CN102270298A - Method and device for detecting laser point/area - Google Patents

Abstract

The invention provides a method and a device for detecting a laser point/area. The method for detecting the laser point/area comprises the following steps of: acquiring a characteristic value of a standard laser point/area; generating a candidate laser point/area from an input image; calculating the offset value of the characteristic value of each candidate laser point/area relative to the characteristic value of the standard laser point/area; and determining the candidate laser point/area with the minimum characteristic offset value as the laser point/area. The method for detecting the laser point/area can support the detection of various laser points/ areas, and has high environmental suitability.

Description

Laser spots/method for detecting area and device

Technical field

The present invention relates to Flame Image Process, relate more specifically to laser spots/method for detecting area and device.

Background technology

Have some and computing machine to carry out remote mutual system and method at present, their purpose is for the user is more effectively carried out as the activities such as demonstration based on computing machine and projector.These system and methods have adopted multiple technologies such as ultrasound wave, infrared etc. have realized remote mutual solution.

Under speech or conferencing environment, laser pen is widely used in pointing to the specific part of view field.Because laser pen is popular and low-cost, some telecounter interactive systems and method adopt laser pen and camera to realize, they provide the operation of long distance laser pen for the user by the following method: detection laser point/zone from the screen picture that camera captures, and then the control operation of analog computer keyboard/mouse.

In order to use laser pen and computing machine to carry out alternately, proposed the laser spots follow-up mechanism and come the position of detection laser point and be used for the operational computations machine.US patent 6275214 B1 have described a kind of interactive strategy: one or more attributes of detection laser point; Thereby any one or a plurality of attribute can combined use generate the right instructions to computing machine.These attributes may comprise intensity, color, shape, size, move mode and position.US patent 2008/0266253 A1 has described a kind of laser spots detection and tracking method of polygon coupling: the parameterized laser polygon of predefine; The profile of each Drawing Object in polygon and the view data is mated; Checking whether the polygon that coupling obtains satisfies the predefine parameter, then is candidate's laser polygon as satisfying.

Article Laser pointer interaction, Dan R.Olsen Jr., Travis Nielsen, Proceedingsof Association for Computing Machinery Computer Human InteractionConference:Human Factors in Computing Systems.Seattle, WA, USA:2001:17～22 have been described under the situation of using convolution filter, and the brightest red luminous point is detected as laser spots.Article LumiPoint:Multi-user Laser-based interaction on Large Tiled Displays, James Davis and Xing Chen, Displays, Vol.23-5, Elsevier Science 2002 has described the use luminance threshold and has carried out pre-service, use down-sampling and prediction to accelerate detection speed then, comprise the stroke server simultaneously and carry out a location estimation.Article Laser Actuated Presentation System, AtulChowdhary, Vivek Agrawal, Subhajit Karmakar, Sandip Sarkar, CoRR Abs/0911.5404: 2009 have described the RGB camera that uses filter redly catches laser spots, and the Blue layer that uses RGB then is as the threshold value attribute.

In the prior art, most of paper/patents are carried out brightness or color attribute as " laser spots/zone is the brightest " or " laser spots/zone is the reddest or the greenest " etc. to image based on assumed conditions and are filtered and realize that always laser spots/zone detects, and have some papers/patent to use additive method such as technology such as form fit, trajectory predictions to realize laser spots/zone detection in addition.In addition, aspect hardware device, some paper/patent is used specific camera head or extra filter, disturbs thereby reduce external environment condition.

As an open loop (Open Loop) problem, laser spots/method for detecting area can not reach 100% verification and measurement ratio in theory.Some object in the two field picture that camera captures may disturb laser spots/zone to detect, even if use the similar environmental interference of specific hardware device effectively to be eliminated.In addition, according to investigation and experiment, under the varying environment condition, there is different property value performances in the laser spots/zone that is captured; Even if in same environmental conditions, laser spots/zone does not have foregoing fixed attribute value tag, and for example in some cases, noise spot/zone shows brightlyer, redder or greener than laser spots/zone.

For the two field picture that noise spot/zone is arranged, done will the lead to errors detection of laser spots/method for detecting area in the existing paper/patent of laser spots/area attribute supposition.

