CN103632380A - On-line game playing card identification method based on key point decision trees - Google Patents
On-line game playing card identification method based on key point decision trees Download PDFInfo
- Publication number
- CN103632380A CN103632380A CN201310529776.8A CN201310529776A CN103632380A CN 103632380 A CN103632380 A CN 103632380A CN 201310529776 A CN201310529776 A CN 201310529776A CN 103632380 A CN103632380 A CN 103632380A
- Authority
- CN
- China
- Prior art keywords
- key point
- color value
- white
- playing card
- board
- 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.)
- Granted
Links
Images
Abstract
The invention puts forward an on-line game playing card identification method based on key point decision trees. The method periodically captures the screenshot of a game interface, sequentially extracts each playing card image according to the coordinate values of matched boundary points, then rapidly identifies the mark of each playing card image according to the three-layered key point decision tree and finally identifies the points of each playing card image according to the ten-layered key point decision tree, so that a final identification result is obtained. The on-line game playing card identification method sufficiently utilizes the characteristics of the decision tree, i.e., low complexity and convenient computation, and the characteristics of on-line game playing cards, i.e., no rotation angles and clear images, the mark and points of each playing card image can be accurately and rapidly identified according to the color values of at most thirteen key points on the two decision trees, and the on-line game playing card identification method has the characteristics of low computational load, good real-timeness and high accuracy, thus overcoming the problems of high computational load, poor real-timeness and the like existing in the prior art and fully meeting the requirement of on-line game playing card identification on real-timeness and accuracy.
Description
Technical field
The present invention relates to image and process and recognition technology field, be specifically related to the playing card recognition methods of processing based on image.
Background technology
Playing card cards game, because of its interest and craftsmenship, is liked by numerous game players deeply.Particularly, along with the fast development of internet, the online snipsnapsnorum such as network fighting landlord, upgrading becomes more and more popular.But there is a large amount of audients' online playing card cards game, but lack so far the virtual playing card method for quickly identifying of requirement of real time.Game on line playing card identification requirement in moment, identify interface tens of boards pattern with count, it is the key components of many intelligent systems such as the process record system of playing a card, residue board prompt system of online snipsnapsnorum.The playing card recognition methods that current playing card recognition methods is mainly divided into the entity playing card recognition methods based on special marking and processes based on image.
Playing card described in China Patent Publication No. CN102989168A and recognition device thereof and recognition methods, first on playing card, print the photosensitive material of different photosensitive values, then utilize photoelectric conversion part that the optical signal sequence of photosensitive material reflection is converted into electrical signal sequence, finally by identification electric signal, reach quick identification playing card.The method of the identification playing card described in publication number CN1531703A is to use the ink stamped mark on playing card can only could showing below specific light, and identification is the object that the sign that goes out each playing card by specific device Identification display reaches quick identification playing card.The common ground of these two kinds of methods is that speed is fast, and accuracy rate is high, but is all that entity playing card based on having special marking are realized, so these two kinds of methods all cannot be applied to the virtual playing card identification of game on line.
The playing card recognition methods of processing based on image is broadly divided into again mode identification method and the method based on template matches based on sorter.A kind of networking full automatic cards distributing machine system described in China Patent Publication No. CN203075619U, wherein use camera to obtain playing card image, extract playing card pattern and count, having used the mode identification method based on nerual network technique to obtain last recognition result.The advantage of mode identification method is that extendability is better, can identify the playing card with rotation angle, but mode identification method needs a large amount of sample training, and computation complexity is larger, is difficult to meet completely the requirement of game on line real-time.In autoamtic card-dealers described in China Patent Publication No. CN100371042C, recognition device is to obtain playing card image by camera, the effective playing card view data obtaining and the standard deck type view data in board type database are contrasted to computing, by similarity, judge current playing card pattern and count.This template matching method is mainly applicable to not have the playing card identification of rotation angle, all standard deck type images in playing card image to be identified and sample database need to be compared one by one, finally find out similarity soprano, therefore its computing cost is very big, is also difficult to meet completely the requirement of game on line real-time.
Summary of the invention
For the problem such as the existing calculation of complex of prior art, speed be slower, the invention provides a kind of game on line playing card recognition methods based on key point decision tree, the method is utilized feature and the key point decision tree technique of game on line playing card irrotationality corner, clear picture, for every sheet playing card image, by the color value comparison of 13 key points, identify accurately and rapidly its pattern and count at the most, meeting game on line playing card completely and be identified in the demand in accuracy rate and real-time.
