US20120238383A1

US20120238383A1 - Device and method for detecting and commenting on a game event in the immediate vicinity of a goal

Info

Publication number: US20120238383A1
Application number: US13/384,392
Authority: US
Inventors: Patrick Simonnet; Francis Bras
Original assignee: OMRI Tarek
Current assignee: OMRI Tarek
Priority date: 2009-07-16
Filing date: 2010-07-16
Publication date: 2012-09-20
Also published as: FR2948031A1; WO2011006989A1; EP2453992A1

Abstract

A device capable of automatically detecting game actions that take place in the vicinity of a goal for sports or activities in which the objective is to make a projectile enter the goal. The device is suitable for triggering, always in an automatic manner and in real time, audio or visual animations suited to the game situation thus detected. A method for processing information suitable for qualifying a game action according to information gathered from various types of sensors placed in the playing area is also described.

Description

The invention belongs to the field of the technical devices for audio-visual broadcast relating to games and play activities practised in a room. More precisely the invention relates to a device capable of detecting, in an automatic manner, actions that are taking place in a space of play and/or in the vicinity of a goal cage for sports or activities whose aim consists in making penetrate a projectile in the aforementioned cage. The device of the invention is capable of triggering, always in an automatic manner and in real time, visual or sound animations adapted to the situation of play thus detected. The invention also relates to a method of data processing capable of qualifying an action of play according to the information collected on various types of sensors of the device.
Prior art describes various devices and methods capable of detecting an action of play for sports such as soccer, said action consisting, for instance, in the crossing by a ball of the entrance plan of a goal cage, or the impact of the aforesaid ball with a goal post. These devices are sometimes capable of locating the input zone of the ball into the goal cage.
Such a device is described for example in the European patent application EP-A-1 911.499. The aforementioned device is used as a support for the training of soccer players while enabling them to file information relating to penalty kicks as an example. This device does not automatically qualify the actions and is limited to recording parameters relating to the trajectory of the ball, said parameters being processed later on, or simply interpreted by the player himself. The French patent application FR-A-2 806 924 describes a method and a device capable of automatically detecting the crossing by the ball of the entrance plan of a soccer goal cage, and to ensure that it is really the ball that crossed this plan. In this example also, the device does not qualify the shooting and is limited to sending to the referee, who will carry out the interpretation, an image of the crossing by the ball of the goal line. Under these conditions a lag of several seconds between the occurence of the event and the data processing is acceptable, and the complexity of the information delivered after this processing is reduced since the aforementioned treatment only aims at making the detection reliable but not at qualifying the event. Consequently, prior art devices rely on a complex signal processing in particular aiming at identifying the shape of the projectile and at calculating its trajectory, the overall objective being, in the two preceding examples, to validate and to characterize an entrance trajectory of the ball. These prior art devices rely on a temporal analysis in particular of images provided by a vision device. The time of acquisition required for conducting this analysis is an incompressible parameter. The detection reliability is obtained with the detriment of the computing time and the complexity of the implemented means. The analysis of image according to prior art, relying in particular on the detection of the shape of the projectile, implies the use of expensive high speed cameras, and of sufficient optical quality to produce a clear image of the projectile penetrating at high speed into the cage.
Within the framework of a play activity practised in a room, the system of detection and of qualification of events of play needs to be very dynamic and needs to propose a broad pallet of contextual animations so as to keep the show. The time separating the detection of an event to its interpretation and the triggering of a suitable animation cannot exceed the second. In addition the device shall be used in an intensive manner and be implemented and maintained by a non-specialized staff, so that its commercial exploitation be profitable.
The invention aims at proposing a robust device associated with a simple and fast signal processing, capable of detecting and to qualifying in real time an action of play, i.e. without perceptible latency for an observer, while using not specialized hardware, in particular for the vision. More precisely, the invention relates to a device for the practice of an activity using a projectile, said device comprising:

- a closed space lit by a source of artificial light comprising a cage whose plan of entrance is delimited by a framework bearing a net;
- a projectile of a given color shade;
- sensors comprising a vision device capable of detecting the color shade of the projectile in a volume comprising the cage;
- the colors, of all the elements likely to be in the scene seen by the vision device, when they are lit by the artificial light source, being selected so that none of the mixtures of the aforesaid colors cannot comprise the color shade of the projectile.

