EP4260284A1 - Procédé de calibrage d'une caméra ultra grand angle - Google Patents
Procédé de calibrage d'une caméra ultra grand angleInfo
- Publication number
- EP4260284A1 EP4260284A1 EP21834774.8A EP21834774A EP4260284A1 EP 4260284 A1 EP4260284 A1 EP 4260284A1 EP 21834774 A EP21834774 A EP 21834774A EP 4260284 A1 EP4260284 A1 EP 4260284A1
- Authority
- EP
- European Patent Office
- Prior art keywords
- sun
- camera
- celestial body
- photograph
- correction
- Prior art date
- Legal status (The legal status 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 status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/80—Analysis of captured images to determine intrinsic or extrinsic camera parameters, i.e. camera calibration
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10016—Video; Image sequence
Definitions
- the present invention relates to ultra wide angle cameras, also called “fisheye” cameras for “fish eye” in English. More particularly, the invention relates to a method for calibrating an ultra-wide-angle camera. Calibration is synonymous with standardization, often called calibration by reference to its English translation.
- the calibration of an ultra-wide-angle camera is of course very useful, even necessary, to enable fine and precise processing of the data acquired by the camera.
- the invention aims to satisfy this need and it achieves this, in whole or in part, thanks to a method for calibrating an ultra-wide-angle camera comprising a sensor and a lens, positioned at a given location while facing the sky, the method comprising the following steps:
- Step a take a series of panoramic photographs of the dome of the sky, in particular at regular intervals, for one or more days, preferably without clouds,
- Step b locate on each photograph the position of a celestial body, in particular the sun, and determine, using said series of photographs, the observed trajectory of the celestial body, in particular the sun, during said said days,
- Step c calculate, from a database of the expected position of the celestial body, in particular of the sun, on the day and at the time of taking the photographs and the position of the camera, the theoretical position of the celestial body, in particular the sun, in space for each photograph,
- the implementation of the final step is based on some or all of the following system of equations representative of the various corrections: with :
- “Ultra-wide-angle camera” means a fisheye type camera, that is to say with an observation angle greater than 140°.
- the celestial body is preferably chosen from the group consisting of the sun, the moon, and any celestial body insofar as the characteristics of the camera allow its identification and its positioning and where the relative stellar positions are predictable with precision during time.
- the celestial body is preferably the sun.
- step a is preferably implemented, in particular between sunrise and sunset at said given location, on one or more days when the course of the sun is the longest, i.e. i.e. a summer day, and one or more days when there are few or no clouds in the sky.
- the celestial body is used as a reference for the calibration, preferably the sun, which is a reliable source insofar as databases are available with high-performance algorithms making it possible to precisely calculate the theoretical position of the celestial body, in particular the sun, on which we wish to rely to calibrate the camera.
- the calibration according to the invention also makes it possible to take into account the lack of parallelization of the camera with respect to the sky. This defect still exists because it is impossible to install the camera perfectly aligned and there is always a camera manufacturing defect such that lens and sensor are not parallel.
- the steps of the calibration method according to the invention thus make it possible to have a precise calibration of the camera.
- step by step to arrive at obtaining the complete model.
- the number of photographs taken during the implementation of step a and processed may be between 20 and 1500 photographs, for example between 100 and 1500 photographs, in particular be between 500 and 1500 photographs, or even between 1000 and 1500 photographs .
- the reliability of the calibration can be improved.
- Steps a, b, c can be carried out simultaneously or not, in this order or not, step a being however necessarily implemented prior to step b.
- Step b may include determining the position of the celestial body, in particular the sun, in each photograph using the intensity of the celestial body, in particular of the sun, in a given region of the photograph, by image processing using in particular a thresholding-type algorithm, step b also preferably comprising the determination of the coordinates (x s , y s ) in pixels of the body celestial, in particular the sun, in the center of said region.
