JP4156960B2 - Underground cavity detector - Google Patents

Description

なお、P_α(l,d)，P_β(l,d)は、それぞれαデータ及びβデータを画像として表示する場合の小領域の濃淡度としても用いることができる。
【００４０】
βデータの作成が完了すると、制御部２０はデータ処理部３６に対し解析処理の開始を指示する。制御部２０にデータ解析の開始を指示されたデータ処理部３６は、αデータ記憶部３３およびβデータ記憶部３５に蓄積されたデータに対し、解析処理を開始する（Ｓ０３０）。その動作について図２を交え説明する。初めにβデータに対して第一候補領域決定手段５１を適用し、第一候補領域を決定する。第一候補領域決定手段５１では、まず、βデータ記憶部３５中の小領域について、小領域の位置(l,d)を中心とした幅Ｒ_βの範囲内に存在する同深度の周辺小領域の平均パワー
【数２】

を算出する。この
【数３】

と、小領域のパワーP_β(l,d)との比が次式を満たす場合、当該小領域を第一候補領域と判定する：
【数４】

なお、パワー比以外にパワーの差分を用いて判定することも可能である。そして、当該第一候補領域の位置情報（l,d）と、小領域のサイズから、当該第二候補領域の左端(START)、右端(END)、上端(TOP)、下端(BOTTOM)位置を算出し、候補領域情報記憶部３７に蓄積する。この手順は、開始地点から終了地点までの範囲内に存在する小領域について実行される。
【００４１】
第一空洞候補領域の選出が完了すると、次に候補領域情報統合部６０による領域情報の統合が行われる（Ｓ０３２）。すなわち、候補領域情報統合部６０中の領域情報統合手段６１において、候補領域情報記憶部３７に記憶された任意の二領域のSTART，ENDのうち最も近接した２つの距離が基準値以下で、かつ、TOP，BOTTOMのうち最も近接した２つの距離が所定の値以下である場合、この二領域を統合する。統合は様々に定義することが可能であるが、例えば統合の対象となる２領域を、
【数５】
(START_１,END_１,TOP_１,BOTTOM_１)
(START_２,END_２,TOP_２,BOTTOM_２)
で表したとき、統合後の領域(START_ＮＥＷ,END_ＮＥＷ,TOP_ＮＥＷ,BOTTOM_ＮＥＷ)を、次式で定義することが可能である：
【数６】
START_ＮＥＷ＝min(START_１,START_２)
END_ＮＥＷ＝max(END_１,END_２)
TOP_ＮＥＷ＝min(TOP_{１ ,}TOP_２)
BOTTOM_ＮＥＷ＝max(BOTTOM_１,BOTTOM_２)
そして、新たな領域情報として前記候補領域情報記憶部３７に記憶するとともに、この二領域の情報を候補領域情報記憶部３７から消去する。領域情報の統合は、統合の対象となる領域が存在しなくなるまで行う。
【００４２】
領域統合によって新たに生成された空洞候補領域に対しては、一意に対応するIDを付与し、当該IDの空洞候補領域を形成する小領域の位置情報（例えば探査位置や上端位置を表すTOＰ等）を保存してもよい。これにより、空洞の形状の詳細な把握や、後段の空洞の候補絞込み過程での利用が容易となる。
【００４３】
領域情報の統合が完了すると、次に第二空洞候補選出部７０による第二空洞候補の選出が行われる（Ｓ０３５）。第二空洞候補選出部７０中の信号強度判定手段７１では、候補領域情報記憶部３７に記憶された領域情報(START,END,TOP,BOTTOM)に従って、設定した範囲において候補領域内の深さdに存在する小領域のパワーのうち最大の値
【数７】

を算出する。また、深さdにおいて候補領域の両端(START,END)から、それぞれ幅Ｒ_αの範囲内に存在する同深度の周辺小領域のパワーのうち最小の値となる
【数８】

を算出する。
【数９】

の算出においては、例えば、候補領域内の深さdに存在する小領域のパワーの平均値を用いても良い。また、
【数１０】

の算出においては、ある程度少ない数の領域の中から最小パワーを算出しても良い。また、候補領域の両端（START,END）から、それぞれ幅Ｒ_αの範囲内に存在する同深度の周辺小領域の平均パワーや、幅Ｒ_αの範囲内に存在する同深度の周辺小領域のパワーのうち下位Ｎ位までのパワーの平均値を
【数１１】