Thus the device according to the invention, detects the presence of the projectile in the cage by the simple presence of its color shade in the zone of visibility of the vision device.
The color shade is defined as a nuance of color. In colorimetric representation a color is defined by the polar co-ordinates of the point corresponding to the aforementioned color on the CIELAB disc. The color shade is quantified by the value of the angle of these polar co-ordinates, the value of the radius of said polar co-ordinates quantifies the color shade saturation. White and black are not color shades.
The detection of the color shade and not of the color makes it possible to detect the presence of the projectile even if the image is fuzzy and when the image of the projectile is limited to a trail.
This device can be implemented according to the advantageous embodiments exposed hereafter, that can be considered individually or according to any technically effective combination.
Advantageously, the color shade of the projectile when lit by the source of artificial light is included in an opposite quadrant of the CIELAB disc compared to the color shade of the other elements delimiting and contained in the closed space. This particular configuration makes it possible to simplify considerably the image processing needed for the detection of the projectile in the scene seen by the vision device, and thus to reduce the time separating the occurence of the event from and its detection, while using a video equipment of average range, cheaper and of simpler maintenance.
According to a preferred embodiment, the saturation of the color shade of the projectile is located in a crown between the outmost part and 80% of the CIELAB disc radius. Thus, the saturation of the color shade of the projectile makes its detection weakly affected by the variations in lighting luminosity.
Advantageously, the ground of the cage, its framework and the net are selected of white or black color, so that the superposition of the color of the projectile with the color of these elements only modifies the saturation of the color shade of the projectile seen by the vision device.
In a particularly advantageous embodiment, the ground of the cage is in white and black checkerworks. Thus the checkerwork provides a reference to these extremes tones in the video image
Advantageously, the optical center of the vision device is contained in a plan parallel to the plan of entrance and is shifted inside the cage from this one by a distance equal or higher to the radius of the projectile used for shooting. Thus the detection of the presence of the color shade of the projectile in a zone of the image displayed by the vision device, attests in a simple way of the crossing of the plan of entrance of the cage by the projectile. By placing the optical center of the vision device in a determined plan, this area is easily delimited by a straight line whatever are the optical distortions of these means.
The device according to the invention can be associated to complementary sensors particularly making it possible to detect other actions of play than the goals.
Advantageously, the device comprises a light barrier capable of detecting the crossing by the projectile of the entrance plan of the cage. Thus the detection of the crossing by the projectile of the entrance plan of the cage is made more reliable because it relies on information coming out of two sensors: the light barrier and the vision device. In addition, the position of the light barrier makes it possible to locate the entering of the projectile into the cage.
In order to further improve the detection reliability of the penetration of the projectile into the cage, the device advantageously comprises a tension sensor of the net, allowing the detection the impacts of the projectile on said net.
Advantageously also, the device comprises at least two impact sensors on the framework of the cage, remote from one to the other. These sensors make it possible to detect impacts of the projectile on the framework, and to locate these impacts thanks to the time analysis of the signals on the at least two sensors.
The device can comprise radio frequency identification means and means for detecting and reading of the aforesaid identification means. According to this embodiment, these identification means, commonly called RFID, are placed via bracelets on the players evolving in the device and the detection means make it possible to recognize the presence of a player in a given area in the space of the device.
Finally, the device comprises means for audio-visual broadcast, the determination of the diffused contents being a function of the information provided by the sensors. Thus it is possible to recreate, in an automatic manner, an illusion of opening of such a blind and confined space, without interfering with the capability and the reliability of the events of play detection. The automation of such a device makes it possible to exploit commercially in the same place and in an economic way, several spaces of this type without requiring many sportcaster.
The invention also relates to a method implementing such a device, said method comprising a step consisting in detecting the presence of the color shade of the projectile in the scene seen by the vision device. This method makes it possible to detect the crossing of the goal line in a reliable manner, using a simple color camera, without calling for means of video of a high cost such as high speed cameras.
The invention also relates to a method for the implementation of the device according to the invention in an embodiment comprising sensors of impact placed on the framework of the cage, said method comprising the steps of:

- defining a threshold level on the signals coming from the sensors of impact;
- assessing the position of the impact on the framework of the cage by the lag separating the instant of crossing in rising edge of this threshold on the signals coming from the sensors of impact.

Finally, the invention also relates to a method for the broadcast of audio-visual contents in the space according to the invention in its embodiment comprising means for audio-visual broadcasting, said method comprising the steps of:

- obtaining a data file, designated as the “sequence file”, comprising pre-recorded audio-visual sequences and associated with triggering keys;
- obtaining a data file, designated as the “state” file, putting in correspondence the aforementioned keys with patterns of combined state of the sensors;
- triggering a stopwatch counting the total time spent since a defined event of play;
- scanning the state of the sensors;
- triggering the broadcast of an audio-visual sequence, selected in the “sequence file”, when the combination of the keys corresponds to an identified pattern of combination.