- Step c preferably comprises the determination, for each photograph, of the theoretical spherical coordinates (azimuth a, zenith ⁇ ) of the celestial body, in particular of the sun, by using a calculation algorithm on the basis of the position in longitude, latitude and camera altitude and based on the time and day the photograph was taken.
- Step c may optionally include the calculation of the theoretical trajectory of the celestial body, in particular of the sun, during said day or days of shooting, or take such a calculation into account.
- the calibration method according to the invention may further comprise a step d, for example after step c, consisting in carrying out, by comparison between the observed and theoretical trajectories of the celestial body, a correction of the center of each photograph, if necessary.
- Step d includes in particular the performance of an optimization, in particular using specific software, so as to make the theoretical trajectory and the observed trajectory of the celestial body, in particular of the sun, coincide with the theoretical position of the celestial body , in particular of the sun, with respect to the center of the photograph at a given moment being expressed by re, the observed position of the celestial body, in particular of the sun, with respect to the center of the photograph at a given moment being expressed by
- r photo ⁇ /(x — x c ) 2 + (y — y c ) 2 , with x c and y c the pixel coordinates of the center of the photograph, the objective of step d consisting preferably to minimize a function fi such that
- step d the offset of the center of the camera can be corrected.
- the center of the photograph is corrected by integrating the inaccuracy of pointing the camera on the point at the zenith of the celestial vault, as well as the inaccuracy of alignment between lens and sensor of the camera, which can be shifted by a few pixels between theory and practice. It should be noted that there is usually no possibility of completely correcting physically the centering of the camera, but it is possible to integrate this correction of the center in the rest of the calculations.
- n, and rphoto are distances expressed in pixels between the position of the celestial body, notably the sun, real and theoretical respectively, at a given moment, and that of the center of the photograph.
- the calibration method according to the invention may further comprise a step e, in particular after step c, consisting in determining, for example from the result of step d when present, a function representative of the distortion correction parameters proper on camera.
- Linear regression step e advantageously makes it possible to find a single solution for the coefficients ki and fe, initially not determined.
- the calibration method according to the invention may further comprise a step f, in particular after step c, for example after step e when present, consisting in determining the angle of azimuthal correction of the orientation of the camera in a universal landmark.
- This step corresponds to a correction made in view of the usual conventions, with the East direction being at the right end of the sky dome of a photograph and the West direction being at the left end of the sky dome in a photograph. 'a photograph.
- Step f therefore amounts to calculating the angle at sunrise of the celestial body, in particular of the sun, and to making the difference to correct the azimuth.
- Steps d, e, f are not necessarily all present, although this is preferable for the quality of the result obtained.
- steps d, e and f are preferably carried out in this order, i.e. step e after step d and step f after step e.
- a correction can be introduced to take into consideration a possible lack of parallelism between sensor and camera lens, with the parameters t, u, v.
- the final step we can perform a complete calibration of the camera, not limited to a local optimum, and taking into account the possible lack of parallelization between lens and sensor within the camera.
- the invention thus makes it possible to define a global optimal and not only a local optimal.
- Steps a to c, and, when present, steps d to f make it possible to determine the starting points of the optimization, to identify the initialization points of the global optimization for a complete and optimal calibration of the camera .
- the optimization is nonlinear and the parameters are interrelated. If one of them changes, the others change too.
- Steps d, e, f make it possible to improve the initialization of the corresponding parameters and thus make it possible to ensure convergence towards a global minimum. This convergence can however take place without these prior initialization steps d, e, and f.
- the invention it is sought in particular to find the following parameters: x c , y c , ki, k2, correctionioria, t, u and v.
- the parameters ki and allow you to correct the distortion, for example.
- the shift effect between the observation plane (the sky) and the sensor plane due to the orientation of the camera relative to the sky and the parallelization between the lens and the sensor is corrected in particular using the parameters x c , y c , /, u and v.
- the orientation is advantageously corrected with correction a .
- the implementation of the steps of the method according to the invention makes it possible to find the parameters and therefore to carry out the calibration of the camera.