Summary of the invention

In view of the problems referred to above of the prior art, the present invention has been proposed.

According to an aspect of the present invention, provide a kind of laser spots/method for detecting area, can may further comprise the steps: the eigenwert of obtaining standard laser point/zone; From input picture, produce laser spots/zone of candidate; The eigenwert in laser spots/zone of calculating each candidate is with respect to the off-set value of the eigenwert in standard laser point/zone; And determine to possess the candidate's laser spots/zone of minimal characteristic off-set value as laser spots/zone.

This laser spots/method for detecting area can also comprise the eigenwert of upgrading standard laser point/zone based on the characteristic information in determined laser spots/zone.

Described laser spots/the zone that produces the candidate from input picture can comprise with the predetermined characteristic threshold value filters to obtain laser spots/zone of candidate two field picture.

This laser spots/method for detecting area can also comprise: the pre-defined predetermined number that does not overlap each other laser spots/area type; The described eigenwert of obtaining standard laser point/zone comprises, obtains the eigenwert in each standard laser point/zone of this predetermined number type respectively; The described laser spots zone that produces the candidate from input picture comprises, utilize with this predetermined number laser spots/area type in every type of characteristic threshold value that is associated, generation is divided into the laser spots/zone of the candidate in predetermined number laser spots/area type; The eigenwert in laser spots/zone of described each candidate of calculating comprises with respect to the off-set value of the eigenwert in standard laser point/zone, calculate this candidate laser spots/zone eigenwert with respect to laser spots/zone of this candidate under the off-set value of eigenwert in standard laser point/zone of the corresponding type of type; Describedly determine that the laser spots/zone that possesses the minimal characteristic off-set value comprises as laser spots/zone, determine laser spots/zone as the type with being divided into the candidate's laser spots/zone that possesses the minimal characteristic off-set value in laser spots/zone of all candidates in laser spots/area type.

The described eigenwert of obtaining standard laser point/zone can comprise at least a in the following operation: the fixed value that the eigenwert in standard laser point/zone is defined as being scheduled to; To every kind of laser spots/area type, the several leading two field picture that comprises respective type laser spots/zone behind the described laser spots/method for detecting area of startup is considered as primary data, obtain the eigenwert in the standard laser point/zone of the type based on this primary data training; From training the eigenwert in every type the standard laser point/zone that obtains based on a large amount of primary datas at every kind of laser spots/area type in advance, obtain the eigenwert in the standard laser point/zone of the type.

Described predetermined characteristic threshold value can comprise one or more in each color attribute threshold value.

Feature can comprise with the next item down or multinomial: color; Size and shape; And position and motion track.

The eigenwert of calculating laser spots/zone of each candidate can comprise with respect to the off-set value of the eigenwert in standard laser point/zone: the color character value that the color character value that comprises based on candidate laser spots/zone and standard laser point/zone have been estimated is calculated the color character off-set value; Size that comprises based on candidate laser spots/zone and shape facility value and size and shape facility value that the standard laser point/zone has been estimated are calculated size and shape facility off-set value; Position that comprises based on candidate laser spots/zone and motion track eigenwert and position and motion track eigenvalue calculation position and motion track feature off-set value that the standard laser point/zone has been estimated; For color, size and shape, position and motion track feature off-set value are provided with weighting factor respectively; The feature off-set value in candidate's laser spots/zone is following three value sums: the color character off-set value multiply by the respective weight factor of setting, size and shape facility off-set value multiply by the respective weight factor of setting, and position and motion track feature off-set value multiply by the respective weight factor of setting.

In addition, if candidate laser spots/zone any in meeting the following conditions, then it can be considered as noise spot/zone, and abandon: the predetermined characteristic property value does not meet predetermined value or preset range; The color character off-set value does not meet predetermined value; Size and shape facility off-set value do not meet predetermined value; Position and motion track feature off-set value do not meet predetermined value; And the feature off-set value in this candidate's laser spots/zone does not meet particular value.

