CN106326934A - Dice throw-out point number judging method and device - Google Patents
Dice throw-out point number judging method and device Download PDFInfo
- Publication number
- CN106326934A CN106326934A CN201610741292.3A CN201610741292A CN106326934A CN 106326934 A CN106326934 A CN 106326934A CN 201610741292 A CN201610741292 A CN 201610741292A CN 106326934 A CN106326934 A CN 106326934A
- Authority
- CN
- China
- Prior art keywords
- points
- image
- point
- dice
- processing unit
- 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; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/74—Image or video pattern matching; Proximity measures in feature spaces
- G06V10/75—Organisation of the matching processes, e.g. simultaneous or sequential comparisons of image or video features; Coarse-fine approaches, e.g. multi-scale approaches; using context analysis; Selection of dictionaries
- G06V10/751—Comparing pixel values or logical combinations thereof, or feature values having positional relevance, e.g. template matching
-
- A—HUMAN NECESSITIES
- A63—SPORTS; GAMES; AMUSEMENTS
- A63F—CARD, BOARD, OR ROULETTE GAMES; INDOOR GAMES USING SMALL MOVING PLAYING BODIES; VIDEO GAMES; GAMES NOT OTHERWISE PROVIDED FOR
- A63F9/00—Games not otherwise provided for
- A63F9/04—Dice; Dice-boxes; Mechanical dice-throwing devices
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/44—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- Multimedia (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Health & Medical Sciences (AREA)
- Computing Systems (AREA)
- Databases & Information Systems (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Software Systems (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a dice point number judging method for a dice game device. The method comprises the steps of construction of a training library, storage of a training library image processing mode, shooting, edge detection and feature extraction, comparison with a training library image and result output, wherein the construction step of the training library comprises the steps of shooting multiple six-side images of a dice on each side from different angles, then sending the images to an image processing unit for processing, and then comparing the six-side images of the dice to find out points of each side of images not superposed with the points of other sides of images under the requirement that the sizes of the points can be stably distinguished from the other points; in the absence of points completely not superposed with the other point images, the points not superposed with part of point images can be selected as feature points; overall judgment can be performed by adopting a combination of multiple points; the combination of multiple points for completely distinguishing six sides can be found by a machine or artificially, and the images can be cut by using part of small images convenient to cut.
Description
Technical field
The present invention relates to dice throw out count determination methods, throw out count judgment means, use it game device and
Dice is thrown out judgement system of counting, particularly to the dice accurately judged by image recognition throw out count determination methods, throw
Go out judgment means of counting, the game device using it and dice to throw out judgement system of counting.
Background technology
In dice game, Random Factor Mahjong, dice is necessary props.Also have in other are played, such as some families
In front yard electronic game machine, Trivial Pursuit Unhinged etc., dice is also conventional instrument.In the past, throwing out when shaking the elbows was counted judgement, all
It is visually carrying out by game participant.Due to size of desktop, dicebox size and the impact of environmental factors, cause mesh
Depending on highly inconvenient.
Improved plan is proposed, and the dice judged of counting thrown out the most electronically carrying out dice is thrown out and is counted
Identify device.In this device, the inside of each of " 1 "～" 6 " of dice is all built-in with Magnetitum, is detected from this by Magnetic Sensor
The magnetic that Magnetitum sends, and the detection signal of this scanning is processed with control and treatment circuit, judge to throw out with this and count.But,
According to the judgment means of counting of throwing out of existing dice, owing to the special dice of built-in Magnetitum sheet must be used, therefore can not
Use the dice sold in toyshop etc..
Somebody proposes a kind of pattern recognition device, and it, by arranging photographic head below dicebox, passes through processor
Control and process and identify the counting of dice, it uses general recognizer, its general flow for taking pictures, image locates in advance
Reason, feature detection and extraction and training storehouse image comparison, crucial point result.But, there is identification and calculate in current pattern recognition device
Method does not considers the particularity of die structure, causes recognizer to be simplified not, and training storehouse quantity is big, and efficiency is not high enough, and cost
High.
Summary of the invention
Therefore, the present invention is directed to the problems referred to above, it is provided that one can improve recognition effect and effect effectively reducing cost simultaneously
The apparatus and method of rate.Further, the present invention provides a kind of and adapts to the identification apparatus and method without model size dice.
It is an object of the invention to provide the method and apparatus identifying dice of a kind of optimization, further, be provided that
Voice broadcast.
The present invention, in order to reach above-mentioned purpose, uses following scheme.
