所述步骤(5.1)中设置的参数具体为：
    采样段大小以及对比段距中心段的间隔：设下方有带状地下目标的山地地表的采样段的长度与宽度各为z个像素；带状地下目标两旁山地地表采样段的长度与宽度也各为z个像素；两旁山地地表采样段中心距离带状地下目标上方采样段中心的距离为s个像素，其中z、s为预设值。
    搜索的方向：通过查阅资料估计出该地区带状地下目标的走向，按照估计出的走向度数设置遍历搜索的中心首尾坐标P0(x0,y0),P1(x1,y1)，这两点的连线通过该区域中点附近的位置；
    每次延展的距离：每搜索一组假设下方存在带状地下目标的山地地表图像与两边对比的山地地表图像就向两边平移距离d再继续新的一组的搜索，直到到达图像的边界，搜索自动停止；
    脉冲阈值：每一段中间假设下方存在带状地下目标的山地地表图像采样段的灰度均值与两边对比的山地地表图像采样段的灰度均值的差的绝对值，均大于该脉冲阈值th的时候，该处的脉冲才算作有效脉冲；若两个差值中任意一个小于脉冲阈值该处的脉冲就是因为信号微弱而被认为无效。
  10. 如权利要求1至3任一项所述的方法，其特征在于，所述步骤(5.2)具体为：
    由(5.1)中确定的遍历搜索的中心首尾坐标P0(x0,y0),P1(x1,y1)， 开始使采样段从点P0(x0,y0)逐像素滚动到点P1(x1,y1)，且每次滚动一个像素后，再左右各平移r个像素，r为预设值，寻找其中中间段与左右两边对比段差值最大的位置，并测试该位置的脉冲是否为有效脉冲，若为有效脉冲，则效脉冲数目增加一；接着计算P0,P1向两边平移每次延展的距离之后的坐标P0’,P1’，在P0’,P1’之间按照P0,P1之间同样的采样方法进行统计，并并计数有效脉冲数目；有效脉冲数目占总数目最大的一组两端坐标P0,P1确定的位置为带状地下目标位置，结果中其余脉冲出现的位置即为虚警位置。
PCT/CN2015/072680 2014-12-30 2015-02-10 一种基于山脊能量校正的山地中带状地下目标的探测方法 WO2016106958A1 (zh)

Priority Applications (1)

Application Number Priority Date Filing Date Title
US15/114,651 US9625611B2 (en) 2014-12-30 2015-02-10 Method for detecting zonal underground target in mountain land based on ridge heat radiation correction

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN2014108513528 2014-12-30
CN201410851352.8A CN104484577B (zh) 2014-12-30 2014-12-30 一种基于山脊能量校正的山地中带状地下目标的探测方法

Publications (1)

Publication Number Publication Date
WO2016106958A1 true WO2016106958A1 (zh) 2016-07-07

Family

ID=52759118

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2015/072680 WO2016106958A1 (zh) 2014-12-30 2015-02-10 一种基于山脊能量校正的山地中带状地下目标的探测方法

Country Status (3)

Country Link
US (1) US9625611B2 (zh)
CN (1) CN104484577B (zh)
WO (1) WO2016106958A1 (zh)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109033032A (zh) * 2018-07-06 2018-12-18 北京兴农丰华科技有限公司 基于农机轨迹和地块格网化计算农机有效作业面积的方法
CN113533695A (zh) * 2021-07-26 2021-10-22 山东省农业机械科学研究院 一种农田墒情数据估计方法及系统
CN114384585A (zh) * 2021-12-30 2022-04-22 西北核技术研究所 基于相对位置及最小埋深的山地地下爆炸绝对定位方法