In order to realize foregoing invention object, the present invention has adopted following technological means: a kind of game on line playing card recognition methods based on key point decision tree, it comprises the steps:
(1) interface is periodically carried out to sectional drawing, according to sectional drawing result adaptation, go out the play a card coordinate figure of location boundary point of the coordinate figure of interface player's hand board frontier point under current interface size and each player;
(2) interface sectional drawing step (1) being obtained, the coordinate figure of the playing card frontier point going out according to adaptation finds position corresponding to every sheet playing card image, extracts successively and comprises pattern and every sheet playing card image of counting;
(3) every sheet playing card image step (2) being extracted, 3 layers of binary decision tree that form according to the color value by 4 key points, by contrasting successively wherein the color value of 3 key points at the most, identify the pattern of playing card fast;
(4) to pattern recognition result in step (3), not every sheet playing card image of large Xiao Wang, 10 layers of binary decision tree that form according to the color value by 13 key points, by contrasting successively wherein the color value of 10 key points at the most, counting of identification playing card, obtains final recognition result fast.
The process of the interface sectional drawing adaptation in step (1) is:
Interface is periodically carried out to sectional drawing, preserve truncated picture; According to sectional drawing result adaptation, go out the play a card coordinate figure of location boundary point of the coordinate figure of interface player's hand board frontier point under current interface size and each player, facilitate follow-up individual playing card image that carries out to extract.
Further, the process that the playing card image in step (2) extracts is:
The play a card coordinate figure of location boundary point of the coordinate figure of the player's hand board frontier point first obtaining according to step (1) interface sectional drawing adaptation and each player, the starting point coordinate that initial left frame of usining the first sheet playing card extracts as playing card, extraction comprises pattern and the first sheet playing card upper left corner rectangular image of the pattern of counting, then starting point coordinate is added to the width value of playing card, and this point is made as to new starting point coordinate, according to color value, judge whether this point is the left frame of playing card, left frame if not playing card, intercepting playing card stop, if the left frame of playing card, continue to extract the rectangular image in next the sheet playing card upper left corner, until all playing card image extracts complete.
Further, pattern identification in step (3) has adopted the technology based on key point decision tree, only need at the most the color value of 3 key points in the every sheet playing card image of comparison, being take in the upper left corner of playing card image to be identified in pattern identification is true origin (0, 0), image level is to the right x axle forward, image is set up rectangular coordinate system for y axle forward straight down, in this coordinate system x axle take playing card image to be identified horizontal width 1/21 be a coordinate unit, y axle take playing card image to be identified vertical height 1/39 be a coordinate unit, 4 key point coordinates that pattern identification relates to are: (5, 22), (8, 29), (13, 27), (9, 24), the pattern recognition result obtaining has 6 kinds: spade, heart, plum blossom, square, king, Xiao Wang, detailed process comprises the following steps:
(3.1) first judge key point (5 in this playing card image, 22) whether color value is white, if key point (5,22) is not white explanation, this board is not king or Xiao Wang, then judges this key point (5,22) whether color value is red, if the color value of key point (5,22) be redness this board for king, if key point (5,22) color value is not that red this board of explanation is Xiao Wang, obtains the pattern recognition result of this board; If color value is white in key point (5,22), need further to judge the pattern of this board;
(3.2) judge key point (8 in this playing card image, 29) whether red color value is, if the color value of key point (8,29) is red, illustrate that this board pattern is square or heart, whether white the color value that judges again key point (9,24) is, if key point (8,29) color value is that this board pattern of white explanation is heart, if the color value of key point (8,29) is not white explanation, this board pattern is not square, obtains the pattern recognition result of this board;
(3.3) if key point (8 in this playing card image, 29) color value is not red, illustrate that this board is spade or plum blossom, judge again whether key point (13,27) is white, if key point (8,29) color value is that this board pattern of white explanation is plum blossom, if the color value of key point (8,29) is not white explanation, this board pattern is not spade, obtains the pattern recognition result of this board.