One thus has a method allowing to quickly associating an animation and a context of play through a simple identification in a preset table.
Advantageously, this method further comprises the steps of:

- obtaining a “history” data file associating a key to an event of play, designated as “current event”, and to another event of play preceding the said “current event”;
- storing for given duration the events of play detected by scanning, and corresponding to an identified pattern in the “state” data file;
- triggering the broadcast of an audio-visual sequence selected in the “history” data file, when the combination of the state of the scanned sensors with the stored event of play corresponds to an identified combination in the “history” data file. Thus it is possible to reinforce the contextual character of the animations and to increase the illusion of spontaneousness of atmosphere for the players while preserving an entirely automated triggering and selection of the broadcasted sequences.

This method is advantageously implemented for all or part of its steps by a computer, which makes it possible to easily vary the contexts of release of animations, by a simple modification of the data files, in order to adapt them to the nature of the practised sport, to the language of broadcasting, to the age of the players, to the season or to the context of topicality etc. . . .
To this end the invention also concerns a medium readable by a computer, possibly totally or partially removable, including CDROM, magnetic media such as a hard disk or a floppy disk, a dongle with a memory, or a transmissible medium, such as an electric or optical signal, characterized in that it includes the instructions of a computer program, for the implementation of any of the steps of the method for audio-visual broadcast according to any of its embodiments when this program is loaded and run by a computer system.
The invention will be now more precisely described within the framework of its preferred embodiments, by no means restrictive, represented on FIGS. 1 to 12, in which:
FIG. 1 represents an overall front view of a goal cage according to the invention, showing an example of installation of the various sensors;
FIG. 2 shows this same goal cage, seen in profile;
FIG. 3 is an overall view in perspective of the playing space;
FIG. 4 represents the scene seen by the video camera placed above one of the goals cages;
FIG. 5 represents the image recorded by the camera during the penetration of the projectile into the cage;
FIG. 6 is a diagrammatic representation of the CIELAB disc;
FIG. 7 represents the temporal evolution of the signals coming from the sensors of impact as well as the concomitant evolution of the state of the logical tag associated to each of these signals;
FIG. 8 is an example of flow chart representing the data processing performed in order to arrive to the diffusion of a suitable audio-visual sequence;
FIG. 9 is an example of an excerpt of the “state” data file relating to the “qualif” key;
FIG. 10 shows the evolution of the “score key” according to the progression of the score of the game;
FIG. 11 is an excerpt of the data file relating to the “time key”;
FIG. 12 is an excerpt example of the data file relating to the “rythm key”.
FIG. 1, according to an embodiment suited for the practice of soccer in a room, the device according to the invention is built around a cage comprising an entrance plan (1) delimited by a framework (10). The device combines sensing means consisting in sensors of impact (21 to 24) attached to the framework (10), infra-red barriers (31 to 35) and a vision device (40) in the form of a video camera. The sensors of impact can be of any type known from the man skilled in the art, made up, for example, of piezoelectric accelerometers, or by strain gauges of the piezoelectric or piezoresistive types, glued on the framework. As an example, piezoelectric accelerometers sold under the denomination of shock sensors by the MURATA® company are well suited to this use.
These sensors are associated, if needed, with a signal conditioner adapted to their technology, for example, with a charge amplifier for the piezoelectric sensors, making it possible to obtain an analogical or numerical output signal proportional to the intensity of the load on the sensor.
The detection of an event, like an impact on the framework (10), is done by the level of this output signal going beyond of a given threshold. Each measurement channel corresponding to each sensor is thus associated with, at least, one logical tag whose state is changed when the conditioned signal emitted by the sensor corresponding to the aforementioned measurement channel exceeds a definite threshold. For instance, the logical tag is set to 0 when the level of the signal is lower than this threshold, and its level passes to level 1 when the level of the signal exceeds this threshold. One associates advantageously to the level 1 of this example, a time of remanence, Tp, predetermined, so that the logical tag, triggered by the crossing of threshold, remains on level 1 during this Tp time before returning to the 0 state.
The infra-red barriers are made of an infra-red diode and a receiver, such as a phototransistor, letting pass the electrical current when it is lit by this type of light. This technology presents a very short reaction time and is only sensitive to a narrow light spectrum, invisible at eye Typically one uses infrared diodes, emitting according to a wavelength of 940 nm under a current of 100 mA. This technology is widely used thus making these components very easy to acquire and to maintain. Each infra-red barrier manages only one quite precise zone, these barriers being not interdependent. The information delivered by each barrier is binary, the state 0 corresponding to a “not crossed sate”, and state 1 to the “was crossed state”. A time of remanence, Ti, is associated to the 1 state, so that if a barrier is crossed, it continues to deliver an information corresponding to the 1 state during a time Ti before returning to the 0 state.
FIG. 2, the device, comprises moreover a net (11) prolonging the framework and capable to damp out and to retain the projectile inside the cage when this one crossed the framework, said net being connected to a sensor (51) which measures its tension.
As an option it is possible to surround the perimeter of the framework by other nets, these nets can also be connected to a device for detecting a motion. Thus, it is possible to detect shots passing close to the cage but without touching the posts, but also to prevent the projectile appearing in the field of vision of the camera (40) without actually penetrating into the cage.
FIG. 3, in order to practise the concerned activity, for instance room soccer, two detection devices (L and V) of this type, are placed at the two ends of a space of activity (60), representing a soccer pitch. Said space of activity is a blind space and is lit only by artificial light sources (63). This space is advantageously equipped with audio-visual broadcasting means, such as loud speakers (61), screens or video projection devices (62).
The device comprises a projectile (100) suited for the practice of the concerned sport, said projectile being of a unique color shade in this space.
According to a particular embodiment, devices (110) of detection and reading of “RFID chips” or “RFID tags” are placed around the space (60) or in the ground of this one. These means make it possible to detect the position of players moving in the space of play (60), the aforementioned players being then equipped with such RFID tags placed in bracelets, shoes or any other element of their clothes.
The audio-visual animations diffused by these means make it possible to create an atmosphere which varies according to the game.
These animations broadcasted by the audio-visual means are triggered according to a temporal context and to an event-driven context. The event-driven context corresponds to events of play, primarily shots on target, in direction of the L and V cages, these events of play being detected and qualified via the sensors laid out in particular in the vicinity of these cages. The events of play are classified in two families:

goals: corresponding to the crossing of the entrance plan (1) of the cage (10) by the ball
other events

The first task of the process of audio-visual broadcasting will be thus to detect a significant event of play and to classify it in one or the other of these families.
This first classification relies upon the information collected by the vision device (40) and the tension sensor of the net (51).
The tension sensor of the net is, for example, a piezoelectric or piezoresistive (FSR) force sensor or a motion sensor placed between the net and a tensioner of this net. When a projectile penetrates into the cage and is stopped by the net, the load on the tensioners is modified compared to its average value. If this modification, detected by the tension sensor (51), is higher than a preset rate, a logical tag changes its state and passes, for example, from 0 to 1. As previously, the 1 state is associated a time of remanence, Tf, so that the logical tag returns to the 0 state only after a Tf time when it turned to the 1 state.
The vision device (40) consists, for instance, of a video color camera, this camera acquires video frames, made up of lines, at a rate of 25 frames a second. As an example, a CCD day/night color camera of the MTV type distributed by the OPTOVISION® Company is suited for this functionnality. This is a standard hardware whose acquisition and maintenance are easy.
This camera is characterized by a cone of visibility (42) depending on the size of the video detector and on the focal length of the lens. The vertex of this cone coincides with the optical center of the camera. The camera is preferentially placed above the cage, sufficiently remote from it, so that the probability of impact between the camera and the projectile used for the play, is low.
FIG. 4, the scene observed by the camera corresponds to a window (43), generally rectangular, inside the projection of the image of the cone of visibility in its focal plan. The focal length of the lens used for the camera is selected so that the image of the cage (10) and of the net (11) be comprised in this window when the camera is placed at the required distance.
The optical center of the camera is placed in a plan (41) parallel with the entrance plan (1) but shifted from this one towards the interior of the cage of a distance at least equal to the radius of the projectile used for the play, ideally equal to the diameter. One defines in the image, a zone inside which, the presence of the projectile attests the crossing by it of the entrance plan. By placing the optical center of the camera in a plan parallel with the entrance plan, the aforementioned zone is delimited by a straight line whatever the barrel or pincushion optical distortions of the lens.
If needed, the window of observation can be defined by a digital processing of the image provided by the camera.
The video line (44) delimiting the zone of observation, divides this image into two parts, one of the two parts being located behind the entrance plan (1) of the cage.
The detection of the presence of the projectile is based on a colorimetric analysis of the image. For this purpose the color of the aforesaid projectile, for example a soccer ball (100), is selected of a specific color shade. The presence of this specific color shade in the considered zone of the image attests the presence of the projectile in this zone of the image. In the example of soccer, the speed of the ball during a direct shot is lower than 60 ms⁻¹. The depth of the cage is 2 meters. With a frequency of acquisition of 25 frames per second, if a ball penetrates into the cage and that it crosses the entrance plan, its color will necessarily influence the contents of an image: either of the image which immediately precedes the detection of the net movement, or of the image which follows the detection of the net movement, and in the majority of the cases, these two images, even if the ball rebounds and comes out of the cage after having penetrated into it. In the case of another sport, the man skilled in the art will adapt the frequency of acquisition of the camera to the rate of travel of the projectile and to the goals dimensions, knowing that the higher the speed and the shallow the goals, the higher the image rate per second.
FIG. 5, a ball penetrating at high speed in the cage will leave on the photosensitive element of the camera a roughly elliptic trail (47) mixing the color of the ball and the background color. The time exposure of each frame is selected so that the ball does not move more than once its diameter during this time. As an example, in the case of soccer, the ball is 22 cm in diameter. At 60 ms⁻¹the ball crosses this distance in 3.6 ms. Consequently a shutter speed (40) of 1/300th of second ensures that the ball will have crossed a distance lower than its diameter, whatever the shot, during the time exposure of the photosensitive element of the camera.
The flood-lighting device will be selected with enough power, particularly in the vicinity of the cages, in order to enable such a short time exposure. The lower the projectile diameter and the higher its travelling speed, the shorter the time exposure.
The presence of the ball is thus detected on the image by the presence of a zone, of sufficient size, whose color shade is sufficiently close to the color shade of the ball. In order to facilitate and to make the detection more reliable, the color shade of the ball is preferably saturated and sufficiently remote from the color shade of the other items likely to be in the scene observed by the camera.
FIG. 6, the color of an item can be quantified by its position on the CIELAB disc. This disc shows, in a quantified manner, any possible color. It is organized according to a vertical axis (402) which goes upwards from blue to yellow and a horizontal axis which goes, from left to right, from green to red. Any color (404) can be represented in this system by its polar co-ordinates by the angle (403), which defines its color shade, and by the distance from to the center (405). The higher this distance (405), the more saturated the color shade. In order to facilitate the detection it is thus preferable to increase the colorimetric contrast between the ball and the other items likely to be in the scene observed by the camera. The colorimetric contrast between two colors is higher when de euclidian distance between them on the CIELAB disc is higher. Preferably, the color of the ball is chosen in shades corresponding to an angle of 0°, 90° or 180° and the colors of the other items likely to be in the scene seen by the camera (players' shirts, shoes . . . ) are chosen in the lower part of the disc in opposition, as far as possible, from the color shade of the ball
Brightness variations influence the saturation of the colors seen by the sensor of the camera. Overexposure or underexposure of this sensor moves closer to the center of the CIELAB disc the colors seen by this item. In order to limit this influence, the color of the ball is chosen in the outmost part of the CIELAB disc, typically in a crown (410) corresponding to the 20% outmost part of the CIELAB disc. From a practical point of view, the colors of the ground of the cage, of the framework and of the nets are such that any possible mix of these colors with the projectile color (404) is filling a surface (406) of the CIELAB disc. The color chosen for the projectile and the other colors shall be such that the mixing together of the other colors, with the ground, with the nets and with the framework of the cage never fills a surface of the CIELAB disc having an intersection with the surface (406) of this disc corresponding to mixes of the projectile color with the ground, the framework or the nets or the cages taken alone or in combination.
Advantageously, the ground of the cage, the color of framework and the net are selected in white or black, so that the superposition of the color of the projectile with the color of these items only modifies its saturation without modifying its color shade. Still more advantageously, they can be with white and black checkerworks and thus providing a reference to these extremes tone for the video image.
As the items in the scene visualized by the camera are reflectors of the ambient light, it is appropriate that this light is determined and appreciably constant in intensity as well as in temperature of color. This is the reason why, the space of play (60) is necessarily lit only by artificial light sources of known characteristics. Although a slight variation of the lighting intensity might be acceptable because of the selectivity of the shades, the lighting device shall be selected so that the light is never lacking, even in an unperceivable manner with the eye, nor it undergoes substantial modification of the quality during the acquisition of an image by the camera, for example it can be carried out by three-phase neon lights. Of course, the automatic white balance of the camera shall be off.
The state pattern of all the sensors is permanently scanned. A stopwatch is triggered at the beginning of the game, so as all of the signal or state acquisitions can be dated according to this stopwatch constituting a common time base for the whole of devices. The processing of an event of the “goal family” is triggered by the detection of the color shade of the projectile in the zone watched by the camera. The change of state of this tag leads to the starting of the image processing. This processing consists in filtering this image, more specifically in an erosion processing and a labelling processing on pixel groups corresponding to the color shade of the ball, and, starting from this processed image, to confirm or not the detection of the ball into the cage. Alternately, the same processing can be started in case of the detection of a motion of the net of the cage. If the detection is confirmed, the event of play is qualified as a goal, and the type of the audio-visual broadcasted sequence is selected in the “sequence file” corresponding to this kind of event. In order to select the right sequence in this file, other conditions are analyzed, in particular the state of the infra-red barriers. The state of these barriers enables to define where the ball has penetrated into the goal, for instance:

in the top corner of the net (31,32)
at ground level (34)
under the cross bar (33)