A kind of dice game device and dice determination methods, this game device includes dice, box 100, photographic head
201, graphics processing unit 202, processor 300, memory element 400, output unit 500；Wherein box 100 includes box body 101,
Lid 102, is provided with lamella lucida in box body, output unit 500 is speaker and/or display device；
It is characterized in that, it is judged that time its perform following steps,
Training storehouse builds: shoot the six-face image of dice by being arranged on the photographic head 201 bottom box body 101, to every
Shoot multiple the most from different perspectives, be then fed into graphics processing unit and process, wherein six-face image is carried out preliminary
The process abated the noise, subsequently by hexahedro for dice figure comparison, finds out the counting not of other face images of counting of each image
There is the point that position overlaps, it is desirable to this spot size can stably realize the differentiation put with other；When not existing and other point-and-figure charts
During the most misaligned point of shape, can select to count the misaligned point of image as characteristic point with part；And can use many
The mode of individual combination carries out overall judgement；By machine or the group that manually finds the multiple points 6 faces can distinguished completely
Close, and little with area and cut convenient partial graphical image is cut, it is stored in training storehouse after extracting characteristics of image；
Training storehouse image procossing mode stores: after selection determines the cutting mode of training storehouse image, will by processor 300
This kind of selection mode is sent to graphics processing unit 202, count in follow-up game identification time, graphics processing unit 202 is right
During the process of image, extract the view data of dice in game in this manner, to facilitate and training storehouse comparison.
Taking pictures: after starting game, dice rotates a period of time stopping in box 100, now by being positioned at cartridge bottom
Dice is shot by photographic head 201 across transparent material；
Feature detection and extraction: after taking pictures, by image transmitting to graphics processing unit 202, graphics processing unit is first
Carry out certain pretreatment, then carry out limb recognition, wherein, by by the color gray value of dice body and dice box
Color gray value formed powerful contrast so that during image recognition, dice edge and background have obvious boundary, with 256 grades of ashes
As a example by degree, gray value differences away from for 50-255, preferred scheme be gray scale difference be 150-255, if dice body is white, and dice
Sub-box inwall is Dark grey or black, or in turn；Subsequently image is positioned, and according to the cutting mode stored before
Carry out cutting image and extract characteristics of image；
Image comparison, result output:
Graphics processing unit 202 will extract image feature data send into processor 300, with preprocessor 300 by its with
The view data in training storehouse is compared, and finds out the point-and-figure chart picture of coupling, and result is sent to by image and/or voice mode
Player.
The further scheme of the present invention is, the hexahedro some size of dice is identical, and arbitrary angle point distance center of 2 .-6
Position identical, now, above-mentioned training storehouse structure concrete analysis step be: 2, the centre of 6 substantially without with other faces
Point exist overlap, can be as characteristic point；1 point, 3 points, 5 dot center's points overlap with 2 points, 4 points, 6 can distinguish, it is contemplated that
As characteristic point；The position of four angle points is symmetrical, independent one can only other as the differentiation with 1, two are adjacent
Angle point can by 1 with 2 points, 3 with 4 points, 5 points, 6 distinguish respectively, collect above-mentioned information subsequently, analyze and understand central point
Can by 1 point, 3 points, 5 with 2 points, 4 points, 6 differentiate；1 can distinguish by two adjacent corner points with 5 with 3, and i.e. two
Individual adjacent corner points does not exists, and is 1 point, to have one be 3 points, has two is 5 points；For 2 points, 4 and the district of 6
Point, then adjacent corner points exists one is 2 points, and having two is 4 or 6 points, in conjunction with the middle any point in 2, centre of 6
Distinguish and 6 points at 4, it is seen that the central point of 1 central point, two adjacent corner points and combines can district completely at one 6
Divide 6 points, followed by one by the straight line of central point, and each the graphic feature using this cutting mode is extracted
It is stored in training storehouse.
Another scheme substituted of the present invention is that the hexahedro some size of dice is identical, and in arbitrary angle point distance of 2 .-5
The position of the heart is identical, 64 angle points relative to other faces angle point be located closer to edge, now, above-mentioned training storehouse structure
The concrete analysis step made is: any point of 6 all will not be completely superposed with the point in other faces, and possesses distinguishing condition, can
Using as characteristic point；1 point, 3 points, 5 dot center's points overlap with 2 points, 4 can distinguish, can be as characteristic point；Remaining four angle points
Position is symmetrical, independent one can only other as the differentiation with 1, two adjacent corner points can by 1 with 2 points, 3 points
With 4 points, 5 distinguish respectively, collect above-mentioned information subsequently, analyze understand central point can by 1 point, 3 points, 5 with 2 points, 4 districts
Do not open；1 can distinguish by two adjacent corner points with 5 with 3, and i.e. two adjacent corner points do not exist, and are 1 point, exist
One is 3 points, to have two be 5 points；For 2 points, the differentiation of 4, then adjacent corner points exists one is 2 points, and having two is 4
Point, can distinguish and other points, then 1 central point, 4 two adjacent corner points and 6 points at 6 in conjunction with any point of 6
A point combine and can distinguish 6 points completely, followed by two diagonal by image division, take 1/4th areas
Image, and by use this cutting mode each graphic feature extract be stored in training storehouse.
Accompanying drawing illustrates:
Fig. 1 is the hexahedro schematic diagram of dice.
Fig. 2 is the structural representation of apparatus of the present invention.
Fig. 3 dice of the present invention identification step.
Schematic diagram after the cutting of Fig. 4 first embodiment image.
Fig. 5 the second embodiment hexahedro schematic diagram of dice.
Schematic diagram after the cutting of Fig. 6 the second embodiment image.