- the calibration according to the invention makes it possible to correct a lack of parallelism between the camera and the sky, in particular if the camera has been incorrectly positioned. Calibration also makes it possible to alert of an accidental movement, or in time, of the camera, which allows a technician to come and put it back in place. Such a displacement can be slight, the precision being able to be within one or several millidegrees.
- the model of the camera is preferably recorded and, during each cloudless day, an evaluation of the model of the camera can be carried out with respect to the celestial body, in particular the sun, detected. If a significant difference is detected, it is deduced that the camera has moved, and it is then necessary to implement once again all or part of the calibration method according to the invention.
- At least one step of the calibration method according to the invention in particular the final step, can be repeated, for example at regular intervals, for example every day, in particular based on the results of the final step previously implemented . This makes it possible to quickly identify and diagnose a possible problem.
- steps a, b, c which characterize the possible evolution of the position or the characteristics of the camera during course of time. This may indeed have been moved or see its characteristics change, for example in connection with vibrations, changes in temperature or other.
- the final step can then be repeated based on the new data from steps a, b, c and taking, as initialization values from the parameters of the final step, the values found during the final step previously implemented.
- the values of the parameters found during steps d, and/or e, and/or f, of the previous iteration can also be used as initialization parameters.
- a further subject of the invention is an installation for calibrating an ultra-wide-angle camera for implementing the calibration method as defined above, the installation comprising an ultra-wide-angle camera oriented towards the sky and configured to photograph the sky, in particular at regular and/or predefined intervals, for example every minute, during one or more predefined days, a memory for recording the photographs taken by the camera, an image processing device configured to process the photographs taken by the camera, a database making it possible to calculate a theoretical position of the celestial body, in particular of the sun, on a photograph according to a position of the camera and a day and an hour given, and a mechanism for adjusting the relative position between the lens and the sensor of the camera.
- Figure 1 partially shows, in cross section and in an exploded schematic view, the block diagram of an example of an ultra-wide-angle camera to be calibrated
- FIG 2 figure 2 schematically represents in block diagram different stages of the calibration process according to the invention
- Figure 3 is a photograph of the sky dome taken by an ultra-wide-angle camera, the photograph illustrating the position of the sun,
- Figure 4 is a graph illustrating the elevation of the sun as a function of azimuth
- FIG 5 figure 5 is a photograph taken by an ultra wide angle camera, on which the theoretical trajectories of the sun are represented for three different days,
- figure 6 is a graph illustrating the theoretical trajectories of figure 5 and the observed trajectory of the sun using the camera, before optimization,
- FIG 7 represents a portion of a photograph taken by the camera on which the observed center and the corrected center after optimization have been illustrated
- figure 8 represents the graph of figure 6 on which the theoretical trajectories and the observed trajectory appear after optimization
- Figure 9 is a graph illustrating the result of a linear regression carried out in step e of the method according to the invention from the theoretical trajectories and observed in Figures 5, 6 and 8,
- Figure 10 is a graph illustrating the superposition of the theoretical and observed trajectories after implementation of the method according to the invention.
- Figure 11 is a bar graph illustrating the error distance of the position of the sun in pixels as a function of the number of occurrences after implementation of the method according to the invention
- Figure 12 is a bar graph illustrating the error distance of the abscissa position in pixels as a function of the number of occurrences after implementation of the method according to the invention.
- Figure 13 is a bar graph illustrating the error distance of the ordinate position in pixels as a function of the number of occurrences after implementation of the method according to the invention.
- FIG 1 Illustrated in Figure 1 a block diagram of camera 1 ultra wide angle also called “fisheye", viewing angle greater than 140 °, the calibration is performed using the calibration method according to invention.
- the hemispherical lens 2 and the sensor 3 are visualized in particular, as well as an optical system 4 arranged between the lens 2 and the sensor 3 to allow the beam trajectories to be reoriented as visible in this figure.