This information is then combined with the event having previously led to the audio-visual broadcasted sequence as well as with the time which has elapsed since this last event and the time lapsed from the beginning of the game in order to select a pre-recorded sequence and to broadcast it.
For instance, if the audio-visual broadcasted sequence is a comment:

a first event can be the crossing by the ball of the entrance plan of the V cage in the first minute of the game:
the tag associated with the signal coming out of the net sensor (51) of the V cage turns to 1
the camera (40) detects the ball in the bottom area (45) of the cage: the goal is taken for certain
scanning the infra-red barriers indicates that the barrier in the top left corner (31) has been crossed.

The triggered comment can be for instance:

“the L team opens the score from the very beginning of the game by a shot in the top left corner”

If within 2 minutes, an event like a goal is detected in the L cage, the comment can be, for instance:

“the V team counter-attacks and equalizes immediately by a shot at the ground level”

The comment section like “the L team” or “the V team” are pre-recorded sequences, in the same way as “a shot in the top corner” or “the score opens”. These pre-recorded sequences are subjected to a semantic combination according to the temporal and event-based context. This temporal and event-based context is described by keys depending on the state of the sensors. Thus, the structure of the comments corresponding to the “goal family” is initially labellized by a “goal key” enabling to classify this event in this family and comprises:

a subject: “the team L”, “the team V” determined by a “cage key” corresponding to the cage in which the goal was detected
a verbal group: “the score opens”, “counter-attacks while equalizing”, “equalizes” etc. . . . the choice of which is determined by a “score key”, a “time key” defining the time lapsed since the beginning of the game, a “rythm key” determining the time lapsed since the last event like the detection of a goal, and a “cage key” related to this last event and corresponding to the nature of this event (goal of the same team or of the opposing team)
a possible adverbial phrase determined by the “rythm key” when the time separating two events is particularly short or particularly long and/or a “time key” when it takes place at the beginning or at the end of the game
an adverbial phrase determined by a “shot key” qualifying the nature of the shot.
eventually, an idiom randomly selected in a group of expressions that are not describing the facts but that are supposed to account for the atmosphere such as: “what an emotion”, “spectacular” etc. . . . in order to vary the messages and to make a lively comment.