According to a further aspect in the invention, provide a kind of laser spots/regional detection device, can comprise: standard laser point/provincial characteristics value is obtained parts, is used to obtain the eigenwert in standard laser point/zone; Candidate's laser spots/regional production part is used for from laser spots/zone of input picture generation candidate; The off-set value calculating unit, the eigenwert in laser spots/zone that is used to calculate each candidate is with respect to the off-set value of the eigenwert in standard laser point/zone; Parts are determined in laser spots/zone, are used to determine to possess the candidate's laser spots/zone of minimal characteristic off-set value as laser spots/zone.

This laser spots/method for detecting area and device can support polymorphic type laser spots/zone to detect, and have better environmental suitability.

Description of drawings

Fig. 1 is the overall flow figure of laser spots/method for detecting area according to an embodiment of the invention;

Fig. 2 is the process flow diagram of laser spots/method for detecting area according to another embodiment of the present invention;

Fig. 3 is the synoptic diagram of the required element of type definition, characterizing definition and the standard laser point/provincial characteristics value initialization in laser spots/zone according to an embodiment of the invention;

Fig. 4 is for supporting the process flow diagram of the method that polymorphic type laser spots/zone is detected according to an embodiment of the invention;

Fig. 5 shows the instantiation of following laser spots/area attribute: COLOR COMPOSITION THROUGH DISTRIBUTION information, size and shape information, position and motion track information;

Fig. 6 and Fig. 7 show two instantiations of laser spots/area type definition respectively;

Fig. 8 is the schematic block diagram of laser spots pick-up unit according to an embodiment of the invention; And

Fig. 9 can put into practice exemplary computer system of the present invention according to an embodiment of the invention.

Embodiment

In order to make those skilled in the art understand the present invention better, the present invention is described in further detail below in conjunction with the drawings and specific embodiments.

Before specifically describing, overall introduction thought of the present invention once at first: the present invention has illustrated a kind of method that can support that polymorphic type laser spots/zone is detected, and this method is intended to increase environmental suitability and minimizing false detection rate under the noise spot situation.The method is based on following supposition: compare with noise spot/zone, laser spots/zone shows more as " laser spots/zone " (being called standard laser point/zone); Under equivalent environment, laser spots/area attribute performance is relatively stable.At first, the definition or estimate that all types of standard laser point/provincial characteristics values is so that do next step detection, for example, the several leading two field picture that only comprises a certain type laser spots/zone after system start-up can be regarded as initial data set, can train or estimates to obtain standard laser point/provincial characteristics value based on this data set.Then, every frame is needed detected image: at first carry out pre-service, for example image is filtered, produce candidate's laser spots/zone (type is distinguished) with some certain threshold level such as luminance threshold; Secondly, according to certain class standard laser spots/provincial characteristics value of having estimated, be the laser spots/regional calculated characteristics off-set value of each generation of respective classes; The 3rd, to each laser spots/area type, the luminous point/zone with minimal characteristic off-set value is detected as the laser spots/zone of this type.In addition, can also upgrade or merge standard laser point/provincial characteristics value that detected laser spots/area characteristic information has been estimated to respective classes, detect for use in the laser spots/zone of next frame image.

Fig. 1 is the overall flow figure of laser spots/method for detecting area 100 according to an embodiment of the invention.

As shown in Figure 1, laser spots/method for detecting area 100 can comprise: the eigenwert obtaining step S110 in standard laser point/zone; Candidate's laser spots/zone generates step S120; Off-set value calculation procedure S130; Laser spots/regional determining step S140.

At step S110, obtain the eigenwert in standard laser point/zone.

Laser spots/zone of this paper is meant the point/regional various piece that has laser in the image, usually, when the part size that has laser hour, be referred to as laser spots habitually, and, then generally be referred to as laser region when its size greatly the time.Laser spots/zone can utilize various attributes or feature to describe.As the example of attribute or feature, color, size and shape, position and motion track etc. are arranged.Generally speaking, with regard to color character, in brightness, laser region is generally bright than the background area, and on color, under the situation of using the red laser pen, laser region is generally red than the background area; With regard to size and shape facility, laser region may show as general so-called laser spot, wire or other shape; With regard to position and motion track, the laser spots motion track of sequential frame image is generally relative smooth track.