In figure: 100-box, 101-box body；102-lid；201-photographic head, 202-graphics processing unit；300-process
Device；400-memory element；500-output unit, 501-speaker, 502 display devices.
Detailed description of the invention:
Fig. 1 shows the dice of the present invention, and general structure is hexahedron, 1-6 point is respectively provided with on one face, its
In 1 the most corresponding with 6 points, 2 and 5 points, 3 and 4, other regular polygon structures can be used the most as required.Dice
The image of sub-surface is fairly simple for head portrait, fingerprint, and dice surface does not the most have big impurity, and dice
Size of counting is relatively large for the small impurities such as dust.
Fig. 2 shows the structural representation of dice game device for the present invention, including: box 100, it is by box body
101, lid 102 forms, and is provided with a lamella lucida in box body；Photographic head 201, is arranged on the lower section of box body 101 inner transparent plate, work
Can shoot the image of dice when making from below, its quantity can be one or more；Graphics processing unit 202, is responsible for shooting
Image process；Processor 300, is responsible for the control to whole device；Memorizer 400；Output unit 500；Output unit
500 can be speaker and/or display device.
Seeing Fig. 3, the workflow of apparatus of the present invention is, training storehouse builds, training storehouse image procossing mode stores, claps
According to, feature detection and extraction and training storehouse image comparison, output result.The present invention has on the product or method of multiple steps
Certain improvement, is specifically shown in explanation below.
(1) training storehouse builds:
Fairly simple in view of dice image information, need identification in game is all fixing one group or several groups of dices,
Carrying out under dice is in quiescent conditions during identification, and the face of identification is substantially stationary with the distance angle of dice, therefore, the present invention trains
The structure in storehouse need not consider the generally recognized in the problem such as the collimation error, just shoot structure with the dice used.
Concrete mode is to be shot the six-face image of dice by photographic head 201, many to every shooting the most from different perspectives
, it is then fed into graphics processing unit and processes, wherein six-face image is carried out the preliminary process abated the noise, subsequently
Carry out feature extraction, and each multiple image is classified as a class.
Identify the graphic collection that General Principle is the different objects that will identify of image, and this classifying mode can be by difference
Image distinguishes, and for dice, the most regular six faces, image is the most significantly different, and general classifying mode is exactly every figure
As sorting out, but this mode does not take into full account the feature of dice image, causes training database data amount relatively big and follow-up
Image comparison data volume big, impact processes time and speed, in order to identify rapidly, then has processor performance higher
Requirement, it is impossible to reduce cost very well.
In the solution of the present invention, all images of every one side is not extracted in training storehouse when building, and only extracts part, this
Sample can improve efficiency, and this just requires that six faces can be accurately distinguished out by the parts of images extracted, and this mode is feasible
Property based on the hexahedro pattern of dice the most simply, the most significantly different and every one side there is again polysymmetrical outward appearance.See
Fig. 1 dice figure understands, and six face graphics distinctions are huge, finds out count (or certain part counted) and other faces of each image
There is not the point that position overlaps in the counting of image, then this count (or part) can be as the characteristic point in this face, it is naturally required that this spy
Levy and must a little be fulfilled for certain size, to be capable of the stable differentiation put with other, and dust impurity can be excluded；When not existing
During the most misaligned with other figures of the counting point, can select with the part misaligned point of image of counting as characteristic point, when
So at this time, it may be necessary to be combined judging with other the most misaligned points.Now, the selection of the misaligned point of some is then
Need some principles, such as, the existence of certain point can clearly exclude point-and-figure charts much more as far as possible as time, then as preferably.
Such as, (placed according to the mode in Fig. 1) for 6 o'clock in the middle of 2 positions of a row substantially without with other
Figure overlaps, or its figure that is longer due to its 3 point one rows and that cause a part for outermost point will not put with other
Overlap, may select for characteristic point, and its central point without point can as with 1,3,5 characteristic points distinguished；For 5 points
For, there is not the point the most misaligned with other dices, its central point can be as the feature distinguished with 2,4,6 dot patterns
Point, any two adjacent corner points can be different with accompanying drawing 2 certainly for each cake number size as the characteristic point of differentiation 1,3
Dice, may find out other different diacritical points；According to this thinking, image of respectively counting is processed, find each counting
The characteristic point of figure, the presence or absence of a point at most can as the judgement in two faces, but the composite figure of multiple point is the most permissible
Distinguish multiple.
Practical operation can find out the combination of a lot of point, and always find convenient segmentation and the point of operand can be reduced
Combination, for all of general dice, the combination of the multiple points on minimum half images can distinguish 6 faces completely, and
For special dice, the combination having the point in the image-region of less area can distinguish 6.
After finding out relevant point, it is considered to rule convenient and can cover the image cutting mode of these positions come selected as
Training storehouse view data, it is possible to be output to display device 502, user judge which type of shape, or by program
Automatically selecting according to certain principle, cut the most for convenience, reduce emotion operand, prioritizing selection passes through limit
Edge angle point, midpoint, the straight line of central point.