Alternately, according to a simpler embodiment though less rich in terms of comments, complete sentences can be recorded and associated to keys combinations or with state of the sensors.
In some shooting circumstances, the projectile can penetrate into the vision space of the camera without resulting in a goal. It is the case when the projectile comes to hit one of the side nets by the outside or when it rests on the exterior of the upper net. These confusing cases can be discriminated by an image processing by analyzing the dimension of the projectile in the image and/or its position It is however more effective to protect the cage by a second net preventing the projectile to reaching these areas. However the shots corresponding to this kind of situation being indeed remarkable shots, these additional nets can be equipped with motion sensors used for the qualification of the shot.
The broadcast of the comment can be accompanied by the broadcast of video images on screens or by projection (62) in the space of play (60). These video sequences can be images of a public or even the retransmission in video, possibly in idle, of the action that just occurred, and which was recorded by an additional camera (70) whose sequence was readjusted in time thanks to the time base given by the stopwatch.
The events that do not belong to the “goal family” but which constitute significant events of play are the targeted shots, and in particular the shots impacting the framework delimiting the cage.
These impacts of the ball on the goal posts are detected and qualified by the sensors of impact (21-24). Accordingly, the signals coming out of these sensors are connected to an acquisition card comprising at least as many synchronized acquistion channels as the number of sensors. The signals got from this card are stored in a buffer corresponding to a given acquisition duration. When the level of a signal coming from one of the sensors exceeds a given threshold, the signals stored in the buffer are analyzed in order to detect a possible crossing of the threshold on one of the other sensors, and the time position of the rising edge related to this threshold crossing. The time lag between the rising edges of the threshold crossing of the signals coming from the different sensors, is used for assessing the rough position of the impact on the framework (10) of the cage.
For instance, on FIG. 7, a threshold crossing is firstly observed on the signal (221) corresponding to the sensor (21) of impact located at the bottom left of the framework of the cage. This threshold crossing leads to a change in the state of the corresponding logical tag (2210) turning from 0 to 1 for a Tp duration. The shockwave propagates in the framework and reaches the sensor (22) located on the left side of the cross bar, leading to a threshold crossing of the associated signal (222) after a T lag, as well as the change of the state of the corresponding logical tag (2220). The shaking propagates in the cross-bar and reaches the sensor (23) located on the right side of this bar. The corresponding signal (223) exhibits a disturbance corresponding to the shaking, but which does not exceed the threshold level, so that the state of the logical tag (2230) corresponding to this signal does not change. A “shot key” that locates the shot on the left post is deduced from the state of the logical tags and from the temporal sequence of their change of state. In a situation where the logical tag (2220) corresponding to the left sensor of the cross-bar (22) would have changed of state before the one (2210) corresponding to the sensor into bottom of the left post (21), all the other conditions being kept, one would have associated to this event a “shot key” that locates the shot in the top left corner.
According to an embodiment seen on FIG. 8, the method of detection and broadcasting is implemented by means of computer programs run by three computers connected in a network. Two computers, known as “capturing computers”, are affected each one with a goal cage. The computer program they are implementing, permanently scans the state of the sensors of their respective cage.
The first step (1000) of the method, implemented by the “capturing computers”, comprises the detection of any event occurring at the level of the monitored cage. An event consists in the change of state of a logical tag assigned to one of the sensors of impact (21 to 24), of motion of the net (51) or the crossing of any light barrier (31 to 36), or the detection of the color shade of the projectile in the area monitored by the camera. As an example, the computer dedicated to the A cage detects the change from 0 to 1 of the logical tag corresponding to the crossing of the light barrier (31) located in the left corner of the cage.
The detection of this event triggers a step (1100) of examination of the state of the other logical tags. At the end of this step, the state of the logical tags corresponding to the detected event is known and used for the performance of the qualification step (1200). This qualification step (1200) triggers, if needed, subroutines of scanning and signal treatment (1210) in particular of the video signal, if one of the logical tags corresponding to the net motion or to the presence of the projectile was triggered.