Further, the identification of counting in step (4) has adopted the technology based on key point decision tree, only need at the most the color value of 10 key points in the every sheet playing card image of comparison, the identification of counting is used and rectangular coordinate system identical in step (3), 13 key point coordinates that relate to are: (17, 12), (11, 11), (3, 11), (7, 12), (2, 9), (2, 20), (4, 4), (3, 15), (9, 9), (9, 16), (9, 15), (4, 12), (6, 12), the recognition result of counting obtaining has 13 kinds: A, 2, 3, 4, 5, 6, 7, 8, 9, 10, J, Q, K, detailed process comprises the following steps:
(4.1) if the result of the pattern in step (3) identification be king or Xiao Wang, do not need so to continue identification and count, if not king or Xiao Wang, need to continue to identify and count;
(4.2) judge in this playing card image whether key point (17,12) is white, if the color value of key point (17,12) is not white, illustrate that this board is a kind of in J, Q, K, need to further identify.If the color value of key point (11,11) is not white, the recognition result of counting of this board is J; If the color value of key point (17,12) is the color value that white needs judgement key point (3,11), if the color value of fruit dot (3,11) is not white, the recognition result of counting of this board is Q, if the recognition result of counting of this board of white is K;
(4.3) if the middle key point (17 of step (4.2), 12) color value is white, judge key point (2,9) whether color value is white, if the color value of key point (2,9) is not white, the recognition result of counting of this board is 10, if being white, the color value of key point (2,9) needs to continue to identify to count;
(4.4) if the middle key point (2 of step (4.3), 9) color value is white, judge key point (2,20) whether color value is white, if the color value of key point (2,20) is not white, the recognition result of counting of this board is A, if being white, the color value of key point (2,20) needs to continue to identify to count;
(4.5) if the middle key point (2 of step (4.4), 20) color value is white, judge key point (4,4) whether color value is white, if the color value of key point (4,4) is not white, the recognition result of counting of this board is 7, if being white, the color value of key point (4,4) needs to continue to identify to count;
(4.6) if the middle key point (4 of step (4.5), 4) color value is white, judge key point (3,15) whether color value is white, if the color value of key point (3,15) is not white, the recognition result of counting of this board is 4, if being white, the color value of key point (3,15) needs to continue to identify to count;
(4.7) if the middle key point (3 of step (4.6), 15) color value is white, judge key point (9,9) whether color value is white, if the color value of key point (9,9) is not white, the recognition result of counting of this board is 5, if being white, the color value of key point (9,9) needs to continue to identify to count;
(4.8) if the middle key point (9 of step (4.7), 9) color value is white, judge key point (9,26) whether color value is white, if the color value of key point (9,26) is not white, the recognition result of counting of this board is 2, if being white, the color value of key point (9,26) needs to continue to identify to count;
(4.9) if the middle key point (9 of step (4.8), 16) color value is white, judge key point (9,15) whether color value is white, if the color value of key point (9,15) is not white, the recognition result of counting of this board is 9, if being white, the color value of key point (9,15) needs to continue to identify to count;
(4.10) if the middle key point (9 of step (4.9), 15) color value is white, judge key point (4,12) whether color value is white, if the color value of key point (4,12) is not white, the recognition result of counting of this board is 6, if being white, the color value of key point (4,12) needs to continue to identify to count;
(4.11) if the middle key point (4 of step (4.10), 12) color value is white, judge key point (6,12) whether color value is white, if the color value of key point (6,12) is not white, the recognition result of counting of this board is 8, if the color value of key point (6,12) is that the recognition result of counting of this board of white is 3.
Compared with prior art, the present invention has following beneficial effect:
1, the game on line playing card recognition methods based on key point decision tree that the present invention adopts, make full use of key point decision tree technique, for every sheet playing card image, by the color value of 13 key points on decision tree, identify accurately and rapidly its pattern and count at the most.With respect to the mode identification method based on sorter and the recognition methods based on template matches, the method neither needs to gather a large amount of sample images, also not needing to carry out one by one similarity with template image compares, calculated amount is very little, application the method on the computing machine of WIN 7 operating systems, Intel Core I7 CPU, 4G internal memory, can in 0.1 second, identify interface in 17 sheet playing cards pattern with count, meet game on line playing card completely and be identified in the demand in real-time.
2, the game on line playing card recognition methods based on key point decision tree that the present invention adopts, the key point of using in pattern identification and the identification of counting makes full use of the feature of game on line playing card irrotationality corner, clear picture, internal feature and the outside difference that can truly reflect each playing card, in the playing card identification of 1000 different interfaces, correctly identify the pattern of all playing card and count, meeting game on line playing card completely and be identified in the demand in accuracy.
Accompanying drawing explanation
Fig. 1 is the schematic flow sheet of the inventive method;
Fig. 2 is the adaptive process flow diagram of interface sectional drawing;
Fig. 3 is that playing card image extracts process flow diagram;
Fig. 4 is 3 layers of pattern recognition decision tree based on 4 key points;
Fig. 5 is 10 layers of recognition decision tree of counting based on 13 key points;
Fig. 6 is the distribution schematic diagram of key point in the identification of playing card pattern;
Fig. 7 is the playing card distribution schematic diagrams of key point in identification of counting;
Fig. 8 is playing card region, game on line interface sectional drawing;
Fig. 9 is the Xiao Wang's playing card image extracting;
Figure 10 is the heart A playing card image extracting.