After having determined image cutting mode, selected view data is stored in system by system
(2) training storehouse image procossing mode recording step:
After the selection of training storehouse determines the cutting mode of image, by processor 300 by this kind of selection mode recording and sending
To graphics processing unit 202, in follow-up game, when graphics processing unit 202 is to the process of image, extract in this manner
The view data of dice in game, to facilitate and training storehouse comparison.
Certainly, processor 300 can by this kind of selection mode record but not graphics processing unit 202, but follow-up enter
By it, image is carried out this kind of cutting process during row dice image recognition.
(3) take pictures:
After starting game, dice rotates a period of time stopping in box 100, now by the photographic head being positioned at cartridge bottom
201 shoot, and dice does not directly contact with photographic head 201, but across layer of transparent body, such as glass, due to shooting
201 immediately below dice shooting and its vertical dimension with dice fix, therefore, the image of shooting does not the most deform or chi
Very little difference, relatively easy during image procossing.
(4) feature detection and extraction:
After taking pictures, by image transmitting to graphics processing unit 202, first graphics processing unit carries out certain pre-place
Reason, then carries out the steps such as limb recognition, image characteristics extraction.
1. limb recognition
When general subject image identification, the edge of target object and background image in general can not good districts
Point, this causes needs complicated algorithm to be identified by object edge, and this method comparison expends time and processor, and
When multiple dices being carried out limb recognition, this workload is bigger simultaneously.The present invention uses a kind of simple effective method to solve
This problem, will form powerful contrast by the color gray value in the color gray value of dice body and dice box so that figure
As when identifying, dice edge and background have obvious boundary, as a example by 256 grades of gray scales, gray value differences is away from for 50-255, more excellent
The scheme of choosing be gray scale difference be 150-255, if dice body be white, and dicebox inwall is Dark grey or black, or instead
Come over.Such scheme makes limb recognition very simple, can simplify algorithm accordingly, save the time.
Image border has direction and two features of amplitude, moves towards along edge, and pixel value change ratio is shallower；And be perpendicular to
Edge moves towards, then pixel value change is relatively more violent, and this step-like that acutely may present, it is also possible to present ramped shaped.And it is right
For dice image in the present invention, its edge is straight, and with background image difference substantially, say, that its limit linearly
Edge moves towards, and pixel value is almost unchanged, and is perpendicular to the trend at edge, changes stepped.
Edge detection method of the present invention two oriented operators (level, is vertical) of employing:
By horizontal operator and vertical operator, image being carried out convolution the most respectively during detection, obtain is two matrixes,
It is M1, M2 an equal amount of with original image, and they represent two partial derivatives in image M at same position respectively.Then
It is added after two numbers square of M1, M2 correspondence position and obtains new a matrix G, G and represent the gray scale of each pixel in image M
Grad (one is approached), is then processed by the threshold values of shade of gray value and obtains edge image.
Find, for the spy that contrast at the dice image border of the present invention is obvious and straight according to research and experimental result
Point, uses above-mentioned two operator simple and effective can extract edge, saves operand and time., save computing
Amount and time.
2. image characteristics extraction:
After having identified edge, image is positioned, and the process of the training storehouse image according to (two) step record above
Mode, extracts characteristics of image, such as extracts certain part of certain half or other modes, can reduce extraction data, carries
The while of high efficiency, accurately realize follow-up contrast.
(5) with training storehouse image comparison, result output
The characteristics of image extracted is sent into processor 300 by graphics processing unit 202, with preprocessor 300 by itself and training
The data in storehouse are compared, and by some conventional algorithms, search out the result of its correspondence, provide concrete counting subsequently.
Counting of multiple dices can be by image and/or voice output to player, it is of course also possible to according to game rule
Then, it is judged that the victory or defeat of each side, export in the lump to player.
Embodiment one:
Below, the working method of the present invention it is described with reference to
Start machine, gather dice view data by photographic head 201, and send into graphics processing unit 202 and be analyzed
Processing, in conjunction with accompanying drawing 3, in figure, 2, the centre of only 6 will not overlap with the point in other faces, can be as characteristic point；1 point,
3 points, 5 dot center's points overlap with 2 points, 4 points, 6 can distinguish, it is contemplated that as characteristic point；Remaining four corner location are mutual
Symmetry, independent one can only other as the differentiation with 1, now can consider with the combination of two or more angle points, two
The angle point combination at individual diagonal angle is almost without distinguishing effect, and two adjacent corner points can be by 1 and 2 points, 3 and 4 points, 5 points, 6 points
Distinguishing respectively, the combination of more angle points also can thinking analysis according to this.Collect above-mentioned information subsequently, with seeing whether they can be by 6 later
Individual face distinguishes the most completely.Central point can by 1 point, 3 points, 5 with 2 points, 4 points, 6 differentiate；1 permissible with 5 with 3
Distinguishing by two adjacent corner points, i.e. two adjacent corner points do not exist, and are 1 point, to have one be 3 points, has two is 5
Point；For 2 points, 4 and the differentiation of 6, then adjacent corner points exists one is 2 points, and having two is 4 or 6 points, in conjunction with 6
The middle any point in 2, centre of point can distinguish and 6 points at 4.These points visible can distinguish 6 points completely, then looks for one
Simple and minimum zone surround central point, two adjacent corner points and the line of 6 dot center's points.