So-called standard laser point/zone herein is to think the laser spots/zone of the benchmark that can be used as laser spots/zone, can may be that candidate's laser spots/zone and this standard laser point/zone in laser spots/zone compares by thinking, judge whether candidate laser spots/zone is laser spots/zone.

Supposing that a laser spots/zone utilizes these features of color, size and shape, position and motion track to describe, can be vectorial F=(C, S with the characterizing definition in a laser spots/zone, P), C is a color information herein, and S is size and shape information, and P is position and motion track information.Color information C is for example by R, G under the RGB model case, every value of B, the descriptions such as every value of the H under the HSV model case, S, V.Size and shape information S for example can utilize the pixel quantity that comprises, given shape as descriptions such as circles/square.Position and motion track information for example can be utilized in coordinate figure, the front and back multiple image coordinate, the direction of motion in laser spots/zone to wait and describe.

The eigenwert in standard laser point/zone can obtain by number of ways:

1, experience presets method: according to user's experience, the eigenwert in standard laser point/zone is defined as predetermined fixed value;

2, real-time coaching method: after laser spots/method for detecting area starts, with the laser pen instruction screen, obtain the n frame image data that camera is taken, this n two field picture is regarded as primary data, therefrom obtain the eigenwert of the mean value of for example color, size and shape attribute as standard laser point/zone, at this moment the position in standard laser point/zone and motion track attribute can only comprise laser spots/regional location and motion track information in former two field pictures;

3, coaching method in advance: mass data is prepared in advance, adopt for example support vector machine (Support Vector Machine of some data training method, SVM) eigenwert of this mass data is trained, the eigenwert that training obtains is stored, after laser spots/method for detecting area brings into operation, can read the eigenwert of being stored eigenwert as standard laser point/zone.

In above-mentioned three kinds of methods enumerating, real-time coaching method is because the environment attribute of the environment attribute warp of a n frame image data of being trained and follow-up input picture is similar, so environmental suitability is best; And coaching method needs relatively large data in advance, and environmental suitability still because the eigenwert in standard laser point/zone has been ready in advance, has therefore been saved the time of laser spots/method for detecting area not as real-time coaching method.Those skilled in the art can determine to choose the settle the standard eigenwert in laser spots/zone of which kind of method in practice according to the requirement to environmental suitability and time etc.

Need to prove that the acquisition methods of the eigenwert in above-mentioned standard laser point/zone only is an example, the present invention is not limited thereto, and any method that can obtain the representative property value in laser spots/zone can be applied to the present invention.

At step S120, from input picture, produce laser spots/zone of candidate.

Can be by two field picture being filtered the laser spots/zone that generates the candidate with the predetermined characteristic threshold value.

The predetermined characteristic threshold value can be the threshold value at some feature or combination of features generated; Perhaps by the particular value of the particular community that the user formulated.For example, the predetermined characteristic threshold value can for

(1) intensity threshold-for example, if the color attribute of image utilizes HSV model description well known in the art, V wherein is the abbreviation of " value ", expression intensity, the threshold value of V can be made as 110, certain intensity will be laser spots/zone of candidate greater than the laser spots/zone of this threshold value 110; Or

(2) the colourity threshold value-for example, under the situation that adopts the HSV model, H is the abbreviation of " hue " [colourity], at the threshold value of H, is made as 0～30 and 330～360 for red laser spots, is made as 40～80 for green laser spots;

(3) shape threshold value-laser spots is shaped as circle, square or pentagon;

But above-mentioned predetermined characteristic threshold value only is an example, can also comprise size threshold value, saturation degree threshold value or the like.

At step S130, the eigenwert in laser spots/zone of calculating each candidate is with respect to the off-set value of the eigenwert in standard laser point/zone.