- Camera 1 used in the calculation example is a camera known as HIKVISION DS-2CD2955FWD-I marketed by the company HIKVISION.
- the camera 1 is positioned at a given location on a fixed support while facing the sky. We know the longitude, latitude and altitude of the camera.
- FIG. 2 schematically illustrates seven steps numbered from a to h of an exemplary method according to the invention.
- Step a consists in taking, and recording in a memory, a series of panoramic photographs of the dome of the sky at regular intervals in this example, for example every minute, for one or more days, in particular from sunrise to sunset, from one celestial body, in this example the sun, preferably cloudless.
- the day or days, in this example the three days, during which the photographs will be taken by the camera are chosen beforehand.
- summer days have been chosen because the path of the sun is longer on these days over the field of view of the camera. For each photograph, the precise time it was taken is recorded by being associated with this photograph.
- This step a advantageously makes it possible to obtain a large quantity of photographs, for example more than 500 photographs, in particular between 1000 and 1500 photographs, which contributes to the collection and storage of a large number of data in order to refine the calibration. of the camera.
- Step b includes marking the position of the sun on each photograph, and determining the observed trajectory of the sun during said day or days.
- step b the position of the sun in each photograph is sought using the intensity of the sun.
- the region 21 of the sun is often saturated by the luminosity of the sun. Thanks to image processing using a thresholding type algorithm, known per se, this region 21 can be determined.
- the position of the sun 22 in pixels is chosen to be the center of this region 21, as can be seen in FIG.
- the position of the sun in pixels is denoted by the Cartesian coordinates x s and y s .
- Step c consists in calculating, from a database of the expected position of the sun on the day and at the time of taking the photographs and the position of the camera, the theoretical position of the sun in the space for each photograph, and to calculate the theoretical trajectory of the sun during said day or days.
- Figure 4 illustrates a graph using data from such a database showing the elevation of the sun as a function of its azimuth, expressed in degrees, for certain pre-selected days.
- the azimuth of the sun is denoted a.
- the elevation corresponds to 90° - 9, with 9 the zenith of the sun.
- zenith 9 90°- elevation.
- this type of data can be found on the website www.sunearthtools.com.
- Step d aims to carry out, by comparison between the observed and theoretical trajectories of the sun, a correction of the center of each photograph, if necessary.
- the camera is typically positioned when it is installed to be centered towards the zenith, that is to say towards the center of the sky vertically from the observation point.
- the observed position of the sun on the/each photograph can be described by the equation: [Math the pixel coordinates of the center of the photograph.
- the optimization is done for example by fmincon type function with the Matlab software.
- the optimization makes it possible to find the new center of the photograph and, after optimization, the cubic function is more stuck on the real curve.
- the corrected center 26' of the photograph is displayed in FIG. 7.
- FIGS. 6 and 8 represent the trajectories, theoretical 25 and observed 27, of the sun, reported in the same frame, with an abscissa and an ordinate expressed in pixels, respectively before and after optimization.
- the curves 25 and 27 are substantially dissociated in Figure 6, before optimization, while they substantially overlap in Figure 8. Only a few aberrations 30 are visible around the single trajectory visible in this figure.
- Step e which follows step d makes it possible to determine a linear function representative of the correction parameters specific to the camera from the result of step d.
- This step aims to determine the distortion correction by an analytical function.
- Step f consists in determining the angle of horizontal correction of the orientation of the camera in a universal frame. Indeed, the camera can be oriented with a different angle from the south. This step thus makes it possible to correct this difference. Since the camera is stationary, this is an easy step.
- Step g makes it possible to introduce a correction to take into consideration a possible lack of parallelism between sensor and lens of the camera.
- This correction of parallelism can be made by a linear correction on the coordinates in x and y of the position of the sun by the introduction of factors t, u, v such as:
- step h involves performing a global optimization of the camera calibration from steps d to f.