Such as figure, shown in 4, can draw a straight line by central point, the general area in cutting dice face can carry out six
The division completely in face, is then stored in training storehouse, and the cutting side this determined by each the graphic feature using this cutting mode
Formula is recorded into graphics processing unit 202 or is stored in CPU300 and calls in graphics processing unit 202 when successive image processes.
Then beginning to game, participate in two sides of game or the most all put in box 1 by multiple dices, such as every side puts 3
Individual, dice rotates in box 100, and after the most static, photographic head 201 shoots dice bottom surface.
Image processing step, the image after collection enters processing unit 202, after carrying out the pretreatment such as denoising, according to before
The cutting mode of record carries out image segmentation and feature extraction, and characteristic image is sent into processor 300.
Process step: the data that graphics processing unit 202 is sended over by processor 300 and memory element 500 store
Training storehouse view data compare, and send the result to output unit 500 by output unit 500.At dice image
The analysis storage of training database data and game utilize this distinguishing condition to be judged that dice belongs to during dice image real time transfer
In which side player
Output step: result is sent to player with sound and/or image mode by output unit.
In conjunction with accompanying drawing 5 and 6, the present invention the second detailed description of the invention, concrete steps and the first basic phase of embodiment are described
With, from figure 5 it can be seen that its dice and the difference in Fig. 1 are that dice is different with size, specifically, wherein 6 point
Four corner location than 5 and four angle points more more outward.With reference to the analysis in embodiment one, central point can by 1 point,
3 points, 5 with 2 points, 4 points, 6 differentiate；1 can distinguish by two adjacent corner points with 5 with 3, i.e. two adjacent angular
Point does not exists, and is 1 point, to have one be 3 points, has two is 5 points；For 2 points, 4 and the differentiation of 6, then adjacent angular
It is 2 points that point exists one, and having two is 4 points；It can be distinguished by one angle point of 6 with other points.These points visible
6 points can be distinguished completely, then look for simple and the encirclement central point of minimum zone, two adjacent corner points and one 6
The line of dot center's point.
(it is respectively the situation of 1 o'clock to 6 o'clock from left to right) as shown in Figure 6, sieve can be cut two diagonal, and take
One of i.e. 1/4 area, can carry out hexahedro division completely, then be deposited by each the graphic feature using this cutting mode
Enter and train storehouse, and the cutting mode this determined is recorded into graphics processing unit 202 or is stored in CPU300 and at successive image
Graphics processing unit 202 is called in during process.
The preferred embodiment of the present invention is described above in conjunction with accompanying drawing content.But, the present invention can be with multiple different
Form is implemented, and should not be considered limited to embodiment presented herein；More precisely, these embodiments are carried
Supplying is for completeness and integrity, and conveys the scope of the invention to those skilled in the art fully.
Claims (7)
1. a dice determination methods in dice game device, this game device includes dice, box (100), photographic head
(201), graphics processing unit (202), processor (300), memory element (400), output unit (500)；Wherein box (100)
Including box body (101), lid (102), being provided with lamella lucida in box body (101), photographic head (201) is saturating in being arranged on box body (101)
The lower section of isotropic disk, output unit (500) is speaker and/or display device；
During work, shooting the image of dice during photographic head (201) work from below, subsequent picture is admitted to graphics processing unit
(202), after image is processed by it and will process after data be sent to processor (300), with preprocessor (300) by institute
State view data to compare with the training storehouse image of storage in memory element (400), and it is single that comparison result is sent to output
Unit (500).
2. a dice determination methods in dice game device, this game device includes dice, box (100), photographic head
(201), graphics processing unit (202), processor (300), memory element (400), output unit (500)；Wherein box (100)
Including box body (101), lid (102), being provided with lamella lucida in box body (101), photographic head (201) is saturating in being arranged on box body (101)
The lower section of isotropic disk, output unit (500) is speaker and/or display device；
It is characterized in that, it is judged that time its perform following steps,
Training storehouse builds: by arranging the six-face image of photographic head (201) shooting dice, many to every shooting the most from different perspectives
, it is then fed into graphics processing unit and processes, wherein six-face image is carried out the preliminary process abated the noise, subsequently
Comparing hexahedro for dice, there is not the point that position is completely superposed in the counting of other face images of counting finding out each image
As characteristic point, it is desirable to this spot size can stably realize the differentiation put with other；Complete with other figures of counting when not existing
During misaligned point, select the point misaligned with image of partly counting as characteristic point；And use multiple somes various combinations
Mode carries out overall judgement；By machine or the combination of manually finding the multiple points 6 faces can distinguished completely, and use area
Little and cut convenient partial graphical image is cut, it is stored in training storehouse after extracting characteristics of image；
Training storehouse image procossing mode stores: after selection determines the cutting mode of training storehouse image, should by processor (300)
Kind of selection mode is sent to graphics processing unit (202), count in follow-up game identification time, graphics processing unit (202)
During to the process of image, extract the view data of dice in game in this manner, to facilitate and training storehouse comparison；
Taking pictures: after starting game, dice rotates a period of time stopping in box (100), now by being positioned at taking the photograph of cartridge bottom
As dice is shot by head (201) across transparent material；
Feature detection and extraction: after taking pictures, by image transmitting to graphics processing unit (202), first graphics processing unit enters
The certain pretreatment of row, then carries out limb recognition, wherein, by by the color gray value of dice body and dice box
Color gray value forms powerful contrast so that during image recognition, dice edge and background have obvious boundary, with 256 grades of gray scales
As a example by, gray value differences away from for 50-255, preferred scheme be gray scale difference be 150-255, if dice body is white, and dice
Box inwall is Dark grey or black, or in turn；Subsequently image is positioned, and come according to the cutting mode stored before
Cutting image extracts characteristics of image；
Image comparison, result output:
The image feature data extracted is sent into processor (300), with preprocessor (300) by it by graphics processing unit (202)
Compare with the view data in training storehouse, find out the point-and-figure chart picture of coupling, and result is sent by image and/or voice mode
To player.