Embodiment
Below in conjunction with accompanying drawing, describe concrete grammar of the present invention in detail.
A kind of game on line playing card recognition methods based on key point decision tree provided by the invention, as shown in Figure 1, comprises the following steps:
(1) interface sectional drawing is adaptive
As shown in Figure 2, first periodically interface is carried out to sectional drawing, preserve the picture of current intercepting, obtain the approximate region of playing card, and go out the play a card coordinate figure of location boundary point of the coordinate figure of interface player's hand board frontier point under current interface size and each player according to sectional drawing result adaptation, facilitate follow-up individual playing card image that carries out to extract.Wherein obtaining of frontier point coordinate figure is that the color value of each pixel and the color value scope of playing card left frame by adaptive area contrasts, if do not mated, current coordinate points adds 1 automatically, if coupling, the left frame that finds playing card is described, continue coupling, until the frontier point of whole playing card of making uniform.
(2) playing card image extracts
As shown in Figure 3, the playing card region first obtaining according to step (1) interface sectional drawing adaptation, upgrade the beginning coordinate that playing card extract, upgrade the starting point coordinate of the left frame of the first sheet playing card, then extract and comprise pattern and individual playing card rectangular image of the pattern of counting, starting point coordinate is added to the width value of playing card, and this point is made as to new starting point coordinate, according to color value, judge whether this point is the left frame of playing card, left frame if not playing card, intercepting playing card stop, if the left frame of playing card, continue to extract the rectangular image of next sheet playing card, until all playing card image extracts complete.
(3) identification of the pattern based on key point decision tree
Every sheet playing card image that step (2) is extracted, one the 3 layers binary decision tree (as shown in Figure 4) that form according to the color value by 4 key points, by contrasting successively wherein the color value of 3 key points at the most, identify the pattern of playing card fast.Being take in the upper left corner of playing card image to be identified in the identification of pattern based on key point decision tree is true origin (0, 0), image level is to the right x axle forward, image is set up rectangular coordinate system for y axle forward straight down, in this coordinate system x axle take playing card image to be identified horizontal width 1/21 be a coordinate unit, y axle take playing card image to be identified vertical height 1/39 be a coordinate unit, 4 key point coordinates that pattern identification relates to are so: (5, 22), (8, 29), (13, 27), (9, 24), as shown in Figure 6, the pattern recognition result obtaining has 6 kinds: spade, heart, plum blossom, square, king, Xiao Wang, detailed process is as follows:
(3.1) first judge key point (5 in this playing card image, 22) whether color value is white, if the color value of key point (5,22) is not white explanation, this board is not king or Xiao Wang, then judges this point (5,22) whether color value is red, if the color value of key point (5,22) is red, the pattern recognition result of this board is king, if the pattern recognition result that the color value of key point (5,22) is not red this board is Xiao Wang; If the color value of key point (5,22) is white, need to continue identification pattern;
(3.2) judge key point (8 in this playing card image, 29) whether red color value is, if the color value of key point (8,29) is red, illustrate that this board pattern is square or heart, whether white the color value that judges again key point (9,24) is, if key point (9,24) color value is white, the pattern recognition result of this board is heart, if the color value of key point (9,24) is not that the white pattern recognition result of this board is square;
(3.3) if key point (8 in this playing card image, 29) color value is not red, illustrate that this board is spade or plum blossom, whether the color value that judges again key point (13,27) is white, if key point (13,27) color value is white, the pattern recognition result of this board is plum blossom, if the color value of key point (13,27) is not that the white pattern recognition result of this board is spade.