Calculating the eigenwert in laser spots/zone of each candidate can specific as followsly carry out with respect to the off-set value of the eigenwert in standard laser point/zone:

The color character value that the color character value that comprises based on candidate laser spots/zone and standard laser point/zone have been estimated is calculated the color character off-set value;

Size that comprises based on candidate laser spots/zone and shape facility value and size and shape facility value that the standard laser point/zone has been estimated are calculated size and shape facility off-set value;

Position that comprises based on candidate laser spots/zone and motion track eigenwert and position and motion track eigenvalue calculation position and motion track feature off-set value that the standard laser point/zone has been estimated;

For color, size and shape, position and motion track feature off-set value are provided with weighting factor respectively;

The feature off-set value in candidate's laser spots/zone is following three value sums: the color character off-set value multiply by the respective weight factor of setting, size and shape facility off-set value multiply by the respective weight factor of setting, and position and motion track feature off-set value multiply by the respective weight factor of setting.

At step S140, determine to possess the candidate's laser spots/zone of minimal characteristic off-set value as laser spots/zone.

Fig. 2 is the process flow diagram of laser spots/method for detecting area 200 according to another embodiment of the present invention.

Step S210, S220 in laser spots/method for detecting area 200 shown in Figure 2, S230, S240 are identical with step S110, S120, S130, S140 in laser spots/method for detecting area 100 shown in Figure 1, here it are no longer described.

The step of updating S250 of the eigenwert that laser spots/method for detecting area 200 shown in Figure 2 and laser spots/method for detecting area 100 shown in Figure 1 different are standard laser point/zone many.

In step S240, obtained after new detected laser spots/zone, can be incorporated into detected laser spots/provincial characteristics property value in standard laser point/provincial characteristics property value, thus the eigenwert in the laser spots/zone after obtaining upgrading.In the laser spots/zone of next frame image was detected, the standard laser point/provincial characteristics value after this renewal was used for new laser spots/zone as benchmark and detects.

By upgrading the eigenwert in standard laser point/zone with the new detected laser spots/eigenwert in zone, make the eigenwert in standard laser point/zone along with the time upgrades, have better environmental suitability.

Be a whole method in common in conjunction with Fig. 1 and the described laser spots/method for detecting area of Fig. 2 above.

The present invention can also support polytype laser spots/zone to detect, described below with reference to Fig. 3, Fig. 4, Fig. 3 is the synoptic diagram of the required element of type definition, characterizing definition and the standard laser point/provincial characteristics value initialization in laser spots/zone according to an embodiment of the invention.Fig. 4 is a method of supporting that according to an embodiment of the invention polymorphic type laser spots/zone is detected.

As shown in Figure 3, the feature 113 in a luminous point/zone can be defined as vectorial F=(P), C is a color information 114 herein for C, S, S be the size and shape information 115, P is position and motion track information 116.

5a among Fig. 5 has carried out indicative icon to 5d respectively to these features.The color (redness) that Fig. 5 a shows a laser spots/zone distributes.Fig. 5 b shows the size in a laser spots/zone, represents with the number of pixels in this laser spots/zone.Fig. 5 c shows the shape in laser spots/zone, is linear among the figure.Fig. 5 d shows the movement locus in laser spots/zone, and in this example, the laser spots movement locus in several two field pictures is level and smooth relatively curvilinear path.

The present invention can support the detection in polytype laser spots/zone 112, the set of laser spots/area type can be represented with Lts, can be expressed as Wherein l is the number of type, Lt _iRepresent i class laser spots/area L t _i，

Illustrate polymorphic type laser spots/zone below.

According to color information, the laser spots type may be defined as red and green two classes, as shown in Figure 6.

According to shape information, laser spots/area type may be defined as circle, square and five jiaos of types (can realize by load diffraction optical element on laser pen), as shown in Figure 7.

In an example of the present invention, the number of laser spots/area type and classification pre-define, and do not have overlapping between each type.But, certainly, can have overlapping between each type, the situation that a laser spots/zone is divided into a plurality of types at this moment may occur, technical scheme at this moment of the present invention stands good.

To each laser spots/area type, the proper vector F in standard laser point/zone _{S (Lt)}Can be described as

$F_{S\left(\mathrm{Lt}\right)}=(C_{S\left(\mathrm{Lt}\right)},S_{S\left(\mathrm{Lt}\right)},P_{S\left(\mathrm{Lt}\right)}),\mathrm{Lt}\∈{\left\{{\mathrm{Lt}}_i\right\}}_{i=1}^l---\left(1\right)$

As shown in Figure 4,, obtain the eigenwert of all kinds of standard areas at step S310, thus obtain eigenwert 301, type 2 standard areas of Class1 standard area eigenwert 302 ... the eigenwert 303 of type l standard area.