3. method as claimed in claim 2, wherein, die structure is: each some size in six faces is identical, and 2 .-6 points
The position of arbitrary angle point distance center identical, now, the concrete analysis step of above-mentioned training storehouse structure is: the centre two of 6
Point will not overlap with the some existence in other faces, and it is as characteristic point；1 point, 3 points, 5 dot center's points overlap with 2 points, 4 points, 6 can
To distinguish, it also serves as characteristic point；The position of four angle points is symmetrical, independent one can only other as the district with 1
Point, the combination of two adjacent corner points can by 1 with 2 points, 3 with 4 points, 5 points, 6 these three groups distinguish respectively；Collect subsequently
State information, it is known that central point can by 1 point, 3 points, 5 with 2 points, 4 points, 6 two groups differentiate；1 can be by two with 5 with 3
Individual adjacent corner points distinguishes, and i.e. two adjacent corner points do not exist, and is 1 point, to have one be 3 points, has two is 5 points；Right
In 2 points, 4 and the differentiation of 6, then adjacent corner points exists one is 2 points, and having two is 4 or 6 points, in conjunction with in 6
Between 2 middle any points can distinguish and 6 points at 4；
By analyzing, the intermediate point of 1 central point, two adjacent corner points and 6 combines and can distinguish 6 completely
The point in individual face, followed by one by the straight line of central point, and extracts each the graphic feature using this cutting mode
It is stored in training storehouse.
4. method as claimed in claim 2, wherein, die structure is: the size of each point in six faces is identical, and 2 .-5
The position of arbitrary angle point distance center of point is identical, 64 angle points relative to other faces angle point be located closer to edge,
Now, the concrete analysis step of above-mentioned training storehouse structure is: any point of 6 all will not be completely superposed with the point in other faces,
It is provided with distinguishing condition, could act as characteristic point；1 point, 3 points, 5 dot center's points overlap and with 2 points, 4 can distinguish, can be spy
Levy a little；The position of four angle points is symmetrical, and two adjacent corner points can be by 1 and 2 points, 3 and 4 points, 5 these three groups difference districts
Separately；Collect above-mentioned information subsequently, it is known that central point can by 1 point, 3 points, 5 with 2 points, 4 differentiate；Two adjacent corner points energy
Distinguishing 1 with 5 with 3, i.e. two adjacent corner points do not exist, and are 1 point, to have one be 3 points, has two is 5
Point；For 2 points, the differentiation of 4, then adjacent corner points exist one be 2 points, having two is 4 points, in conjunction with 6 any one
Point can distinguish 6 with other points, then combine can for a point of 1 central point, two adjacent corner points of 4 and 6
To distinguish 6 points completely, followed by two diagonal by image division, take 1/4th area images, and this is cut by employing
Cut each graphic feature of mode to extract and be stored in training storehouse.