(4) identification of counting based on key point decision tree
To pattern recognition result in step (3), it not every sheet playing card image of king, Xiao Wang, one the 10 layers binary decision tree (as shown in Figure 5) that form according to the color value by 13 key points, by contrasting successively wherein the color value of 10 key points at the most, counting of identification playing card, obtains final recognition result fast.The identification of counting based on key point decision tree is used and rectangular coordinate system identical in step (3), 13 key point coordinates that the identification of counting so relates to are: (17,12), (11,11), (3,11), (7,12), (2,9), (2,20), (4,4), (3,15), (9,9), (9,16), (9,15), (4,12), (6,12), as shown in Figure 7, the recognition result of counting obtaining has 13 kinds: A, 2,3,4,5,6,7,8,9,10, J, Q, K, and detailed process is as follows:
(4.1) if the result of the pattern in step (3) identification be king or Xiao Wang, do not need so to continue identification and count, if not king or Xiao Wang, need to continue to identify and count;
(4.2) judge in this playing card image whether key point (17,12) is white, if the color value of key point (17,12) is not white, illustrate that this board is a kind of in J, Q, K, need to further identify.If the color value of key point (11,11) is not white, the recognition result of counting of this board is J; If the color value of key point (17,12) is the color value that white needs judgement key point (3,11), if the color value of fruit dot (3,11) is not white, the recognition result of counting of this board is Q, if the recognition result of counting of this board of white is K;
(4.3) if the middle key point (17 of step (4.2), 12) color value is white, judge key point (2,9) whether color value is white, if the color value of key point (2,9) is not white, the recognition result of counting of this board is 10, if being white, the color value of key point (2,9) needs to continue to identify to count;
(4.4) if the middle key point (2 of step (4.3), 9) color value is white, judge key point (2,20) whether color value is white, if the color value of key point (2,20) is not white, the recognition result of counting of this board is A, if being white, the color value of key point (2,20) needs to continue to identify to count;
(4.5) if the middle key point (2 of step (4.4), 20) color value is white, judge key point (4,4) whether color value is white, if the color value of key point (4,4) is not white, the recognition result of counting of this board is 7, if being white, the color value of key point (4,4) needs to continue to identify to count;
(4.6) if the middle key point (4 of step (4.5), 4) color value is white, judge key point (3,15) whether color value is white, if the color value of key point (3,15) is not white, the recognition result of counting of this board is 4, if being white, the color value of key point (3,15) needs to continue to identify to count;
(4.7) if the middle key point (3 of step (4.6), 15) color value is white, judge key point (9,9) whether color value is white, if the color value of key point (9,9) is not white, the recognition result of counting of this board is 5, if being white, the color value of key point (9,9) needs to continue to identify to count;
(4.8) if the middle key point (9 of step (4.7), 9) color value is white, judge key point (9,26) whether color value is white, if the color value of key point (9,26) is not white, the recognition result of counting of this board is 2, if being white, the color value of key point (9,26) needs to continue to identify to count;
(4.9) if the middle key point (9 of step (4.8), 16) color value is white, judge key point (9,15) whether color value is white, if the color value of key point (9,15) is not white, the recognition result of counting of this board is 9, if being white, the color value of key point (9,15) needs to continue to identify to count;
(4.10) if the middle key point (9 of step (4.9), 15) color value is white, judge key point (4,12) whether color value is white, if the color value of key point (4,12) is not white, the recognition result of counting of this board is 6, if being white, the color value of key point (4,12) needs to continue to identify to count;
(4.11) if the middle key point (4 of step (4.10), 12) color value is white, judge key point (6,12) whether color value is white, if the color value of key point (6,12) is not white, the recognition result of counting of this board is 8, if the color value of key point (6,12) is that the recognition result of counting of this board of white is 3.
By instantiation, further illustrate application of the present invention below.
Step 1: interface sectional drawing is adaptive
According to the sectional drawing to games window, obtain the approximate region of playing card as shown in Figure 8, then contrast the color value of playing card left frame frontier point, find the left frame of playing card.
Step 2: playing card image extracts
The the first sheet playing card left frame finding according to step 1, intercept the first sheet playing card upper left corner rectangular image, then starting point coordinate is added to the width value of intercepting playing card, and this point is made as to starting point coordinate, according to color value, judge whether this point is the left frame of playing card, if not the left frame of playing card, intercepting playing card stop, if the left frame of playing card continues the rectangular image in next the sheet playing card upper left corner of intercepting, until all playing card intercepting is complete.
Step 3: the pattern identification based on key point decision tree
The playing card image that step 2 is extracted, Xiao Wang and the heart A of take in Fig. 9 and Figure 10 are example explanation, first 3 layers of pattern recognition decision tree based on 4 key points are to the key point (5 in Fig. 9,22) color value is judged, judgement key point (5,22) whether white color value is, through judging key point (5,22) color value is not white, now according to binary decision tree, judges that more whether red the color value of key point (5,22) is, through judging key point (5,22) color value is not red, judges that playing card result is as Xiao Wang, obtains the pattern recognition result of Fig. 9.
For Figure 10, first 3 layers of pattern recognition decision tree based on 4 key points judge key points (5, 22) whether color value is white, through judging key point (5, 22) color value is white, illustrate and be not king or Xiao Wang, continue judgement key point (8, 29) whether color value is red, through judging key point (8, 29) color value is red, illustrate that these playing card are heart or square, continue judgement key point (9, 24) whether color value is white, through judging key point (9, 24) color value is white, the pattern that obtains these playing card is heart, pattern end of identification to Figure 10.