Families Citing this family (14)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104501959B (zh) * 2014-12-30 2016-08-17 华中科技大学 一种红外图谱关联智能探测方法及装置
CN105426881B (zh) * 2015-12-24 2017-04-12 华中科技大学 山体背景热场模型约束的地下热源昼间遥感探测定位方法
CN105654477B (zh) * 2015-12-25 2017-11-28 华中科技大学 一种条带状地下目标的探测定位方法
CN107092803B (zh) * 2017-05-12 2020-07-07 长安大学 一种基于三维线激光技术的道路积水区识别方法
CN108182724B (zh) * 2017-12-19 2021-08-24 深圳先进技术研究院 一种高精度城市热岛检测方法、设备及存储设备
CN110532986A (zh) * 2019-09-04 2019-12-03 云南电网有限责任公司带电作业分公司 基于modis遥感影像的山火检测算法、系统及其存储介质
CN110927816B (zh) * 2019-12-11 2021-08-24 中国地质科学院岩溶地质研究所 一种岩溶地下河系统的探测方法
CN111784725B (zh) * 2020-06-29 2023-06-20 易思维（杭州）科技有限公司 光条中心提取方法
CN112082655B (zh) * 2020-08-12 2022-08-12 华北电力大学 一种基于横向剪切干涉信号测量体温的方法
CN113268085B (zh) * 2021-07-16 2021-11-09 成都纵横大鹏无人机科技有限公司 一种航线规划方法、装置及机载激光雷达的飞行设备
CN113592032B (zh) * 2021-08-18 2023-04-18 电子科技大学 一种基于物理模型约束的红外成像虚警源分类方法
CN114112069B (zh) * 2022-01-27 2022-04-26 华中科技大学 地质约束的城市深埋条带通道红外成像探测方法及系统
CN115453521B (zh) * 2022-09-05 2024-05-24 西安电子工程研究所 一种二维相扫雷达地形探测方法
CN115953297B (zh) * 2022-12-27 2023-12-22 二十一世纪空间技术应用股份有限公司 一种遥感图像超分辨率重建和增强方法及装置

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102567702A (zh) * 2010-12-08 2012-07-11 中国科学院地理科学与资源研究所 一种基于嫦娥dem数据的自动识别山谷和山脊线的方法
CN103148842A (zh) * 2013-02-04 2013-06-12 国家海洋局第二海洋研究所 一种基于遥感图像特征的浅海沙波区多波束测深地形重构方法
CN103177258A (zh) * 2013-03-29 2013-06-26 河南理工大学 一种根据矢量等高线数据自动提取地性线的方法
CN103745191A (zh) * 2013-11-15 2014-04-23 中国科学院遥感与数字地球研究所 一种基于地形分析的黄土地区塬梁峁自动识别方法

Family Cites Families (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5086396A (en) * 1989-02-02 1992-02-04 Honeywell Inc. Apparatus and method for an aircraft navigation system having improved mission management and survivability capabilities
US7164883B2 (en) * 2001-02-14 2007-01-16 Motorola. Inc. Method and system for modeling and managing terrain, buildings, and infrastructure
CN102214410B (zh) * 2002-11-05 2014-01-01 亚洲航测株式会社 倾斜红色化立体图像制作装置
JP6120687B2 (ja) * 2012-06-14 2017-04-26 アジア航測株式会社 ラスター画像立体化処理装置及びラスター画像立体化方法並びにラスター画像立体化プログラム
CN103884431B (zh) * 2013-12-31 2015-09-09 华中科技大学 平面地表环境中地下建筑的红外成像探测定位方法
CN103744124B (zh) * 2013-12-31 2015-07-22 华中科技大学 一种平面地形中地下管状设施红外成像探测定位方法

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102567702A (zh) * 2010-12-08 2012-07-11 中国科学院地理科学与资源研究所 一种基于嫦娥dem数据的自动识别山谷和山脊线的方法
CN103148842A (zh) * 2013-02-04 2013-06-12 国家海洋局第二海洋研究所 一种基于遥感图像特征的浅海沙波区多波束测深地形重构方法
CN103177258A (zh) * 2013-03-29 2013-06-26 河南理工大学 一种根据矢量等高线数据自动提取地性线的方法
CN103745191A (zh) * 2013-11-15 2014-04-23 中国科学院遥感与数字地球研究所 一种基于地形分析的黄土地区塬梁峁自动识别方法