としても良い。設定した範囲における
【数１２】

と
【数１３】

との比の最大値が次式を満たす場合、この小領域を空洞候補領域と判定する：
【数１４】

なお、パワー比以外にパワーの差分を用いて判定することも可能である。
【００４４】
信号強度判定手段７１で空洞候補と判定された領域情報は、レーダデータ置換手段７２に渡される。レーダデータ置換手段７２では、第二空洞候補選出部７０で選出された第二空洞候補領域の位置情報(START,END,TOP,BOTTOM)に基づき、データバッファ記憶部３２中のバッファデータを候補領域近傍の平均値で置換する（Ｓ０３６）。
【００４５】
以上の手順によって選出した候補領域の位置情報は、上記の通り候補領域情報記憶部３７に蓄積される。候補領域の選出が完了すると、次に空洞候補絞込み部８０による空洞候補の絞込みが行われる（Ｓ０４０）。空洞候補の絞込みは、候補領域情報記憶部３７に蓄積された全ての候補領域に対して、極性判定手段８１、マンホール検出手段８２、候補領域絞込み手段８３を任意の順で適用する。適用順序は制御部２０で制御する。例えば、極性判定手段８１、マンホール検出手段８２、候補領域絞込み手段８３の適用手順としたり、マンホール検出手段８２、極性判定手段８１、候補領域絞込み手段８３の適用手順としたりすることができる。空洞候補絞込み部８０による候補絞り込みが完了すると、候補領域情報記憶部３７に蓄積された空洞候補領域の位置情報を候補領域情報一時記憶部３８に転送するとともに（Ｓ０４５）、候補領域情報記憶部３７に蓄積された全ての候補領域情報を消去する（Ｓ０５０）。
【００４６】
極性判定手段８１では、候補領域情報記憶部３７に蓄積された空洞候補のαデータについて、極性の判定を行う。極性の判定方法については、パターン認識的な手法を用いることができる。例えば、位相の反転部分を黒、非反転部分を白で表したとき、空洞領域からの反射パターン（正極性とする）は白から黒へ変化するパターンとなり、非空洞領域からの反射パターン（負極性とする）は黒から白へ変化するパターンで表現される。また、前記候補領域を形成する小領域位置と小領域部分を含むレーダデータの周波数スペクトルを解析する方法などが利用できる。なお、空洞領域における反射は、正確には、その境界において起こると言えるので、電磁波の波長が空洞領域の大きさよりも十分短い場合には、この位置のずれが問題となる可能性がある。この場合には、極性判定手段８１の処理を、空洞領域よりもやや広い範囲について行うようにすればよい。
【００４７】
マンホール検出手段８２は、空洞と同様の極性を示すマンホールを候補領域から除去する為の手段である。候補領域情報記憶部３７に記憶された領域情報とαデータから、例えば路面１の深度（位置）と、当該候補領域のTOP位置を比較し、当該領域のTOP位置が該地表面位置より上にあり、かつ、当該領域のTOPからBOTTOMまでの長さが基準値以上である領域を候補領域情報記憶部３７から消去する。例えば、路面１の深度d_surface(l)は、距離lにおけるαバッファの信号u_α(l,y)について、所定の値を超える極大値を与えるyのうち最小のyをd’(l)としたとき、次式で求めることができる：
【数１５】
d_surface(l)＝d’(l)−η
ここで、ηは電磁波パルスの波長に基づき決定される定数である。なお、消去した領域の情報を、別途記憶することで、マンホール候補領域のみのリストを作成することも可能である。
【００４８】
候補領域絞込み手段８３では、前記候補領域情報記憶部３７に蓄積された領域のうち、大きさが一定の基準値以下であるものを空洞候補から除外して、空洞候補領域を絞込む。例えば、小領域の横幅を８０ｃｍとして処理する場合に、領域情報統合手段６１で領域情報を統合した後で、候補領域の幅が小領域の幅以下の候補を除外すると、有効に空洞候補領域の絞込みを実行できることを確認している。
【００４９】
以上の手順（Ｓ００５−Ｓ０５０）を完了した時点で、これまで完了した解析回数カウンタＭを１加算する（Ｓ０５５）。このＭと所定の解析回数Nを比較し、Ｍ＞Ｎならば解析終了処理を行い、そうでなければ再解析作業（第２回目以降の解析作業）に移る（Ｓ０６０）。第２回目以降の再解析作業では、既にデータバッファ３２中に蓄積されているバッファデータを利用してＳ０２０−Ｓ０５０の解析処理を行う。
【００５０】
解析作業回数Ｎの設定としては、１以上の整数を指定する。ただし、１を設定した場合、上記の通りＳ００５−Ｓ０５０の手順を１回実行しただけでは、地中の水溜り等、空洞領域と同等以上に反射波の強度が大きい箇所が空洞領域の近傍に存在する場合に、この空洞領域が信号強度判定により非空洞領域と判定されて見落とされてしまう恐れがある。そこでここでは、Ｎを２以上に設定することで、レーダデータ置換手段７２によって検出済みの候補箇所を除いたデータを用いて、再度、Ｓ０２０−Ｓ０５０の手順で空洞候補の選出を行い、空洞領域の見逃しを抑止する。検出済みの候補箇所を除いたデータについては、レーダデータ置換手段７２によって生成されてデータバッファ３２に蓄積されている。そこで、このデータバッファ３２中のバッファデータから、新たにαデータ・βデータを生成し、データ解析装置３０による解析処理を行えばよい。なお、解析作業回数Ｎを大きな値に設定しておき、新たな空洞候補が選出されなくなるまで空洞候補の再選出を行うこともできるが、多くの場合Ｎ＝３程度とすれば有効な結果が得られることを確認している。
【００５１】
解析終了判定Ｓ０６０においてＭ＞Ｎである場合には、解析終了処理に移る。解析終了処理では、まず、候補領域情報一時記憶部３８に蓄積された候補領域情報を候補領域情報記憶部３７に転送する（Ｓ０６５）。そして、候補領域情報記憶部３７に転送・蓄積された候補領域情報に対し、領域情報統合手段６１により候補領域情報の統合を行う（Ｓ０７０）。候補領域情報の統合（Ｓ０７０）が完了した時点で、候補領域情報記憶部３７に蓄積された候補領域情報を、データ解析装置３０の処理結果として、表示部３９または出力部４０に出力する（Ｓ０７５）。
【００５２】
以上、図３に従い本実施の形態の動作フローを説明したが、第一空洞候補選出部５０に続く候補領域情報統合部６０、第二空洞候補選出部７０、空洞候補絞込み部８０の適用順序はこの限りではない。例えば、第一空洞候補選出部５０、空洞候補絞込み部８０の極性判定手段８１、候補領域情報統合部６０、第二空洞候補選出部７０、空洞候補絞込み部８０の適用順としてもよい。なお、このように適用順序を変更した場合には、各部において候補領域情報記憶部３７に記憶された空洞領域に対する処理も、直前の処理で更新された空洞領域に対して行うように変更されることは言うまでもない。
【００５３】
また、以上の説明においては、第一空洞候補選出部５０は、小領域に対して、同深度の周辺領域の（あるいは平均的）反射強度に基づいて第一空洞候補を決定した。しかし、こうした周辺領域の反射強度が、深度方向にある程度一様である場合には、鉛直方向の周辺領域を考慮することができる為、同深度という限定は必ずしも必要ではない。
【００５４】
【発明の効果】
本発明の地中空洞検出装置により、空洞領域の自動検出の精度を向上させることが可能となる。
【図面の簡単な説明】
【図１】 本実施の形態の装置構成を示すブロック図である。
【図２】 図１のデータ処理部３６の構成を示すブロック図である。
【図３】 処理の流れの一例を示すフローチャートである。
【符号の説明】
１０ 電磁波レーダ、２０ 制御部、３０ データ解析装置、３１ 前処理フィルタ、３２ データバッファ、３３ αデータ記憶部、３４ 高域通過型フィルタ、３５ βデータ記憶部、３６ データ処理部、３７ 候補領域情報記憶部、５０ 第一空洞候補選出部、６０ 候補領域情報統合部、７０ 第二空洞候補選出部、８０ 空洞候補絞込み部。[0001]
BACKGROUND OF THE INVENTION
The present invention relates to an underground cavity detection apparatus, and more particularly to an apparatus for detecting an underground cavity using received data of reflected waves of electromagnetic waves radiated into the ground.
[0002]
[Prior art]
In order to search for a cavity existing under a road surface, an electromagnetic wave radar device mounted on a search vehicle is used. The electromagnetic wave radar device irradiates an electromagnetic wave pulse under a road surface and observes a reflected wave to generate a two-dimensional digital image. In the past, an inspector (expert) visually observed the acquired image and specified a place that seems to be a cavity. Then, in the specified place, a hole is actually made, and if there is a cavity, the cavity is filled.
[0003]
Patent Document 1 describes a road cavity exploration radar system mounted on a vehicle. Patent Document 2 describes a cavity exploration method ranging from investigation by a radar to boring drilling. Patent Document 3 discloses a nondestructive inspection method for obtaining the shape or position of a target in the ground or concrete using a radar image processing system.
[0004]
[Patent Document 1]
Japanese Patent Laid-Open No. 4-194893
[Patent Document 2]
JP-A-5-87945
[Patent Document 3]
JP 9-292350 A
[0005]
[Problems to be solved by the invention]
However, the conventional method of visual inspection by an inspector cannot always clearly identify the cavity region. In addition, in the image processing disclosed in Patent Document 3 and the like, the cavity region cannot be accurately extracted.
[0006]
An object of the present invention is to improve the automatic detection accuracy of a cavity region.
[0007]
[Means for Solving the Problems]
  In order to solve the above problems, the underground cavity detection device of the present inventionElectromagnetic waveRadiateTheReflected waveReflection intensity in the depth direction of the ground was measured by radar receivingreceived dataFromUnderground cavity detector for detecting underground cavitiesBecause,A reception data storage unit that stores the reception data sequentially measured at a plurality of points along a predetermined traveling direction as radar data; and a data analysis unit that processes the radar data to detect a cavity, and The data analysis unit includes a pre-processing filter that generates a data by performing a predetermined filter process for noise removal on the radar data, and a high-pass filter that removes a signal that constantly exists in the traveling direction. A process is performed on the α data, a high-pass filter that generates β data, and a plurality of small regions are set in the depth direction and the traveling direction, and the β data in the small regionReflection intensity around the same depthOf the β data inReflection intensity or at approximately the same depthOf the β dataAverageNaCompared with reflection intensityThe small region is a hollow regionWhether or not1st cavity candidate selectionMeans,Compare the reflection intensity of the α data in the small area with the reflection intensity of the α data in the peripheral area at approximately the same depth or the average reflection intensity of the α data at approximately the same depth. A second cavity candidate selection means for determining whether or notIncludingA cavity is detected in a small area determined as a cavity area by the first cavity candidate selecting means and the second cavity candidate selecting means.
[0008]