Step 4: the identification of counting based on key point decision tree
To Fig. 9 and Figure 10 identification of counting, because the pattern of Fig. 9 has been identified as Xiao Wang, without the identification of counting again, because Figure 10 is not king or Xiao Wang, need be to Figure 10 identification of counting, according to key point (17 in 10 layers of recognition decision tree judgement Figure 10 that counts based on 13 key points, 12) whether color value is white, through judging, be white, continue judgement key point (2, 9) whether color value is white, through judging key point (2, 9) color value is white, continue judgement key point (2, 20) whether color value is white, through judging key point (2, 20) color value is not white, obtain counting as A of these playing card, integrating step 3 is known, Figure 10 is heart A, to Figure 10 end of identification.
Those skilled in the art will readily understand; the foregoing is only preferred embodiment of the present invention; not in order to limit the present invention, all any modifications of doing within the spirit and principles in the present invention, be equal to and replace and improvement etc., within all should being included in protection scope of the present invention.
Claims (5)
1. the game on line playing card recognition methods based on key point decision tree, is characterized in that, described method comprises following steps:
(1) interface is periodically carried out to sectional drawing, according to sectional drawing result adaptation, go out the play a card coordinate figure of location boundary point of the coordinate figure of interface player's hand board frontier point under current interface size and each player;
(2) interface sectional drawing step (1) being obtained, the coordinate figure of the playing card frontier point going out according to adaptation finds position corresponding to every sheet playing card image, extracts successively and comprises pattern and every sheet playing card image of counting;
(3) every sheet playing card image step (2) being extracted, 3 layers of binary decision tree that form according to the color value by 4 key points, by contrasting successively wherein the color value of 3 key points at the most, identify the pattern of playing card fast;
(4) to pattern recognition result in step (3), not every sheet playing card image of large Xiao Wang, 10 layers of binary decision tree that form according to the color value by 13 key points, by contrasting successively wherein the color value of 10 key points at the most, counting of identification playing card, obtains final recognition result fast.
2. a kind of game on line playing card recognition methods based on key point decision tree according to claim 1, is characterized in that, the process of the interface sectional drawing adaptation in step (1) is:
Interface is periodically carried out to sectional drawing, preserve truncated picture; According to sectional drawing result adaptation, go out the play a card coordinate figure of location boundary point of the coordinate figure of interface player's hand board frontier point under current interface size and each player, facilitate follow-up individual playing card image that carries out to extract.
3. a kind of game on line playing card recognition methods based on key point decision tree according to claim 1, is characterized in that, the process that the playing card image in step (2) extracts is:
The play a card coordinate figure of location boundary point of the coordinate figure of the player's hand board frontier point first obtaining according to step (1) interface sectional drawing adaptation and each player, the starting point coordinate that initial left frame of usining the first sheet playing card extracts as playing card, extraction comprises pattern and the first sheet playing card upper left corner rectangular image of the pattern of counting, then starting point coordinate is added to the width value of playing card, and this point is made as to new starting point coordinate, according to color value, judge whether this point is the left frame of playing card, left frame if not playing card, intercepting playing card stop, if the left frame of playing card, continue to extract the rectangular image in next the sheet playing card upper left corner, until all playing card image extracts complete.
4. a kind of game on line playing card recognition methods based on key point decision tree according to claim 1, it is characterized in that, pattern identification in step (3) has adopted the technology based on key point decision tree, only need at the most the color value of 3 key points in the every sheet playing card image of comparison, being take in the upper left corner of playing card image to be identified in pattern identification is true origin (0, 0), image level is to the right x axle forward, image is set up rectangular coordinate system for y axle forward straight down, in this coordinate system x axle take playing card image to be identified horizontal width 1/21 be a coordinate unit, y axle take playing card image to be identified vertical height 1/39 be a coordinate unit, 4 key point coordinates that pattern identification relates to are: (5, 22), (8, 29), (13, 27), (9, 24), the pattern recognition result obtaining has 6 kinds: spade, heart, plum blossom, square, king, Xiao Wang, detailed process comprises the following steps:
(3.1) first judge key point (5 in this playing card image, 22) whether color value is white, if key point (5,22) is not white explanation, this board is not king or Xiao Wang, then judges this key point (5,22) whether color value is red, if the color value of key point (5,22) be redness this board for king, if key point (5,22) color value is not that red this board of explanation is Xiao Wang, obtains the pattern recognition result of this board; If color value is white in key point (5,22), need further to judge the pattern of this board;
(3.2) judge key point (8 in this playing card image, 29) whether red color value is, if the color value of key point (8,29) is red, illustrate that this board pattern is square or heart, whether white the color value that judges again key point (9,24) is, if key point (8,29) color value is that this board pattern of white explanation is heart, if the color value of key point (8,29) is not white explanation, this board pattern is not square, obtains the pattern recognition result of this board;
(3.3) if key point (8 in this playing card image, 29) color value is not red, illustrate that this board is spade or plum blossom, judge again whether key point (13,27) is white, if key point (8,29) color value is that this board pattern of white explanation is plum blossom, if the color value of key point (8,29) is not white explanation, this board pattern is not spade, obtains the pattern recognition result of this board.