Such a subroutine can also aim at locating the impact of the projectile having struck the posts starting from the examination of the acquisition buffer of the signals coming out of the sensors of impact. The whole of the state of tags and of the sensors is thus analyzed making it possible to identify the state of the “shot keys” (603), “goal key” (601) and “qualif key” (602), by comparison of the state of the logical tags or of the sensors with the cases indexed in the corresponding “state” data files (503, 502).
FIG. 9 is an excerpt example of the data file “state” relating to the light barriers (503) making it possible to determine the state of the “qualif key” (602). The first 6 columns (C1 with C6) contain the potential states of the 6 logical tags corresponding to the 6 infra-red barriers (31 to 36) that are installed on the cage (10), in the form of binary logical variables. The seventh column (C7) comprises the value of the “qualif key” corresponding to the combinations of the logical variables according to the lines of the table. The value of the key is coded according to a suitable format which can be numerical or alphanumeric. This table (503) sets up a hierarchy in the values of key according to the practised sport or according to the actual difficulty of the shot. As an example, the first line of the table (L1) corresponds to the crossing, during the time of remanence, Ti, of the light barrier (31) located in the top left corner, of that located along the left post (35) and of the one located under the cross bar (33). There is no doubt in this case that the ball has entered by the top left corner and the“qualif key” is coded to qualify the shot as “top left corner shot”. The second line (L2) of the table corresponds to the crossing of the light barrier in the top left corner (31) and to that monitoring the zone under the cross bar (33). In this situation, the shot can be described as “under the cross bar” or “on top corner”. According to this example it was deliberately selected to qualify it as “top left corner shot” since it is considered that a comment associated with this situation is more prestigious for the player.
At the end of the qualification step, the keys are combined in the form of a suitable file and dispatched via the computer network towards the processing computer while associating it with a “cage key” that determines the goal (V, L) in which this combination of keys was detected.
A similar treatment scheme is performed for the other goal.
The program stored in the processing computer will combine the event-driven context, given by the state of the keys sent from the “capturing computers”, with the temporal context and will start the broadcast of audio-visual animations according to these two contexts.
A first step of analysis of the received keys (3000) will, on the one hand, store the event in the form of its keys combination on the top of a stack of a buffer file of events (3010), and on the other hand, will perform other conditioning processes, and in particular, will check the consistency of the information received from the two “capturing computers”, for instance, checking that the two sensor devices did not detect a goal at the same time. The presence of a “goal key” indicating that a goal was scored, will trigger a particular processing relating to the score which will fill a “score key” (604).
FIG. 10 is a processing example of the “score key” (604) whose states (A, B, C, D, E, F, G) are determined by the way the score is moving.
A stopwatch (3500) is started at the beginning of the game and makes it possible to dating the whole of events. The time determined by the stopwatch is shared between the computers via the network as well as a time-sharing protocol so as define a common date basis.
A step of time processing (3510) combines the time information of stopwatch with the state of the keys coming from the “capturing computers”, from the “score key” (604), from the stack of events (3010), by unstacking the latter from bottom (method known as FIFO) and generates a “rythm key ” (605) and a “time key” (606) by identification in the data tables “history” (505) and “time” (504) .
FIG. 11 gives an excerpt example of the data file “state” relating to the “time key”. The initially defined duration of the game is split in 4periods corresponding to values of the “time key”. FIG. 12 is an example of the data file “history” (505) relating to the “rythm key” (605).
A step of selection of the sequence (3200) combines the values of the whole of the keys so as to select an audio-visual broadcasted sequence by identification of the value of the keys in a data file “sequences” (700) and with the associated preset sequences. Alternately the broadcasted sequences can be assembled by a semantic module according to the value of the keys. The audio-visual sequence being selected, it is broadcasted (4000) using suitable audio-visual means (61,62).
The foregoing description illustrates clearly that by its various characteristics and their advantages, the present invention achieves the goals that were laid down. In particular, it makes it possible to detect and to qualify an event of play, then to trigger in an automatic manner the broadcasting of an audio-visual sequence adapted to the temporal and event-driven context of the action of play.