According to this configuration, the reflection intensity from the area is compared with the reflection intensity from the peripheral area at substantially the same depth or the average reflection intensity at substantially the same depth to determine the presence of the cavity area. At substantially the same depth, unless there is a hollow region, generally, conditions such as electromagnetic wave spreading and scattering are equivalent, and soil conditions such as the formation are often equivalent. For this reason, it is possible to detect a cavity region having a reflection intensity different from that of a normal region. Note that the same depth here naturally allows a slight error, and it goes without saying that, for example, an error of a thickness (in the depth direction) of the region is sufficiently allowed.
[0009]
  Preferably, in the underground cavity detection device of the present invention,Data analysis departmentIsArea information for integrating a plurality of small areas existing within a predetermined distance among the small areas determined to be hollow areasIntegration meansMoreIncluding. This makes it possible to identify a cavity region existing within a predetermined distance as one large cavity region. The definition of the predetermined distance can be variously determined, and only the objects that touch or overlap each other may be targeted, or a range that is several times larger than the area may be selected. Further, the definition may be changed every time it is implemented.
[0010]
  Preferably, in the underground cavity detection device of the present invention,Data analysis departmentIsThe polarity of the α data is determined in the vicinity of the small area determined to be a hollow area, and the small area determined to have the polarity reversed is excluded from the detection targets of the cavity.Polarity judgment meansMoreIncluding. GeneralInvite toFrom a medium with a high electric power to a medium with a low dielectric constantElectromagnetic wavesWhen incident, a reflected wave that does not invert the phase (polarity) is generated at the boundary surface. Since the cavity region has a smaller dielectric constant than a solid part such as soil, a reflected wave whose polarity is not reversed can be detected. Therefore, it is possible to distinguish the buried object from the hollow region.
[0011]
  Preferably, in the underground cavity detection device of the present invention,Data analysis departmentIsRadar data replacing means for generating new radar data by replacing the reflection intensity of the radar data with the average intensity of the surrounding area in the small area where the cavity is detected, and generated by the radar data replacing means A new cavity is detected based on the radar data.With this configuration, it is possible to prevent detection of the hollow area from being detected. Of course, the reanalyzing means can be repeated a plurality of times. Note that the integration unit may be implemented after the reanalysis unit is implemented.
[0012]
  Preferably, in the underground cavity detection device of the present invention,Data analysis departmentIs the small region where the shallowest part is shallower than the predetermined depth and the deepest part is deeper than the predetermined depth.Cavity detection as a manholeExclude from the targetFurther manhole detection meansIncluding. This configuration is suitable for removing special cavities such as manholes. That is, the exclusion means can be implemented as a manhole detection means.
[0013]
  Preferably, in the underground cavity detection device of the present invention,Data analysis departmentSaidArea informationIntegrated by means of integrationThe small area that was later determined to be a hollow areaLess than a predetermined size for the areasmallAreaFrom the cavity detection targetexcludeCandidate areaNarrow down meansMoreIncluding. As a result, the remaining hollow region is an integration of at least two small regions. Therefore, it can be said that an area with high detection reliability remains. If the size of the small area is set slightly smaller than the size of the cavity area of interest, there is no worry of removing the cavity area of interest, and more calculations are performed unnecessarily. There is nothing.
[0015]
With the configuration described above, the cavity region can be automatically detected without visual inspection by the inspector. This can be expected to reduce human and time costs. In addition, human factors (ambiguity) are removed when candidate positions are determined, so that cavities and buried objects can be distinguished accurately, and the number and size of cavities can be detected with high accuracy, so that detailed reexamination and drilling can be performed. It becomes possible to reduce the work.
[0016]
The reception data of the reflected wave used in the present invention may be three-dimensional data or two-dimensional data composed of depth and a specific horizontal direction. In the case of three-dimensional data, the cavity may be detected two-dimensionally using the means described above, but it goes without saying that the three-dimensional data can also be detected.
[0017]
DETAILED DESCRIPTION OF THE INVENTION
Preferred embodiments of the present invention will be described below with reference to the drawings.
[0018]
The block diagram of FIG. 1 shows the configuration of the underground cavity detection device of the present embodiment. In this example, a cavity is detected using an exploration vehicle. In other words, the exploration vehicle travels on the road surface 1 in the direction of travel, and aims to detect the cavity 2 in the ground below the road surface 1. Therefore, the underground cavity detection device 3 is mounted on the vehicle and is installed so that the cavity can be detected while the exploration vehicle is stopped and traveling. The underground cavity detecting device 3 is inputted with speed information from the speedometer 4 of the exploration vehicle. In addition, for example, position information can be input from a position measuring device such as a GPS.
[0019]
The underground cavity detection device 3 includes an electromagnetic wave radar 10, a control unit 20, and a data analysis device 30 as main devices. The electromagnetic wave radar 10 functions as a measurement unit for investigating the cavity. The control unit 20 and the data analysis device 30 can be realized by operating a general computer and peripheral devices in cooperation with an application program that instructs each unit of the present embodiment.
[0020]
The electromagnetic wave radar 10 is provided with a radar control unit 11. The radar control unit 11 controls the transmission unit 12 and the reception unit 13 of the radar. That is, the radar control unit 11 controls the radar output interval of the transmission unit 12 and the reception interval of the associated reception unit 13 according to the speed information sent from the speedometer 4 of the exploration vehicle, and a constant distance interval (for example, 5 cm). We are measuring with.
[0021]