- Step h includes solving all of the following equations:
- Correction and y c° rrectlon form the coordinates of the sun as a function of the zenith 9, on the photograph and in a theoretical way taking into account a correction of a possible lack of parallelization between the lens and the sensor within the camera.
- Figure 10 shows the trajectories of the theoretical and observed sun superimposed, after optimization in a Cartesian coordinate system in pixels. It can be seen that only a few aberrations 31 remain, due for example to a cloudy period or to dirt on the camera when the photograph was taken.
- Figure 11 illustrates the number of occurrences as a function of the deviation from the expected distance which is 0. We see an expected peak at 0 and a slight dispersion on either side of this peak, which corresponds to the error distance of the pixel sun.
- Figure 12 details these deviations on the X axis, illustrating the number of occurrences as a function of distance deviation.
- Figure 13 details these deviations on the Y axis by illustrating the number of occurrences as a function of the distance deviation.
- steps a, b, c which characterize the possible evolution of the position or the characteristics of the camera during course of time. This may indeed have been moved or see its characteristics change, for example in connection with vibrations, changes in temperature or other.
- the values of the parameters found during steps d, and/or e, and/or f, of the previous iteration can also be used as initialization parameters.
- the invention is not limited to the example which has just been described.
- a celestial body other than the sun for example the moon or any celestial body in relative motion capable of being identified and followed in its movement by the field of view of the camera.
- Steps d, e and f may not be implemented, in which case only steps a, b and c and the final step are implemented to perform the calibration, also including step g. Only one or some of steps d, e and f can be implemented.
Landscapes
- Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Studio Devices (AREA)
- Stereoscopic And Panoramic Photography (AREA)
- Testing, Inspecting, Measuring Of Stereoscopic Televisions And Televisions (AREA)
Abstract
Description
Claims
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| FR2013148A FR3117653B1 (fr) | 2020-12-14 | 2020-12-14 | Procédé de calibrage d’une caméra ultra grand angle |
| PCT/EP2021/084770 WO2022128677A1 (fr) | 2020-12-14 | 2021-12-08 | Procédé de calibrage d'une caméra ultra grand angle |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| EP4260284A1 true EP4260284A1 (fr) | 2023-10-18 |
Family
ID=74554069
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| EP21834774.8A Pending EP4260284A1 (fr) | 2020-12-14 | 2021-12-08 | Procédé de calibrage d'une caméra ultra grand angle |
Country Status (3)
| Country | Link |
|---|---|
| EP (1) | EP4260284A1 (fr) |
| FR (1) | FR3117653B1 (fr) |
| WO (1) | WO2022128677A1 (fr) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN118967834B (zh) * | 2024-08-08 | 2025-12-09 | 蜻蜓智能视界科技(深圳)有限公司 | 单芯片标定多超广角镜头实现多图像融合无畸变的方法 |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2018055619A1 (fr) * | 2016-09-25 | 2018-03-29 | Israel Aerospace Industries Ltd. | Compas céleste et méthode d'étalonnage |
-
2020
- 2020-12-14 FR FR2013148A patent/FR3117653B1/fr active Active
-
2021
- 2021-12-08 EP EP21834774.8A patent/EP4260284A1/fr active Pending
- 2021-12-08 WO PCT/EP2021/084770 patent/WO2022128677A1/fr not_active Ceased
Also Published As
| Publication number | Publication date |
|---|---|
| WO2022128677A1 (fr) | 2022-06-23 |
| FR3117653A1 (fr) | 2022-06-17 |
| FR3117653B1 (fr) | 2023-04-14 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| EP3317736B1 (fr) | Procédé de pilotage de l'orientation d'un suiveur solaire basé sur des modèles cartographiques | |