Claims

1. A device for the practice of an activity using a projectile, wherein said device comprises:

a closed space lit by a source of artificial light, said space comprising a cage whose plan of entrance is delimited by a framework bearing a net;

a projectile of a given color shade;

sensors comprising a vision device capable of detecting the color shade of the projectile in a volume comprising the cage;

the colors of all the elements likely to be in the scene seen by the vision device, when they are lit by the artificial light source, being selected so that none of the mixtures of the aforesaid colors cannot comprise the color shade of the projectile.

2. The device according to claim 1, wherein the color shade of the projectile, when lit by the source of artificial light, is included in an opposite quadrant of the CIELAB disc compared to the color shade of the other elements delimiting and contained in the closed space.

3. The device according to claim 1 wherein the saturation of the color shade of the projectile is located in a crown ranging between the outmost part and 80% of the radius of the CIELAB disc.

4. The device according to claim 1 wherein the ground of the cage, its framework and the net are selected of white or black color, so that the superposition of the color shade of the projectile with the color of these elements only modifies the saturation of the color shade of the projectile seen by the vision device.

5. The device according to claim 1 wherein the ground of the cage is with white and black checkerworks.

6. The device according to claim 1 wherein the optical center of the vision device is contained in a plan parallel with the entrance plan and is shifted from this one inside the cage of a distance equal or higher than the radius of the projectile.

7. The device according to claim 1, wherein a light barrier is capable to detect the crossing by the projectile of the entrance plan of the cage

8. The device according to claim 1, wherein a sensor measures the tension of the net of the cage.

9. The device according to claim 1, comprising at least two sensors of impact remote from one of the other on the framework of the cage.

10. The device according to claim 1, comprising identification means by radio frequency and detection means for reading of the aforesaid identification means.

11. The device according to claim 1, comprising audio-visual broadcasting means, the determination of the broadcasted content being a function of the information provided by the sensors.

12. A method for the detection of the crossing by the projectile of the entrance plan of the cage of the device according to claim 1, comprising a step consisting in detecting the presence of the color shade of the projectile in the scene seen by the vision device.

13. A method for the localization of an impact on the framework of a cage of the device according to claim 9, comprising the steps of:

defining a threshold level on the signals coming from the sensors of impact;

assessing the position of the impact on the framework of the cage by the lag separating the instant of crossing in rising edge of this threshold on the signals coming from the sensors of impact.

14. A method for broadcasting an audio-visual content by the device according to claim 11, comprising the steps of:

obtaining a data file, designated as the sequence file, comprising pre-recorded audio-visual sequences, associated with triggering keys;

obtaining a data file, designated as the “state file”, putting into a correspondence the aforementioned keys with patterns of combined state of the sensors;

triggering a stopwatch to counting the total time spent since a defined event of play;

scanning the state of the sensors;

triggering the broadcasting of an audio-visual sequence, selected in the “sequence file”, when the combination of the keys corresponds to a identified pattern of combination.

15. The method according to claim 14, further comprising the steps of:

-obtaining a data file “history”, associating a key to an event of play, designated as “current event”, and to another event of play, preceding this “current event”;

storing during a given time the events of play detected by scanning and corresponding to an identified pattern in the “state data file”;

triggering the broadcast of an audio-visual sequence selected in the “history” data file when the combination of the state of the scanned sensors with the stored event of play corresponds to an identified combination in the “history” data file.