5. a kind of game on line playing card recognition methods based on key point decision tree according to claim 1, it is characterized in that, the identification of counting in step (4) has adopted the technology based on key point decision tree, only need at the most the color value of 10 key points in the every sheet playing card image of comparison, the identification of counting is used and rectangular coordinate system identical in step (3), 13 key point coordinates that relate to are: (17, 12), (11, 11), (3, 11), (7, 12), (2, 9), (2, 20), (4, 4), (3, 15), (9, 9), (9, 16), (9, 15), (4, 12), (6, 12), the recognition result of counting obtaining has 13 kinds: A, 2, 3, 4, 5, 6, 7, 8, 9, 10, J, Q, K, detailed process comprises the following steps:
(4.1) if the result of the identification of the pattern in step (3) is king or Xiao Wang, do not need so to continue identification and count, if not king or Xiao Wang, need to continue identification and count;
(4.2) judge in this playing card image whether key point (17,12) is white, if the color value of key point (17,12) is not white, illustrate that this board is a kind of in J, Q, K, need to further identify;
If the color value of key point (11,11) is not white, the recognition result of counting of this board is J; If the color value of key point (17,12) is white, need to judge key point (3,11) color value, if the color value of fruit dot (3,11) is not white, the recognition result of counting of this board is Q, if white, the recognition result of counting of this board is K;
(4.3) if the middle key point (17 of step (4.2), 12) color value is white, judge key point (2,9) whether color value is white, if the color value of key point (2,9) is not white, the recognition result of counting of this board is 10, if the color value of key point (2,9) is white, needs to continue identification and count;
(4.4) if the middle key point (2 of step (4.3), 9) color value is white, judge key point (2,20) whether color value is white, if the color value of key point (2,20) is not white, the recognition result of counting of this board is A, if the color value of key point (2,20) is white, needs to continue identification and count;
(4.5) if the middle key point (2 of step (4.4), 20) color value is white, judge key point (4,4) whether color value is white, if the color value of key point (4,4) is not white, the recognition result of counting of this board is 7, if the color value of key point (4,4) is white, needs to continue identification and count;
(4.6) if the middle key point (4 of step (4.5), 4) color value is white, judge key point (3,15) whether color value is white, if the color value of key point (3,15) is not white, the recognition result of counting of this board is 4, if the color value of key point (3,15) is white, needs to continue identification and count;
(4.7) if the middle key point (3 of step (4.6), 15) color value is white, judge key point (9,9) whether color value is white, if the color value of key point (9,9) is not white, the recognition result of counting of this board is 5, if the color value of key point (9,9) is white, needs to continue identification and count;
(4.8) if the middle key point (9 of step (4.7), 9) color value is white, judge key point (9,26) whether color value is white, if the color value of key point (9,26) is not white, the recognition result of counting of this board is 2, if the color value of key point (9,26) is white, needs to continue identification and count;
(4.9) if the middle key point (9 of step (4.8), 16) color value is white, judge key point (9,15) whether color value is white, if the color value of key point (9,15) is not white, the recognition result of counting of this board is 9, if the color value of key point (9,15) is white, needs to continue identification and count;
(4.10) if the middle key point (9 of step (4.9), 15) color value is white, judge key point (4,12) whether color value is white, if the color value of key point (4,12) is not white, the recognition result of counting of this board is 6, if the color value of key point (4,12) is white, needs to continue identification and count;
(4.11) if the middle key point (4 of step (4.10), 12) color value is white, judge key point (6,12) whether color value is white, if the color value of key point (6,12) is not white, the recognition result of counting of this board is 8, if the color value of key point (6,12) is white, the recognition result of counting of this board is 3.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310529776.8A CN103632380B (en) | 2013-11-01 | 2013-11-01 | On-line game playing card identification method based on key point decision trees |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310529776.8A CN103632380B (en) | 2013-11-01 | 2013-11-01 | On-line game playing card identification method based on key point decision trees |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN103632380A true CN103632380A (en) | 2014-03-12 |
| CN103632380B CN103632380B (en) | 2017-04-12 |
Family
ID=50213396