With preamble with reference to the class of operation of figure 1 described step S110 seemingly, can take equally experience preset method, in real time coaching method, coaching method obtains the eigenwert in all kinds of standard laser points/zone in advance.

Particularly, under the situation that adopts real-time coaching method, after laser spots/method for detecting area starts, for every kind of laser spots/area type Lt, laser pen instruction screen with corresponding such laser spots/zone, obtain the n frame image data that camera is taken, this n two field picture is regarded as primary data, therefrom obtain for example color, the weighted mean value of size and shape attribute is as the eigenwert in standard laser point/zone, and at this moment the position in standard laser point/zone and motion track attribute can only comprise laser spots/regional location and motion track information in former two field pictures.

If the primary data of laser spots/area type Lt is with D _{Init (Lt)}Expression, then D _{Init (Lt)}Can be defined as the set of the proper vector of this n two field picture by following formula (2):

$D_{\mathrm{Init}\left(\mathrm{Lt}\right)}=\left\{F_{\mathrm{Init}(\mathrm{Lt},i)}\right|F_{\mathrm{Init}(\mathrm{Lt},i)}=(C_{\mathrm{Init}(\mathrm{Lt},i)},S_{\mathrm{Init}(\mathrm{Lt},i)},P_{\mathrm{Init}(\mathrm{Lt},i)}){\}}_{i=1}^n,n\≥1,\mathrm{Lt}\∈{\left\{{\mathrm{Lt}}_i\right\}}_{i=1}^l---\left(2\right)$

F wherein _{Init (Lt, i)}The proper vector of representing the laser spots/zone in the i two field picture, C _{Init (Lt, i)}Represent color information, S in the proper vector in the laser spots/zone in the i two field picture _{Init (Lt, i)}Represent size in the proper vector in the laser spots/zone in the i two field picture and shape information, P _{Init (Lt, i)}Represent position and motion track information in the proper vector in the laser spots/zone in the i two field picture.

Wherein under the situation of training, preferably each two field picture comprises and only comprises a Lt class laser spots/zone (noiseless point).

For example, to each laser spots/area type, primary data can obtain by following operation: with laser pen (state that keeps laser to light) statically point operation shield, the running time is at least n/FPS second, FPS represents the frame per second of camera herein.

At the primary data D that has obtained formula (2) expression _{Init (Lt)}After, as example, proper vector F that can following estimation type Lt standard laser point/zone _{S (Lt)}

To F _{S (Lt)}Each proper vector component, V _{S (Lt)}=C _{S (Lt)}∨ S _{S (Lt)}∨ P _{S (Lt)}:

Wherein, with formula (2) correspondence, n is the frame number in the primary data, V _{S (Lt)}Corresponding one by one with V, i.e. V _{S (Lt)}Be C _{S (Lt)}The time V should be C _Init, V _{S (Lt)}Be S _{S (Lt)}The time V should be S _Init, V _{S (Lt)}Be P _{S (Lt)}The time V should be P _Init

Component C for proper vector _{S (Lt)}Or S _{S (Lt)}, utilize first minor in the formula (3) to try to achieve, the proper vector property value of primary data is asked weighted mean, α respectively _iRepresentation feature vector sample V _{(Lt, i)}Weight, preferred, its span be 0 to 1 and

Component P for proper vector _{S (Lt)}Usually with regard to the time, current processed frame image possessed bigger reference significance apart from the laser spots/regional location trace information in the nearest several two field pictures of current processed frame, when therefore the individual features in initialization standard laser point/zone is vectorial, the time of can writing down in the primary data is gone up in the later image laser spots/regional location trace information in order to the reference of follow-up several leading frame processed frame, second minor in the formula (3) is promptly expressed this implication, wherein t represents the t frame in the n frame, will choose the reference as subsequent treatment of position from the t frame to the n frame and motion track information.