5. a dice game device, including dice, box (100), photographic head (201), graphics processing unit (202), process
Device (300), memory element (400), output unit (500)；Wherein box (100) includes box body (101), lid (102), box body
(101) being provided with lamella lucida in, photographic head (201) is arranged on the lower section of box body (101) interior lamella lucida, and output unit (500) is to raise
Sound device and/or display device；
Can shoot the image of dice during photographic head (201) work from below, photographic head quantity can be one or more, photographic head
(201) connecting graphics processing unit (202), the image that photographic head (201) shoots processes by it, and the data after processing
It is sent to processor (300), with preprocessor (300) by described view data and the training storehouse of storage in memory element (400)
Image is compared, and comparison result is sent to output unit (500)；
It is characterized in that, it is judged that time its perform following steps,
Training storehouse builds: by being arranged on the six-face image of photographic head (201) the shooting dice of box body (101) bottom, to every
Shoot multiple the most from different perspectives, be then fed into graphics processing unit and process, wherein six-face image is carried out preliminary
The process abated the noise, compares hexahedro for dice figure subsequently, finds out the point of other face images of counting of each image
There is not the point that position overlaps in number, it is desirable to this spot size can stably realize the differentiation put with other；When not existing and other points
During the number the most misaligned point of figures, the optional point misaligned with image of partly counting is as characteristic point；And use multiple
The mode of some various combination carries out overall judgement；The multiple points 6 faces can distinguished completely by machine or manually find
Combination, and use Area comparison little and image is carried out by the partial graphical of combination that cutting easily can cover above-mentioned multiple point
Cutting, is stored in training storehouse after extracting characteristics of image；
Training storehouse image procossing mode stores: after training storehouse selects to determine the cutting mode of image, should by processor (300)
Kind of selection mode is sent to graphics processing unit (202), count in follow-up game identification time, graphics processing unit (202)
During to the process of image, extract the view data of dice in game in this manner, to facilitate and training storehouse comparison；
Taking pictures: after starting game, dice rotates a period of time stopping in box (100), now by being positioned at taking the photograph of cartridge bottom
As dice is shot by head (201) across transparent material；
Feature detection and extraction: after taking pictures, by image transmitting to graphics processing unit (202), first graphics processing unit enters
The certain pretreatment of row, then carries out limb recognition, wherein, by by the color gray value of dice body and dice box
Color gray value forms powerful contrast so that during image recognition, dice edge and background have obvious boundary, with 256 grades of gray scales
As a example by, gray value differences away from for 50-255, preferred scheme be gray scale difference be 150-255, if dice body is white, and dice
Box inwall is Dark grey or black, or in turn；Subsequently image is positioned, and come according to the cutting mode stored before
Cutting image extracts characteristics of image；
Image comparison, result output:
The image feature data extracted is sent into processor (300), with preprocessor (300) by it by graphics processing unit (202)
Compare with the view data in training storehouse, find out the point-and-figure chart picture of coupling, and result is sent by image and/or voice mode
To player.
6. dice game device as claimed in claim 5, wherein: die structure is, each some size in six faces is identical, and
The position of arbitrary angle point distance center of 2 .-6 is identical.
7. dice game device as claimed in claim 5, wherein: die structure is, each some size in six faces is identical, and
The position of arbitrary angle point distance center of 2 .-5 is identical, and 64 angle points are relative to being located closer to of angle point in other faces
Edge.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610741292.3A CN106326934A (en) | 2016-08-27 | 2016-08-27 | Dice throw-out point number judging method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610741292.3A CN106326934A (en) | 2016-08-27 | 2016-08-27 | Dice throw-out point number judging method and device |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN106326934A true CN106326934A (en) | 2017-01-11 |
Family
ID=57791215
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610741292.3A Pending CN106326934A (en) | 2016-08-27 | 2016-08-27 | Dice throw-out point number judging method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106326934A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111414918A (en) * | 2020-03-23 | 2020-07-14 | 哈工大机器人（合肥）国际创新研究院 | Dice point number identification method |
| CN112016479A (en) * | 2020-08-31 | 2020-12-01 | 重庆五诶科技有限公司 | Chess and card game auxiliary system |
Citations (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CA2344064A1 (en) * | 1998-09-15 | 2000-03-23 | Michail Order | Device for automatic detection of the number of spots on the top side of a dice for use on a professional basis, in particular for the game "craps" |
| CN1663654A (en) * | 2004-03-04 | 2005-09-07 | 赫德森索夫特株式会社 | Dice spot number determination method, dice spot number determination apparatus, game apparatus using same, and dice spot number determination game system |
| HK1073717A1 (en) * | 2003-12-04 | 2005-10-14 | Hudson Soft Co Ltd | Method and apparatus for dice eye number determination |
| CN1832787A (en) * | 2003-08-11 | 2006-09-13 | 大平技研工业株式会社 | Dice game machine |