The electromagnetic wave pulse 14 radiated from the transmission unit 12 is mainly reflected at that point to be a reflected wave 15 when the dielectric constant changes, and is received by the reception unit 13. For this reason, this reflected wave 15 is mainly comprised as a synthetic | combination wave from boundary surfaces, such as the cavity 2 which exists under the road surface 1, and a buried object. The time difference t until the electromagnetic wave pulse radiated from the transmission unit 12 is observed by the reception unit 13 is determined by the propagation speed of the electromagnetic wave in the medium below the road surface 1 and the depth of the boundary surface where the reflection occurs. .
[0022]
The received reflected wave 15 is accumulated as radar data in the reception data storage unit 16 via the radar control unit 11. The reception data storage unit 16 may be a recording medium such as a CD-R and a DVD-R in addition to a volatile / nonvolatile memory. In the example according to the present embodiment, the electromagnetic wave radar 10 does not have sufficient resolution in the vehicle width direction (horizontal direction orthogonal to the traveling direction). Therefore, the information in the vehicle width direction is compressed, and the radar data is stored as a two-dimensional gray image having the traveling direction (distance) and the depth direction (depth) of the exploration vehicle as two axes. If the electromagnetic wave radar 10 has sufficient resolution in the vehicle width direction (horizontal direction orthogonal to the traveling direction), the radar data may be stored as three-dimensional data. Further, the cavity analysis method described below can be easily performed on this three-dimensional data.
[0023]
The control unit 20 is realized by using a function of a calculation unit centering on a CPU of a computer. Then, the observation operation in the electromagnetic wave radar 10 and the data transfer timing to the data analysis device 30 are instructed. In addition, the data analysis device 30 is controlled for various analyzes described below.
[0024]
The radar data accumulated in the received data storage unit 16 is sent to the preprocessing filter 31 of the data analysis device 30, noise clipping is removed using an intermediate value filter and a low-pass filter, and the data is transferred to the data buffer 32. Is done. The data buffer 32 accumulates the transferred radar data, and transfers the radar data to the α data storage unit 33 in accordance with an instruction from the control unit 20. The α data storage unit 33 accumulates radar data transferred from the data buffer 32. At this time, every time the data transfer amount from the data buffer 32 to the α data storage unit 33 reaches a certain value, the control unit 20 is notified of this fact. In accordance with the instruction from the control unit 20, the high-pass filter 34 performs a filtering process on the data stored in the α data storage unit 33 to remove stationary signals parallel to the traveling direction such as roads and strata. The type of filter is not particularly limited. For example, a moving average may be taken with an appropriate width in the traveling direction, and the value may be subtracted from the original data. The filtered radar data is transferred to the β data storage unit 35. The β data storage unit 35 accumulates data transferred from the high-pass filter 34.
[0025]
The data processing unit 36 has an arithmetic processing function, performs detection processing of a cavity candidate region using radar data accumulated in the α data storage unit 33 and the β data storage unit 35, and stores it in the candidate region information storage unit 37. To do. The data in the data buffer 32 is also rewritten based on the detected cavity candidate area. The data stored in the candidate area information storage unit 37 can be temporarily stored in the candidate area information temporary storage unit 38. In addition to the series of parts related to the detection processing of the cavity candidate region, that is, the data processing unit 36, the preprocessing filter 31, the data buffer 32, the α data storage unit 33, the high-pass filter 34, the β data storage unit 35, It can be said that the candidate area information storage unit 37 and the candidate area information temporary storage unit 38 function as a cavity detection unit.
[0026]
  In addition, the data analysis device 30 is provided with a display unit 39 configured by a liquid crystal display, a CRT display, or the like, and an output unit 40 that can output data to an external device. Data can be output from the candidate area information storage unit 37 to the display unit 39 and the output unit 40. Display section39Then, it is also possible to transmit and display radar data before data analysis from the received data storage unit 16.
[0027]
Next, the configuration of the data processing unit 36 will be described with reference to the block diagram of FIG. The data processing unit 36 includes a first cavity candidate selection unit 50, a candidate area information integration unit 60, a second cavity candidate selection unit 70, and a cavity candidate narrowing unit 80. The first cavity candidate selection unit 50 functions as a small area determination unit that determines whether the small area is a hollow area, and uses the area information of the small area determined as the hollow area as the first cavity candidate candidate area information. Accumulate in the storage unit 37. The candidate area information integration unit 60 is an integration unit that integrates the hollow areas stored in the candidate area information storage unit 37. As a result, a part of the cavity regions of the first cavity candidate is integrated and redefined as a new cavity region. In the second cavity candidate selection unit 70, the cavity region of the first cavity candidate stored in the candidate region information storage unit 37 is analyzed based on the reflection intensity of the radar data accumulated in the α data storage unit. It is determined whether it is appropriate. If it is determined as a hollow region, it is accumulated in the candidate region information storage unit 37 as a second hollow candidate. That is, the second cavity candidate selection unit 70 can be said to be a cavity region strength determination unit. The cavity candidate narrowing unit 80 narrows the second cavity candidate cavity regions stored in the candidate region information storage unit 37 based on the radar data accumulated in the α data storage unit, and reduces the number thereof.
[0028]
The first candidate area determination means 51 included in the first cavity candidate selection unit 50 sets a rectangular area of an arbitrary size as a small area in the radar data accumulated in the β data storage device 35. The size of the rectangle may be set based on the minimum size of the cavity to be detected. For example, the width is about 80 cm and the depth is about 30 cm. Then, the reflection intensity with the region of the same depth is compared, and if the difference is greater than the set value, this small region is determined as a hollow region. The region of the same depth may be a peripheral region such as adjacent to this rectangular region, and an average reflection intensity at this depth can be used. The determined area information of the cavity area is accumulated in the candidate area information storage unit 37 as the first cavity candidate.
[0029]
In the region information integration unit 61 in the candidate region information integration unit 60, the distance between the regions is equal to or less than a certain reference (for example, 3 m) with respect to the cavity region that is the first cavity candidate accumulated in the candidate region information storage unit 37. In some cases, these are integrated to redefine the cavity region, and the first cavity region candidate is updated.
[0030]
In the signal strength determination means 71 in the second cavity candidate selection unit 70, the radar data accumulated in the α data storage unit 33 is the same as the cavity region of the first cavity candidate accumulated in the candidate region information storage unit 37. A rectangular area corresponding to the position and the same size is set, and it is determined whether or not the rectangular area can be a hollow area based on a power ratio with a peripheral area of the same depth. The determined cavity region is accumulated in the candidate region information storage unit 37 as a second cavity candidate. Although it is possible to not include a region that has not been determined as a candidate, the integration of the cavity region by the region information integration unit 61 is canceled, and signal strength determination is performed again for each of the canceled cavity regions. May be.
[0031]
In the radar data replacement means 72, the reflection intensity of the radar data in the data buffer 32 corresponding to the cavity area selected as the second cavity candidate by the second cavity candidate selection unit 70 is used as a representative reflection intensity around this cavity area. Replace with. As a representative reflection intensity, for example, an average value of the reflection intensity in the peripheral region may be used. This replacement is performed to reanalyze the detection of a series of cavity regions using the received data of the replaced reflected wave.
[0032]
The cavity candidate narrowing unit 80 narrows down the second cavity candidates stored in the candidate area information storage unit 37 based on the radar data accumulated in the α data storage unit, thereby reducing the number of cavity candidates. That is, the polarity determination unit 81 of the cavity candidate narrowing unit 80 determines whether the area is a cavity area or other embedded object by determining the polarity of the radar signal waveform near the cavity area. The manhole detecting means 82 has a depth at the upper end position (the shallowest part) of the area accumulated in the candidate area information storage unit 37 that is shallower than a predetermined reference and the deepest part is equal to or greater than a predetermined reference value. In some cases, it is an exclusion means for excluding the region as a non-hollow region (manhole). Further, the candidate area narrowing means 83 removes the area from the cavity area and reduces the number of cavity candidates when the size of the area accumulated in the candidate area information storage unit 37 is less than the size of the object of interest. It is a narrowing part.
[0033]
Next, the flow of processing will be described using the flowchart of FIG. There are two types of processing that can be executed in the present embodiment: real-time processing and batch processing. In the real-time processing, whenever data under the road surface 1 is acquired, radar data is transferred from the electromagnetic wave radar 10 to the data analysis device 30 for analysis processing. Further, the batch processing is to perform analysis processing in a lump after accumulating all the data of the analysis target route in the received data storage unit 16. However, in any case, the operation of the data analysis device 30 is the same.
[0034]
In these processes, one or more analysis operations are performed with respect to one road survey data. In the following description, the number of analysis operations is N. The analysis work number N may be set in advance before the analysis work, or may be determined adaptively during the analysis work.
[0035]
In the first analysis work, analysis is performed using radar data acquired by the electromagnetic wave radar 10, but analysis is performed using data generated in the data analysis device 30 from the second time. For this reason, the operation flow of processing differs between the first analysis work and the second and subsequent times (S001). Here, first, processing in the first analysis work will be described.
[0036]
Radar data acquired by the electromagnetic wave radar 10 (S005) and accumulated in the reception data storage unit 16 is filtered by the preprocessing filter 31 using an intermediate value filter or a low-pass filter (S010), and stored in the data buffer 32. Transferred (S015).
[0037]
The data buffer 32 accumulates the radar data transferred from the preprocessing filter 31 and transfers the radar data to the α data storage unit 33 in accordance with an instruction from the control unit 20 (S020). The data transferred from the data buffer 32 is accumulated in the α data storage unit 33. Hereinafter, the data in the data buffer is referred to as “buffer data”, the data in the α data storage unit 33 is referred to as “α data”, and the buffer data and α data at the distance l and the depth d are u (l ,, d), respectively. , U_αRepresented by (l, d). The α data storage unit 33 includes a counter that counts the amount of data transferred from the preprocessing filter 31. When the counter reaches a predetermined value, the control unit 20 is notified of this and the counter is reset. . When the control unit 20 receives the notification from the α data storage unit 33, the control unit 20 stops the data transfer from the data buffer 32 to the α data storage unit 33, and the high-pass filter 34 stores the data in the α data storage unit 33. Instructs filtering of alpha data.
[0038]
Data processed by the high-pass filter 34 is accumulated in the β data storage unit 35 (S025). Hereinafter, the data in the β data storage unit 35 is referred to as “β data”, and the β data at the distance l and the depth d is u._βRepresented by (l, d).
[0039]
In the α data and β data, horizontal X_r(Traveling direction) x vertical Y_rA small region (depth direction) is set. Where X_r, Y_rIs a constant representing the size of the small area, and is set according to the distance interval recorded in the received data storage unit during the search. Also, the power P as the reflection intensity of the small area in the α data at the position of l in the traveling direction and d in the depth direction from the exploration start point._α(l, d), and power P of the small area in β data_βCalculate (l, d). There are several ways to define this power, for example it can be defined by the following formula:
[Expression 1]