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109033032A (zh) * 2018-07-06 2018-12-18 北京兴农丰华科技有限公司 基于农机轨迹和地块格网化计算农机有效作业面积的方法
CN109033032B (zh) * 2018-07-06 2023-10-10 北京兴农丰华科技有限公司 基于农机轨迹和地块格网化计算农机有效作业面积的方法
CN113533695A (zh) * 2021-07-26 2021-10-22 山东省农业机械科学研究院 一种农田墒情数据估计方法及系统
CN114384585A (zh) * 2021-12-30 2022-04-22 西北核技术研究所 基于相对位置及最小埋深的山地地下爆炸绝对定位方法

Also Published As

Publication number Publication date
CN104484577B (zh) 2017-06-16
US9625611B2 (en) 2017-04-18
CN104484577A (zh) 2015-04-01
US20160356920A1 (en) 2016-12-08

Similar Documents

Publication Publication Date Title
WO2016106958A1 (zh) 一种基于山脊能量校正的山地中带状地下目标的探测方法
CN101566692B (zh) 利用卫星遥感数据中的云影信息检测云高的方法
CN104637073B (zh) 一种基于太阳照射阴影补偿的带状地下结构探测方法
Chen et al. Augmenting a deep-learning algorithm with canal inspection knowledge for reliable water leak detection from multispectral satellite images
ES2604807B1 (es) Método y sistema para buscar fugas de agua a través de análisis de imágenes generadas mediante sistemas de detección remota
Xu An approach to analyzing the intensity of the daytime surface urban heat island effect at a local scale
Jiang et al. Archeological crop marks identified from Cosmo-SkyMed time series: the case of Han-Wei capital city, Luoyang, China
Statella et al. Image processing algorithm for the identification of Martian dust devil tracks in MOC and HiRISE images
Xiong et al. Outgoing longwave radiation anomalies analysis associated with different types of seismic activity
Yu et al. Coal fire identification and state assessment by integrating multitemporal thermal infrared and InSAR remote sensing data: A case study of Midong District, Urumqi, China
CN105654477A (zh) 一种条带状地下目标的探测定位方法
Su et al. Detect and identify earth rock embankment leakage based on UAV visible and infrared images
CN114387253A (zh) 外墙外保温层缺陷红外图像处理方法、装置及存储介质
RU2428722C2 (ru) Способ дистанционной диагностики магистральных трубопроводов
Wu et al. Towards automated 3D evaluation of water leakage on a tunnel face via improved GAN and self-attention DL model
Su et al. A framework for RQD calculation based on deep learning
Sellami et al. A modern method for building damage evaluation using deep learning approach-Case study: Flash flooding in Derna, Libya
Bescoby Detecting Roman land boundaries in aerial photographs using Radon transforms
CN113011368A (zh) 矿井开采地表导通采空区裂隙识别方法及电子设备
CN115880597B (zh) 一种基于遥感技术的岩溶区落水洞提取方法
Xiong et al. Automatic defect detection in operational high-speed railway tunnels guided by train-mounted ground penetrating radar data
Parmar et al. Land use land cover change detection and its impact on land surface temperature of malana watershed kullu, Himachal Pradesh, India
Kour et al. Influence of shadow on the thermal and optical snow indices and their interrelationship
Liu et al. Semi-supervised deep neural network-based cross-frequency ground-penetrating radar data inversion
Wang et al. Monitoring urban expansion of the Greater Toronto area from 1985 to 2013 using Landsat images

Legal Events

Date Code Title Description
WWE Wipo information: entry into national phase

Ref document number: 15114651

Country of ref document: US

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

Ref document number: 15874651

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: 15874651

Country of ref document: EP

Kind code of ref document: A1

122 Ep: pct application non-entry in european phase

Ref document number: 15874651

Country of ref document: EP

Kind code of ref document: A1