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201310529776.8A Active CN103632380B (en) | 2013-11-01 | 2013-11-01 | On-line game playing card identification method based on key point decision trees |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103632380B (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104874173A (en) * | 2015-05-25 | 2015-09-02 | 厦门大学 | Automatic bridge calling recognition method and system based on deep belief network |
| CN110852436A (en) * | 2019-10-18 | 2020-02-28 | 桂林力港网络科技股份有限公司 | Data processing method, device and storage medium for electronic poker game |
| CN112463992A (en) * | 2021-02-04 | 2021-03-09 | 中至江西智能技术有限公司 | Decision-making auxiliary automatic question-answering method and system based on knowledge graph in mahjong field |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20040094895A1 (en) * | 2002-11-20 | 2004-05-20 | Ronald Nasco | Multi-player card game and method of playing the same |
| CN102722736A (en) * | 2012-06-13 | 2012-10-10 | 合肥工业大学 | Method for splitting and identifying character strings at complex interference |
-
2013
- 2013-11-01 CN CN201310529776.8A patent/CN103632380B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20040094895A1 (en) * | 2002-11-20 | 2004-05-20 | Ronald Nasco | Multi-player card game and method of playing the same |
| CN102722736A (en) * | 2012-06-13 | 2012-10-10 | 合肥工业大学 | Method for splitting and identifying character strings at complex interference |
Non-Patent Citations (2)
| Title |
|---|
| SIJIE WU等: "Forest Cover Classification from Multi-temporal MODIS Images in Southeast Asia Using Decision Tree", 《 COMPUTER AND COMPUTING TECHNOLOGIES IN AGRICULTURE Ⅴ》 * |
| 寇永军: "图像识别在发牌机上的应用研究", 《中国优秀硕士学位论文全文数据库信息科技辑》 * |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104874173A (en) * | 2015-05-25 | 2015-09-02 | 厦门大学 | Automatic bridge calling recognition method and system based on deep belief network |
| CN110852436A (en) * | 2019-10-18 | 2020-02-28 | 桂林力港网络科技股份有限公司 | Data processing method, device and storage medium for electronic poker game |
| CN112463992A (en) * | 2021-02-04 | 2021-03-09 | 中至江西智能技术有限公司 | Decision-making auxiliary automatic question-answering method and system based on knowledge graph in mahjong field |
| CN112463992B (en) * | 2021-02-04 | 2021-06-11 | 中至江西智能技术有限公司 | Decision-making auxiliary automatic question-answering method and system based on knowledge graph in mahjong field |
Also Published As
| Publication number | Publication date |
|---|---|
| CN103632380B (en) | 2017-04-12 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108229509B (en) | Method and device for identifying object class and electronic equipment | |
| CN110232311B (en) | Method and device for segmenting hand image and computer equipment | |
| WO2020151489A1 (en) | Living body detection method based on facial recognition, and electronic device and storage medium | |
| CN107019901B (en) | Method for establishing chess and card game automatic gaming robot based on image recognition and automatic control | |
| CN111971685A (en) | Machine learning driven object detection system and method | |
| CN104463250B (en) | A kind of Sign Language Recognition interpretation method based on Davinci technology | |
| CN102704215B (en) | Automatic cutting method of embroidery cloth based on combination of DST file parsing and machine vision | |
| CN103065134A (en) | Fingerprint identification device and method with prompt information | |
| CN104517101A (en) | Game poker card recognition method based on pixel square difference matching | |
| KR20120089452A (en) | System and method for object extraction | |
| CN109190643A (en) | Based on the recognition methods of convolutional neural networks Chinese medicine and electronic equipment | |
| CN109409377B (en) | Method and device for detecting characters in image | |
| CN106127205A (en) | A kind of recognition methods of the digital instrument image being applicable to indoor track machine people | |
| CN104346816A (en) | Depth determining method and device and electronic equipment | |
| CN103198304A (en) | Palm print extraction and identification method | |
| CN109559362B (en) | Image subject face replacing method and device | |
| CN107194361A (en) | Two-dimentional pose detection method and device | |
| CN110119680A (en) | A kind of electrical cabinet wiring automatic errordetecting system based on image recognition | |
| CN104951940A (en) | Mobile payment verification method based on palmprint recognition | |
| CN106483130B (en) | A kind of detection method and its automatic detection device of rice disease | |
| CN106339707A (en) | Instrument pointer image recognition method based on symmetrical characteristics | |
| WO2018032626A1 (en) | Teaching toy kit and number identification method thereof | |
| CN111445459A (en) | Image defect detection method and system based on depth twin network | |
| CN107506738A (en) | Feature extracting method, image-recognizing method, device and electronic equipment | |
| CN102184404B (en) | Method and device for acquiring palm region in palm image |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |