FR2992414A1 - METHOD FOR ESTABLISHING A GEOLOCALIZED RECOVERY PLAN, COMPUTER PROGRAM PRODUCT AND CORRESPONDING STORAGE MEDIUM - Google Patents

Abstract

There is provided a method of establishing a network evaluation plan, the network comprising a plurality of markup targets disposed in the vicinity of an apparent element of said network, the method comprising the following steps: - obtaining a plurality of representative images, from different angles of view, of an inspection scene comprising said element; photogrammetric processing of said plurality of images obtained, consisting in: detecting and locating markup targets in the images, orienting the images and determining in the space of locations and orientations of the camera having served capturing images, modeling the network element, and determining, by geo-referencing, an absolute positioning of the network element in a topographic repository.

Description

FIELD OF THE INVENTION FIELD OF THE INVENTION The field of the invention is that of civil engineering.

More specifically, the invention relates to a technique for establishing a geolocated inventory plan. A test is an operation to establish a geographical representation of the configuration of a network, such as a network of water, gas, electricity, or data for example.

The invention applies particularly, but not exclusively, to networks in open excavations. 2. TECHNOLOGICAL BACKGROUND Traditionally, reworking operations require the simultaneous presence, at the excavation site, diggers and surveyors (or a deferred action of surveyors without closing the trench), which makes these operations complex to organize. since they involve at least two trades. In addition, the necessary presence of surveyors at the excavation site during the time required for their surveys makes the execution of this operation relatively costly, given the high degree of qualification of at least some of the surveyors. stakeholders. In addition, the excavations must remain open until the surveyors have completed their surveys, which can cause a durable gene and therefore important for residents and users of the road in which the excavations were conducted. This procedure can therefore be particularly long, binding and relatively expensive. Consequently, it appears particularly interesting to be able to propose a technique that makes it possible to simplify and make less costly the operations of evaluation in open excavations. OBJECTIVES OF THE INVENTION The invention, in at least one embodiment, has the particular objective of overcoming these various disadvantages of the state of the art.

More specifically, in at least one embodiment of the invention, an objective is to provide a technique that simplifies and reduces the time required for the proofing operations, and in particular to speed up the closing of open excavations.

At least one embodiment of the invention also aims to provide a technique that does not require the intervention of a surveyor at the excavation site during its implementation. SUMMARY OF THE INVENTION In a particular embodiment of the invention, there is provided a method for establishing a network evaluation plan, the network comprising a plurality of markup targets arranged in the vicinity of the network. an apparent element of said network, the method comprising the steps of: obtaining a plurality of representative images, from different viewing angles, of an inspection scene comprising said element; photogrammetric processing of said plurality of obtained images, comprising: identifying markup targets in each of the resulting images; orienting, in space, by triangulation, at least two of said images obtained according to said identified markup targets; determining, in space, locations and shooting orientations used to obtain the images, according to said at least two oriented images; provide a three-dimensional modeling of said network element; determining, by geo-referencing, an absolute positioning of said network element in a topographic reference system; establishing a network evaluation plan comprising said modeled and georeferenced network element. The general principle is based on a virtual reconstruction, by photogrammetry, of an inspection scene of a network comprising at least one apparent element, with a view to producing a three-dimensional network evaluation plan in a given topographic reference frame (such as for example, the national reference system). Thanks to the invention, the establishment of the location plan is fully automatic and can be done asynchronously, that is to say possibly away from the inspection scene, and therefore possibly after it was rendered inaccessible after filling the excavations.

According to an advantageous characteristic, the identification step comprises the following steps: detecting and locating, in each of the images of said plurality of images, markup targets, based on graphic information supported by the targets; identifying, in each of the images of said plurality of images, markup targets, based on alphanumeric information supported by the targets. This makes it possible to obtain, for each identified target, the two-dimensional coordinates of each of the vertices of said target. Advantageously, the orientation step comprises a step of determining, for each identified target, the position in the space of said target identified as being common to said at least two images. This makes it possible to obtain, for each identified target, the three-dimensional coordinates of each of the vertices of said target. In another embodiment of the invention there is provided a computer program product which comprises program code instructions for implementing the aforesaid method (in any one of its various embodiments), when said program is run on a computer. In another embodiment of the invention, there is provided a computer-readable and non-transitory storage medium, storing a computer program comprising a set of instructions executable by a computer for implementing the aforementioned method (in any of its different embodiments). 5. LIST OF FIGURES Other features and advantages of the invention will become apparent on reading the following description, given by way of indicative and nonlimiting example, and the appended drawings, in which: FIG. 1 shows a flowchart a particular embodiment of the method according to the invention; FIG. 2 shows a simplified diagram of the structure of a device implementing the method of establishing a proofing plan according to the invention in the particular embodiment presented in FIG. 1; FIG. 3 shows an example of an inspection scene whose production plan is to be obtained by means of the method according to the invention; FIG. 4 represents a virtual representation of the result of the execution of an image orientation step according to a particular embodiment of the invention; FIG. 5 represents the inspection scene superimposed on a virtual representation established by implementing the method according to the invention, after the position of the network elements has been determined; FIG. 6 illustrates the virtual representation with a 3D reference specific to the inspection scene obtained after modeling; FIG. 7 illustrates a three-dimensional proofing plan obtained after a network element has been modeled and georeferenced in a topographic reference system (3D); FIG. 8 represents the superimposed inspection scene with the recovery plan obtained in FIG. 7. DETAILED DESCRIPTION In all the figures of the present document, the elements and identical steps are designated by the same numerical reference. The general principle is based on a virtual reconstruction, by photogrammetry, of an inspection scene of a network comprising at least one apparent element, with a view to producing a three-dimensional network evaluation plan in a given topographic reference frame (such as for example, the national reference system). Thanks to the invention, the establishment of the location plan is fully automatic and can be done asynchronously, that is to say possibly away from the inspection scene, and therefore possibly after it was rendered inaccessible after filling the excavations.

FIG. 1 presents a flowchart of a particular embodiment of the method according to the invention.

The method of the invention aims to establish, in a fully automatic manner, a plan for collecting a geolocated network. It can be broken down into two main phases: a first location of images, and a second phase of dimensional exploitation of images.

The first location phase, referenced 110 (steps 10 to 60), aims to determine the six degrees of freedom (that is to say the location and orientation in space) of the camera with served to capture images from the inspection scene including a buried network. An example of an inspection scene 300 of an open-excavation network whose production plan is to be obtained using the method according to the invention is illustrated in FIG. 3. The second phase of dimensional exploitation, referenced 120 (steps 70 and 80), aims to determine the absolute positioning of one or more apparent elements belonging to a buried network (such as ducts, ducts or one for example), positioning whose coordinates are expressed in a topographic reference system by means of a geo-referencing step. The main steps of the resuming plan generation algorithm described in FIG. 1 are described below. In a step 10, a set of n representative images is first obtained in a step 10. different viewing angles, of an inspection scene 300. All the images have been previously captured by an operator according to convergent shots of the inspection scene 300 by means of a digital camera. classic. The configuration of the inspection scene 300 is stored, it is then possible to fill the excavations and cover the NTE network element that is to be modeled on a geolocated location plan, its modeling can be, thanks to the invention, performed subsequently (asynchronously), by performing steps 10 to 80 of photogrammetric processing of the images of the inspection scene 300 that have been captured.

Then, in a step 20, detection and localization, in each of the images obtained in the previous step 10, markup targets present on the inspection scene and whose principle is illustrated below in FIG. 3 represents an inspection scene 300 of an open-excavation network which has been captured from different angles of view and which it is desired to obtain a recovery plan according to the method of the invention. . The network comprises a plurality of markup targets DBi (for i = 0 to 6) that have been arranged in the vicinity of an apparent NTE network element. The latter consists here of two substantially parallel branches and an open loop. Each markup target DBi is of tetrahedral form, each of whose faces serves as support for alphanumeric information, consisting of a target number Oi (for i = 0 to 6), a number j (for j = 0 to 2) of the face of said target Oi. Each of the faces of the target serves to support graphic information having a color distinct from the colors present in the background of the inspection scene, in this case a colored outline of each face of red color, intended to allow better distinguish this face from the background that will appear on the captured images where the target will appear. The target is a spatial marker constituting a singularity in the inspection scene. It can therefore be detected automatically on a captured image. Thus, in step 20 of FIG. 1, the detection, in each of the n images obtained, of the markup targets is carried out by means of a colorimetric analysis of said image. For this, each of the faces of each target supports graphic information having a sufficiently detectable detection color with respect to the environment of the inspection scene. In other words, the algorithm performs a search, in each captured image, of triangles of predetermined color (for example a red color detection triangle). The location (or positioning) of the face of the target in each image is then determined by means of a contour detection of the graphic information (detection of outline of the red detection triangle for example). For this, the algorithm determines, by edge detection of each detected triangle, a set of coordinates for each point forming a vertex of a detected triangle. For each detected triangle, we obtain three sets of two-dimensional coordinates (x1, y1, x2, y2, x3, y3).

Next, the algorithm proceeds, in a step 30, to the identification of each target by correlation between the alphanumeric information located inside the detection triangle (s) and reference information. To do this, the alphanumeric information is extracted, for each captured image, triangles detected and compared to reference information for identifying the corresponding targets (recognition of the number Oi (for i = 0 to 6) of the target). At the end of this step, each captured image comprises a plurality of triangles, each triangle having an identification number making it possible to identify the markup target to which it belongs and comprising three points each forming an apex of said triangle, each point being associated with a couple of two-dimensional coordinates (x, y). Steps 10 to 30 constitute a first sub-phase of the image localization phase, consisting of the extraction, by image processing (2D), of vertices of the tetrahedral targets detected and identified in each of the captured images. Steps 40 to 60 detailed below constitute a second sub-phase of the image localization phase aiming at a three-dimensional (3D) reconstruction of an inspection scene. In step 40, the algorithm proceeds to a relative orientation of the images captured from the coordinates of the vertices of the targets identified previously in each of the captured images. This orientation step aims at determining locations and orientations of the camera used to capture the images. The algorithm selects for this purpose a set of at least two images from the n images acquired and determines the relative position of the so-called right image with respect to the so-called left image: it detects, by triangulation of the said at least two images, targets identified as being common to said at least two images. This makes it possible to calculate, for each target identified and detected as being common to the said at least two images, the three-dimensional coordinates in the image frame (x, y, z) of the vertices belonging to said target. The result of the execution of step 40 is illustrated in FIG. 4.

More particularly, the algorithm implements the following steps: (0 step of triangulation of new vertices belonging to the targets, (ii) step of raising new unselected images among the n captured images, (iii) step of triangulation of new vertices belonging to the targets; (iv) adjustment of beams of a first type.

Steps (i) to (iv) above are performed iteratively as the system evolves (propagation of the geometric model). The evolution of the system means that new points (vertices) and / or new images are in turn positioned and integrated into the model. Triangulation here means determining the three-dimensional coordinates of a point (or vertex) seen by at least two images (that is to say, identified by the algorithm as being the projection of the same point on at least two images) among the n captured images, the two-dimensional coordinates of which were determined in step 20. By raising is meant here the determination of the 6 degrees of freedom of a captured image having points in common ( at least four points (or vertices)), the two-dimensional coordinates of which were determined in step 20. This raising step allows to triangulate new points whose coordinates have not yet been calculated or new images that do not have not been located yet.

By beam adjustment is meant here the determination of a parameter vector from a system of measurement equations. This deterministic modeling exercise is based on the existence of functional relationships uniting thus observed variables and determined parameters. The vector of parameters is broken down into three sub-vectors: sub-vector of the three-dimensional coordinates (x, y, z) of the points; - sub-vector of the 6 degrees of freedom of the images; - sub-vector of the internal parameters of the digital camera. The internal settings of a camera (or video camera) include optical parameters, such as focal length, sensor characteristics (CCD / CMOS, size, resolution, etc.) and distortion parameters . The redundancy of the number of equations with respect to the size of the parameter vector makes it possible to reject the measurement errors and to estimate the reliability of the results obtained. The observation vector consists of the two-dimensional coordinates of the points on the images; these coordinates are associated with uncertainties (which are a function of the extraction process: manual or image processing operator). The mechanism of error rejection obviously takes into account these uncertainties. Thus, at the beginning of the 3D reconstruction, when the geometric model is still under development (during propagation, determined by a low number of equations), it is absolutely necessary not to be too restrictive on these uncertainties. The beam adjustment performed in step (iv) provides an uncertainty in calculating vertex coordinates equal to about 10 pixels. The algorithm proceeds, in step 50, to the scaling of the geometric model obtained. A beam adjustment is made to a near homothetic factor (intrinsic to the equations of the photogrammetric functional model). It is therefore necessary to scale the geometric model obtained, simply by relying on known distances between measured points. The algorithm based on the inter-vertex distances of the tetrahedral targets, whose geometry through their manufacture is controlled. The non-suitability of the length of these inter-vertex distances with the geometrical extent of the measured volume must be taken into account as a function of the following characteristics: - very large number of inter-vertex distances (use preferably of at least ten targets); precision customer objective much lower than the measurement uncertainty; scaling possible when geo-referencing the measurement. The algorithm proceeds, in step 60, to a beam adjustment of a second type. This precisely of beams of second type makes it possible to determine the geometric model on the scale, by integrating all the observations (of all the images) and thus rejecting any inaccurate and erroneous observations. The model is expressed in an arbitrary repository called measurement reference. It is recalled here that the main objective of these steps 10 to 60 is the determination of the location of the images during the shooting and the mastery of the internal parameters of the digital camera, in order to exploit these images so to be able to position the business objects and thus establish a plan of geolocalised and 3D recovery. In addition, the algorithm proceeds to the dimensional exploitation phase (steps 70 and 80). The principle of implementation of these steps is also illustrated in Figures 6, 7 and 8. Step 70 consists of a three-dimensional modeling of apparent elements of the network. This modeling is based on the three-dimensional virtual representation of the inspection scene obtained at the end of steps 10 to 60 of the method. It consists in digitizing particular points on apparent elements (cables, ducts, pipes, boxes for example) that one seeks to position in a given topographical reference frame. Given the absence of homologous points (identifiable point on different images), the epipolar curve and digitalization on apparent contours are very widely used processes. Step 80 is to establish a geo-referenced proofing plan. Georeferencing is the expression of the measurements obtained in the French cartographic reference system (Lambert system). The proofing plan includes, in three dimensions, the apparent elements of the network. We determine the transformation (rigid or not) that allows to go from the arbitrary image reference to the cartographic reference. For this, we have measured points whose geographic coordinates are known (provided by a surveyor or extracted from existing data). This transformation (3D similarity) uses all or part of the coordinates of the geographical points (selection of the equations X, Y or Z), in order to take into account the topographic reality and thus, to obtain the best possible result. Determining this similarity by also calculating the scale factor (non-rigid transformation) amounts to scaling the geometric model. Comparing the scaling of georeferencing with that performed with tetrahedral targets is an excellent pre-delivery data check. FIG. 2 shows the simplified structure of a device implementing the method of establishing a proofing plan according to the invention (for example the particular embodiment described above in relation with FIG. 1). This device comprises a RAM 23, a processing unit 21, equipped for example with a processor, and controlled by a computer program stored in a ROM 22. At initialization, the code instructions of the program of For example, the computer is loaded into the RAM 23 before being executed by the processor of the processing unit 21. The processing unit 21 receives as input a plurality of images representing, from different angles, a scene of inspection, the images having been captured by means of a camera. The processor of the processing unit 21 processes the obtained images 20 and outputs a three-dimensional and geolocated proofing plan 24 in a topographic reference system, according to the instructions of the program 22. This FIG. 2 only illustrates one particular way, among several possible , to achieve the algorithm detailed above, in connection with Figure 1. Indeed, the technique of the invention is carried out indifferently: - on a reprogrammable calculation machine (a PC computer, a DSP processor or a microcontroller) executing a program comprising a sequence of instructions, or - on a dedicated computing machine (for example a set of logical gates such as an FPGA or an ASIC, or any other hardware module). In the case where the invention is implemented on a reprogrammable calculation machine, the corresponding program (that is to say the sequence of instructions) can be stored in a removable storage medium (such as for example a diskette, a CD-ROM or a DVD-ROM) or not, this storage medium being readable partially or totally by a computer or a processor. FIG. 4 is a virtual representation illustrating the result of the execution of an orientation step (the principle of which is detailed above in relation to step 40 of FIG. 1) of a set of images in FIG. a three-dimensional space (x, y, z). During this orientation step, locations POk (for k = 1 to 10) and directions of the camera used to capture the images are determined in space.

In the 3D virtual representation thus established, a set of particular points (singularities) representative of the position of an apparent element of the network is then determined, with a view to modeling this network element, as illustrated in FIGS. and 6, briefly described below. FIG. 5 represents the inspection scene superimposed on the virtual representation established by implementing the method according to the invention, after the position of the network element has been modeled. We observe in particular a perfect superposition between the positions of the vertices of the markup targets, for example the vertices S01, 502 and 503 of the markup target DBO and the points Nd1, Nd2 and Nd3 respectively representing these vertices in the virtual representation of the inspection scene given in Figure 6. It also observed the set of points Pt1, Pt2, Pt3 and Pt4 that the algorithm according to the invention has detected in the virtual representation as being representative of the network elements to be modeled. FIG. 7 illustrates a three-dimensional proofing plan obtained after an NTE network element has been modeled and georeferenced in a topographic reference frame (3D). The term "georeferencing" is understood to mean a step during which absolute positioning of a network element in a given topographic reference system (for example, the national reference system) is determined. The result of such a step appears in FIG. 6. FIG. 8 illustrates, in superposition with one of the images of the inspection scene 300, a virtual model of the NTE network element resulting from the determination of its position. vis-à-vis said 3D reference specific to the inspection scene. It is observed that the superimposition of the modeled element NTE (illustrated in FIG. 6) using the method according to the invention is superimposed perfectly on the network element in the inspection scene 300.25.

Claims (5)

REVENDICATIONS1. A method of establishing a network evaluation plan, the network comprising a plurality of markup targets disposed in the vicinity of an apparent element of said network, characterized in that it comprises the following steps: - obtaining a plurality of representative images, from different angles of view, of an inspection scene comprising said element; photogrammetric processing of said plurality of images obtained, consisting in: identifying markup targets in each of the images obtained; orienting, in space, by triangulation, at least two of said images obtained according to said identified markup targets; determining, in space, locations and shooting orientations used to obtain the images, according to said at least two oriented images; provide a three-dimensional modeling of said network element; determining, by geo-referencing, an absolute positioning of said network element in a topographic reference system; establishing a network evaluation plan comprising said modeled and georeferenced network element. 2. The method according to claim 1, wherein the identification step comprises the following steps: detecting and locating, in each of the images of said plurality of images, markup targets, based on graphic information supported by the targets; identifying, in each of the images of said plurality of images, markup targets from alphanumeric information supported by the targets. 3. Method according to any one of claims 1 and 2, wherein the orientation step comprises a step of determining, for each identified target, the position in the space of said target identified as being common to said at least two images. . A computer program product, comprising program code instructions for carrying out the method according to at least one of the claims 3, when said program is run on a computer. A computer-readable and non-transitory storage medium storing a computer program comprising a set of instructions executable by a computer or a processor for carrying out the method according to at least one of claims 1 to 3.