P_α(l, d), P_β(l, d) can also be used as the shade of a small area when α data and β data are displayed as images, respectively.
[0040]
When the creation of β data is completed, the control unit 20 instructs the data processing unit 36 to start analysis processing. The data processing unit 36 instructed by the control unit 20 to start data analysis starts analysis processing on the data stored in the α data storage unit 33 and the β data storage unit 35 (S030). The operation will be described with reference to FIG. First, the first candidate area determining means 51 is applied to the β data to determine the first candidate area. In the first candidate area determination means 51, first, for the small area in the β data storage unit 35, the width R centered on the position (l, d) of the small area._βThe average power of a small area around the same depth
[Expression 2]

Is calculated. this
[Equation 3]

And small area power P_βIf the ratio with (l, d) satisfies the following equation, the small region is determined as the first candidate region:
[Expression 4]

It is also possible to make a determination using a power difference other than the power ratio. From the position information (l, d) of the first candidate area and the size of the small area, the left end (START), right end (END), upper end (TOP), and lower end (BOTTOM) positions of the second candidate area are determined. Calculate and accumulate in the candidate area information storage unit 37. This procedure is executed for a small area existing in the range from the start point to the end point.
[0041]
When the selection of the first cavity candidate region is completed, region information integration is performed by the candidate region information integration unit 60 (S032). That is, in the region information integration unit 61 in the candidate region information integration unit 60, the two closest distances among START and END of any two regions stored in the candidate region information storage unit 37 are equal to or less than the reference value, and When the two closest distances among TOP, BOTTOM are less than a predetermined value, the two areas are integrated. Integration can be defined in various ways. For example, two areas to be integrated are
[Equation 5]
(START₁, END₁, TOP₁, BOTTOM₁)
(START₂, END₂, TOP₂, BOTTOM₂)
Area after integration (START_NEW, END_NEW, TOP_NEW, BOTTOM_NEW) Can be defined by:
[Formula 6]
START_NEW= Min (START₁, START₂)
END_NEW= Max (END₁, END₂)
TOP_NEW= Min (TOP_{1 ,}TOP₂)
BOTTOM_NEW= Max (BOTTOM₁, BOTTOM₂)
Then, the new region information is stored in the candidate region information storage unit 37 and the information on the two regions is deleted from the candidate region information storage unit 37. The integration of the area information is performed until there is no area to be integrated.
[0042]
A unique candidate ID is assigned to a cavity candidate area newly generated by area integration, and position information of a small area that forms the cavity candidate area of the ID (for example, a TOP that indicates a search position or an upper end position) ) May be saved. As a result, it becomes easy to grasp the shape of the cavity in detail and use it in the process of narrowing down candidate cavities in the subsequent stage.
[0043]
When the integration of the area information is completed, the second cavity candidate selection unit 70 selects a second cavity candidate (S035). In the signal strength determination means 71 in the second cavity candidate selection unit 70, the depth d in the candidate region in the set range according to the region information (START, END, TOP, BOTTOM) stored in the candidate region information storage unit 37. The maximum value of the power of the small area existing in
[Expression 7]

Is calculated. In addition, the width R from each end (START, END) of the candidate area at the depth d_αIt becomes the minimum value of the power of the surrounding small area of the same depth that exists in the range of
[Equation 8]

Is calculated.
[Equation 9]

In this calculation, for example, an average value of the powers of the small areas existing at the depth d in the candidate area may be used. Also,
[Expression 10]

In this calculation, the minimum power may be calculated from a relatively small number of regions. In addition, the width R from each end (START, END) of the candidate area_αThe average power and width R of the surrounding small area of the same depth within the range_αThe average value of the power to the lower N ranks of the power of the surrounding small area of the same depth existing in the range of
[Expression 11]

It is also good. Within the set range
[Expression 12]

When
[Formula 13]

If the maximum value of the ratio satisfies the following equation, this small region is determined as a cavity candidate region:
[Expression 14]

It is also possible to make a determination using a power difference other than the power ratio.
[0044]
The area information determined as the cavity candidate by the signal strength determination unit 71 is passed to the radar data replacement unit 72. In the radar data replacement means 72, based on the position information (START, END, TOP, BOTTOM) of the second cavity candidate region selected by the second cavity candidate selection unit 70, the buffer data in the data buffer storage unit 32 is converted into the candidate region. Replace with the average value in the vicinity (S036).
[0045]
The position information of the candidate area selected by the above procedure is accumulated in the candidate area information storage unit 37 as described above. When the selection of the candidate area is completed, the cavity candidate narrowing unit 80 narrows down the cavity candidates (S040). For narrowing down the cavity candidates, the polarity determination means 81, the manhole detection means 82, and the candidate area narrowing means 83 are applied to all candidate areas stored in the candidate area information storage unit 37 in any order. The application order is controlled by the control unit 20. For example, an application procedure of the polarity determination unit 81, the manhole detection unit 82, and the candidate area narrowing unit 83, or an application procedure of the manhole detection unit 82, the polarity determination unit 81, and the candidate area narrowing unit 83 can be used. When candidate narrowing by the cavity candidate narrowing unit 80 is completed, the position information of the cavity candidate region accumulated in the candidate region information storage unit 37 is transferred to the candidate region information temporary storage unit 38 (S045), and the candidate region information storage unit 37 All the candidate area information stored in (1) is deleted (S050).
[0046]
The polarity determination means 81 determines the polarity of the cavity candidate α data accumulated in the candidate area information storage unit 37. A pattern recognition method can be used as the polarity determination method. For example, when the phase inversion part is represented by black and the non-inversion part is represented by white, the reflection pattern from the cavity region (with positive polarity) changes from white to black, and the reflection pattern from the non-cavity region (negative electrode) Is expressed in a pattern that changes from black to white. In addition, a method for analyzing the frequency spectrum of radar data including the small region position and the small region part forming the candidate region can be used. In addition, since it can be said that the reflection in the cavity region occurs exactly at the boundary, if the wavelength of the electromagnetic wave is sufficiently shorter than the size of the cavity region, this positional shift may be a problem. In this case, the processing of the polarity determination unit 81 may be performed for a range slightly wider than the hollow region.
[0047]
The manhole detection means 82 is a means for removing a manhole having the same polarity as that of the cavity from the candidate area. From the area information and α data stored in the candidate area information storage unit 37, for example, the depth (position) of the road surface 1 is compared with the TOP position of the candidate area, and the TOP position of the area is above the ground surface position. There is an area where the length from the TOP to the BOTTOM of the area is equal to or larger than the reference value, and is deleted from the candidate area information storage unit 37. For example, the depth d of the road surface 1_surface(l) is the α buffer signal u at distance l_αWith regard to (l, y), when the minimum y among the y giving the maximum value exceeding the predetermined value is d ′ (l), it can be obtained by the following equation:
[Expression 15]
d_surface(l) = d ′ (l) −η
Here, η is a constant determined based on the wavelength of the electromagnetic wave pulse. Note that it is also possible to create a list of only manhole candidate areas by separately storing information on the erased areas.
[0048]
The candidate area narrowing means 83 narrows down the cavity candidate areas by excluding the areas accumulated in the candidate area information storage unit 37 whose size is not more than a certain reference value from the cavity candidates. For example, when processing the width of a small region as 80 cm, after integrating the region information by the region information integration unit 61, if candidates whose width of the candidate region is equal to or smaller than the width of the small region are excluded, the cavity candidate region is effectively Confirm that the filtering can be executed.
[0049]
When the above procedure (S005-S050) is completed, 1 is added to the analysis number counter M completed so far (S055). This M is compared with a predetermined number of analysis times N. If M> N, an analysis end process is performed. Otherwise, the re-analysis operation (the second and subsequent analysis operations) is performed (S060). In the second and subsequent reanalysis operations, the analysis processing of S020 to S050 is performed using the buffer data already accumulated in the data buffer 32.
[0050]
As the setting of the number N of analysis operations, an integer of 1 or more is designated. However, when 1 is set, if the procedure from S005 to S050 is executed only once as described above, a place where the intensity of the reflected wave is equal to or greater than that of the hollow area, such as a puddle in the ground, is in the vicinity of the hollow area. If present, this hollow region may be determined as a non-hollow region by signal strength determination and overlooked. Therefore, here, by setting N to 2 or more, the cavity candidate is selected again by the procedure of S020-S050 using the data excluding the candidate portion detected by the radar data replacement means 72, and the cavity region is selected. Deter oversight. Data excluding the detected candidate portion is generated by the radar data replacement means 72 and stored in the data buffer 32. Therefore, α data and β data may be newly generated from the buffer data in the data buffer 32 and analyzed by the data analysis device 30. It should be noted that the number N of analysis operations can be set to a large value and re-selection of the cavity candidates can be performed until no new cavity candidate is selected, but in many cases, if N = about 3, effective results are obtained. It is confirmed that it can be obtained.
[0051]
If M> N in the analysis end determination S060, the process proceeds to an analysis end process. In the analysis end process, first, the candidate area information accumulated in the candidate area information temporary storage unit 38 is transferred to the candidate area information storage unit 37 (S065). Then, the candidate area information is integrated by the area information integration means 61 with respect to the candidate area information transferred / stored in the candidate area information storage unit 37 (S070). When the integration of the candidate area information (S070) is completed, the candidate area information accumulated in the candidate area information storage unit 37 is output to the display unit 39 or the output unit 40 as the processing result of the data analysis device 30 (S075). ).
[0052]
As described above, the operation flow of the present embodiment has been described according to FIG. 3. However, the application order of the candidate area information integration unit 60, the second cavity candidate selection unit 70, and the cavity candidate narrowing unit 80 subsequent to the first cavity candidate selection unit 50 is as follows. This is not the case. For example, the order of application of the first cavity candidate selection unit 50, the polarity determination unit 81 of the cavity candidate narrowing unit 80, the candidate region information integration unit 60, the second cavity candidate selection unit 70, and the cavity candidate narrowing unit 80 may be adopted. Note that when the application order is changed in this way, the processing for the hollow region stored in the candidate region information storage unit 37 in each unit is also changed to be performed on the hollow region updated in the immediately preceding processing. Needless to say.
[0053]
In the above description, the first cavity candidate selection unit 50 determines the first cavity candidate based on the (or average) reflection intensity of the surrounding area at the same depth with respect to the small area. However, when the reflection intensity in such a peripheral region is uniform to some extent in the depth direction, the peripheral region in the vertical direction can be taken into consideration, and thus the limitation of the same depth is not necessarily required.
[0054]
【The invention's effect】
With the underground cavity detection device of the present invention, it is possible to improve the accuracy of automatic detection of the cavity region.
[Brief description of the drawings]
FIG. 1 is a block diagram showing an apparatus configuration of an embodiment.
2 is a block diagram showing a configuration of a data processing unit 36 in FIG.
FIG. 3 is a flowchart illustrating an example of a processing flow.
[Explanation of symbols]
10 electromagnetic wave radar, 20 control unit, 30 data analysis device, 31 preprocessing filter, 32 data buffer, 33 α data storage unit, 34 high-pass filter, 35 β data storage unit, 36 data processing unit, 37 candidate region information Storage unit, 50 first cavity candidate selection unit, 60 candidate area information integration unit, 70 second cavity candidate selection unit, 80 cavity candidate narrowing unit.

Claims (6)

From the received data reflection intensity of the underground in the depth direction by the radar it has been measured for receiving a reflected wave by radiating electromagnetic waves into the ground a ground cavity detecting device for detecting a cavity in the ground,
A reception data storage unit for storing the reception data sequentially measured at a plurality of points along a predetermined traveling direction as radar data;
A data analysis unit for processing the radar data and detecting a cavity;
With
The data analysis unit
A preprocessing filter for generating α data by applying a predetermined filter process for noise removal to the radar data;
A high-pass filter that removes a signal that constantly exists in the traveling direction is applied to the α data and generates β data; and
The small area is more set in the depth direction and the traveling direction, the reflection intensity of the β data in a small area, approximately in the β data in the reflected intensity or approximately the same depth of the β data in the peripheral region of the same depth A first cavity candidate selection means for determining whether or not the small area is a cavity area as compared with an average reflection intensity;
Compare the reflection intensity of the α data in the small area with the reflection intensity of the α data in the peripheral area at approximately the same depth or the average reflection intensity of the α data at approximately the same depth. A second cavity candidate selection means for determining whether or not
Including
An underground cavity detecting apparatus , wherein a cavity is detected in a small area determined as a cavity area by the first cavity candidate selecting means and the second cavity candidate selecting means . In the underground cavity detecting device according to claim 1,
The data analysis unit further includes an area information integration unit that integrates a plurality of small areas existing within a predetermined distance among the small areas determined to be a hollow area. . In the underground cavity detecting device according to claim 1 or 2,
The data analyzing unit further polarity determination means for determining the polarity of the α data in the small region near it is determined that the cavity region, excluding the small regions inversion is determined of the polarity from the detection target of the cavity An underground cavity detector characterized by including. In the underground cavity detection device according to any one of claims 1 to 3,
The data analysis unit further includes radar data replacement means for generating new radar data by replacing the reflection intensity of the radar data with the average intensity of the surrounding area in the small area where the cavity is detected,
An underground cavity detecting apparatus , wherein a cavity is newly detected based on radar data generated by the radar data replacing means . The underground cavity detection device according to any one of claims 1 to 4,
The data analysis unit further includes a manhole detecting means for excluding the small region whose shallowest part is shallower than a predetermined depth and whose deepest depth is deeper than the predetermined depth as a manhole from a detection target of the cavity. Detection device. In the underground cavity detection device according to any one of claims 2 to 5,
The data analysis unit further includes candidate area narrowing means for excluding a small area smaller than a predetermined size from a hollow detection target with respect to the small area determined as a hollow area after being integrated by the area information integration means. An underground cavity detector characterized by that.

Images