| EP1828992A1 (fr) | Procede de traitement d'images mettant en oeuvre le georeferencement automatique d'images issues d'un couple d'images pris dans le meme plan focal | |
| EP3628939A1 (fr) | Procédé de pilotage prédictif de l'orientation d'un suiveur solaire | |
| FR3054897A1 (fr) | Procede d'elaboration d'une image numerique, produit programme d'ordinateur et systeme optique associes | |
| EP3290891B1 (fr) | Procédé et dispositif de caractérisation des aberrations d'un système optique | |
| EP3112800A2 (fr) | Procede de mesure d'au moins une dimension d'un objet | |
| CN112857356A (zh) | 无人机水体环境调查和航线生成方法 | |
| EP2695377B1 (fr) | Detection de debris spatiaux | |
| CN108291742A (zh) | 校准热电太阳能发电厂的定日镜 | |
| EP1168831A1 (fr) | Procédé et dispositif de calibration de caméra | |
| WO2022128677A1 (fr) | Procédé de calibrage d'une caméra ultra grand angle | |
| FR3006296A1 (fr) | Drone comprenant un dispositif imageur multispectral pour la generation de cartes representant un etat vegetal d'une culture | |
| FR3015649A1 (fr) | Procede d'evaluation des parametres de pilotage d'un suiveur solaire | |
| FR3047829A1 (fr) | Procede d'estimation de la position du disque solaire dans une image de ciel | |
| FR2985307A1 (fr) | Procede d'etalonnage des biais d'alignement d'un systeme d'observation de la terre exploitant des prises de vue symetriques | |
| EP3018625B1 (fr) | Procédé de calibration d'un système de visée | |
| FR3028989A1 (fr) | Procede de calibration des parametres extrinseques d'une camera d'un vehicule automobile et systeme d'assistance a la conduite associe | |
| FR3032820A1 (fr) | Procede de calibration d'une camera montee sur le pare-brise d'un vehicule | |
| EP2746830A1 (fr) | Mise au point optique d'un instrument de saisie d'image | |
| EP3999865A1 (fr) | Procede de determination de parametres d'etalonnage extrinseques d'un systeme de mesure | |
| FR3047830A1 (fr) | Procede de determination de la direction de deplacement d'objets dans une scene | |
| FR2611266A1 (fr) | Dispositif et procede de localisation de vues au sol prises d'un satellite | |
| CN105571828B (zh) | 一种用于太阳光球或色球望远镜平场测量的装置及方法 | |
| FR3063559A1 (fr) | Procede et dispositif electronique de determination d'un modele en perspective d'un objet, programme d'ordinateur et appareil electronique d'affichage associes | |
| FR3151085A1 (fr) | Procédé de guidage vers une destination |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| STAA | Information on the status of an ep patent application or granted ep patent |
Free format text: STATUS: UNKNOWN |
|
| STAA | Information on the status of an ep patent application or granted ep patent |
Free format text: STATUS: THE INTERNATIONAL PUBLICATION HAS BEEN MADE |
|
| PUAI | Public reference made under article 153(3) epc to a published international application that has entered the european phase |
Free format text: ORIGINAL CODE: 0009012 |
|
| STAA | Information on the status of an ep patent application or granted ep patent |
Free format text: STATUS: REQUEST FOR EXAMINATION WAS MADE |
|
| 17P | Request for examination filed |
Effective date: 20230524 |
|
| AK | Designated contracting states |
Kind code of ref document: A1 Designated state(s): AL AT BE BG CH CY CZ DE DK EE ES FI FR GB GR HR HU IE IS IT LI LT LU LV MC MK MT NL NO PL PT RO RS SE SI SK SM TR |
|
| DAV | Request for validation of the european patent (deleted) | ||
| DAX | Request for extension of the european patent (deleted) | ||
| RAP3 | Party data changed (applicant data changed or rights of an application transferred) |
Owner name: COMMISSARIAT A L'ENERGIE ATOMIQUE ET AUX ENERGIESALTERNATIVES |
|
| GRAP | Despatch of communication of intention to grant a patent |
Free format text: ORIGINAL CODE: EPIDOSNIGR1 |
|
| STAA | Information on the status of an ep patent application or granted ep patent |
Free format text: STATUS: GRANT OF PATENT IS INTENDED |