| TW200915192A (en) * | 2007-09-21 | 2009-04-01 | Jumbo Technology Co Ltd | Method, apparatus, application program, and computer readable medium thereof for identifying points of at least one dice |
| US20090263008A1 (en) * | 2008-04-21 | 2009-10-22 | Tien-Shu Hsu | Method For Recognizing Dice Dots |
| SG142324A1 (en) * | 2008-04-28 | 2009-11-26 | Tien Shu Hsu | Method and apparatus for recognizing dice dots |
| CN102509383A (en) * | 2011-11-28 | 2012-06-20 | 哈尔滨工业大学深圳研究生院 | Feature detection and template matching-based mixed number identification method |
| CN103279745A (en) * | 2013-05-28 | 2013-09-04 | 东南大学 | Face identification method based on half-face multi-feature fusion |
| CN105844205A (en) * | 2015-01-15 | 2016-08-10 | 新天科技股份有限公司 | Character information identification method based on image processing |
| CN106334317A (en) * | 2016-08-27 | 2017-01-18 | 厦门市朗星节能照明股份有限公司 | Dice game device |
| CN106709484A (en) * | 2015-11-13 | 2017-05-24 | 国网吉林省电力有限公司检修公司 | Number identification method of digital instrument |
-
2016
- 2016-08-27 CN CN201610741292.3A patent/CN106326934A/en active Pending
Patent Citations (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CA2344064A1 (en) * | 1998-09-15 | 2000-03-23 | Michail Order | Device for automatic detection of the number of spots on the top side of a dice for use on a professional basis, in particular for the game "craps" |
| US6609710B1 (en) * | 1998-09-15 | 2003-08-26 | Michail Order | Device for automatic detection of the number of spots on the top side of a dice for use on a professional basis |
| CN1832787A (en) * | 2003-08-11 | 2006-09-13 | 大平技研工业株式会社 | Dice game machine |
| HK1073717A1 (en) * | 2003-12-04 | 2005-10-14 | Hudson Soft Co Ltd | Method and apparatus for dice eye number determination |
| CN1663654A (en) * | 2004-03-04 | 2005-09-07 | 赫德森索夫特株式会社 | Dice spot number determination method, dice spot number determination apparatus, game apparatus using same, and dice spot number determination game system |
| HK1075853A1 (en) * | 2004-03-04 | 2005-12-30 | Hudson Soft Co Ltd | Dice spot number determination method, dice spot number determinatioin apparatus, game apparatus using same, and dice spot number determination game system |
| TW200915192A (en) * | 2007-09-21 | 2009-04-01 | Jumbo Technology Co Ltd | Method, apparatus, application program, and computer readable medium thereof for identifying points of at least one dice |
| US20090263008A1 (en) * | 2008-04-21 | 2009-10-22 | Tien-Shu Hsu | Method For Recognizing Dice Dots |
| SG142324A1 (en) * | 2008-04-28 | 2009-11-26 | Tien Shu Hsu | Method and apparatus for recognizing dice dots |
| CN102509383A (en) * | 2011-11-28 | 2012-06-20 | 哈尔滨工业大学深圳研究生院 | Feature detection and template matching-based mixed number identification method |
| CN103279745A (en) * | 2013-05-28 | 2013-09-04 | 东南大学 | Face identification method based on half-face multi-feature fusion |
| CN105844205A (en) * | 2015-01-15 | 2016-08-10 | 新天科技股份有限公司 | Character information identification method based on image processing |
| CN106709484A (en) * | 2015-11-13 | 2017-05-24 | 国网吉林省电力有限公司检修公司 | Number identification method of digital instrument |
| CN106334317A (en) * | 2016-08-27 | 2017-01-18 | 厦门市朗星节能照明股份有限公司 | Dice game device |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111414918A (en) * | 2020-03-23 | 2020-07-14 | 哈工大机器人（合肥）国际创新研究院 | Dice point number identification method |
| CN111414918B (en) * | 2020-03-23 | 2023-09-08 | 哈工大机器人（合肥）国际创新研究院 | Dice point identification method |
| CN112016479A (en) * | 2020-08-31 | 2020-12-01 | 重庆五诶科技有限公司 | Chess and card game auxiliary system |
| CN112016479B (en) * | 2020-08-31 | 2024-05-31 | 重庆五诶科技有限公司 | Chess and card game auxiliary system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| KR102462409B1 (en) | Systems and Methods for Automated Table Game Activity Recognition | |
| US20240005726A1 (en) | System for counting quantity of game tokens | |
| US8016665B2 (en) | Table game tracking | |
| US8300935B2 (en) | Method and system for the detection and the classification of events during motion actions | |
| US20090135188A1 (en) | Method and system of live detection based on physiological motion on human face | |
| CN109446942A (en) | Method for tracking target, device and system | |
| EP2424207A1 (en) | Monitoring system | |
| US20060252554A1 (en) | Gaming object position analysis and tracking | |
| CN109635768B (en) | Method and system for detecting parking space state in image frame and related equipment | |
| US20130265434A1 (en) | Image processing apparatus and image processing method | |
| CN107025652A (en) | A kind of flame detecting method based on kinetic characteristic and color space time information | |
| CN102592147A (en) | Method and device for detecting human face | |
| CN105893946A (en) | Front face image detection method | |
| EP1107166A3 (en) | Device and method for face image extraction, and recording medium having recorded program for the method | |
| CN103092930B (en) | Method of generation of video abstract and device of generation of video abstract | |
| CN110648352A (en) | Abnormal event detection method and device and electronic equipment | |
| CN102799854A (en) | Image identification device and image identification method | |
| CN104091388A (en) | Paper currency authentic identification method and device based on magnetic images | |
| CN101782526A (en) | Method and device for automatically restoring, measuring and classifying steel dimple images | |
| JP2002344946A (en) | Monitoring system | |
| CN103279744A (en) | Multi-scale tri-mode texture feature-based method and system for detecting counterfeit fingerprints | |
| WO2011127808A1 (en) | Coin image identification method and device | |
| CN104616006A (en) | Surveillance video oriented bearded face detection method | |
| CN106326934A (en) | Dice throw-out point number judging method and device | |
| CN114140745A (en) | Method, system, device and medium for detecting personnel attributes of construction site |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20170111 |