Adopting under the situation of coaching method in advance, laser spots/zone for every type, mass data is prepared in advance, adopt for example support vector machine (Support Vector Machine of some data training method, SVM) eigenwert of this mass data is trained, the eigenwert storage that training is obtained after laser spots/method for detecting area brings into operation, can be read all kinds of eigenwert of the being stored eigenwert as all kinds of standard laser points/zone.

Turn back to Fig. 4, at step S320, utilize with this predetermined number laser spots/area type in every type of characteristic threshold value that is associated, generation is divided into the laser spots/zone of the candidate in predetermined number laser spots/area type, thus obtain candidate's laser spots/zone 304, the type 2 of Class1 candidate's laser spots/zone 305 ..., candidate's laser spots/zone 306 of type l.

At candidate laser spots/zone of each type l,, can carry out image filtering with the certain threshold level of the type l, the set D in candidate's laser spots/zone of the type l that is generated in order to obtain being divided into candidate laser spots/zone of the type l _{P (Lt)}Be expressed as follows

Wherein m is illustrated in the number that type is candidate's laser spots/zone of Lt; F _{P (Lt, i)}The proper vector of representing candidate's laser spots/zone of i type Lt; C _{P (Lt, i)}, S _{P (Lt, i)}, P _{P (Lt, i)}The proper vector F that represents candidate's laser spots/zone of i type Lt respectively _{P (Lt, i)}Color information, size and shape information, position and motion track information.

The certain threshold level of type l can be that the user is predefined, or from the acquistion of mass data middle school to.For example, certain threshold level can for:

Intensity threshold-HSV model, " value " [intensity] threshold value is made as 110; Or

Colourity threshold value-HSV model, " hue " [colourity] threshold value is that red laser spots is made as 0～30 and 330～360, for green laser spots is made as 40～80;

Shape threshold value-laser spots is shaped as circle, square or pentagon.

Accordingly, the example of all types of candidate's laser spots/zone generations is as follows:

Laser spots type-redness, green: it be the red type laser spots of candidate in 0～30 or 330～360 scopes that intensity level surpasses intensity threshold 110 and chromatic value, and intensity level is candidate's green type laser spots above intensity threshold～110 and chromatic value in 40～80 scopes;

Laser spots type-circle, square and pentagon: what intensity level surpassed intensity threshold 110 and was shaped as circle is the candidate circles laser spots, intensity level surpasses intensity threshold 110 and is shaped as the square square laser spots of candidate that is, intensity level surpasses intensity threshold 110 and is shaped as being of five jiaos of types five jiaos of type laser spots of candidate (can finish the shape threshold filtering by predefined standard shape type and light spot form are mated);

Laser spots type-redness is circular, red square, red pentagon, green circular, green box, green pentagon, intensity level surpasses intensity threshold 110, chromatic value is in 0～30 or 330～360 scopes and be shaped as the circular red circular laser spots of candidate that is, intensity level surpasses intensity threshold 110, chromatic value is in 0～30 or 330～360 scopes and be shaped as the square red square laser spots of candidate that is, intensity level surpasses intensity threshold 110, chromatic value is in 0～30 or 330～360 scopes and be shaped as the pentagonal red pentagon laser spots of candidate that is, intensity level surpasses intensity threshold 110, chromatic value is in 40～80 scopes and be shaped as the circular green circular laser spots of candidate that is, intensity level surpasses intensity threshold 110, chromatic value is in 40～80 scopes and be shaped as square candidate's green box laser spots that is, intensity level surpasses intensity threshold 110, chromatic value is in 40～80 scopes and be shaped as pentagonal be the green pentagon laser spots of candidate (can finish the shape threshold filtering by predefined standard shape type and light spot form are mated).

The pre-defined of above-mentioned laser spots/area type only is a kind of example, and those skilled in the art can define various laser spots/area types according to the design needs.

Need to prove,, can have following situation,, do not have the candidate laser spots/zone of the type for certain laser spots/area type in candidate's laser spots/area dividing step.

At step S330, for each the candidate laser spots/zone in each type, calculate this candidate laser spots/zone eigenwert with respect to the off-set value of the eigenwert in the standard laser point/zone of the type.

Can utilize following formula (5) to calculate each candidate's laser spots zone F of each type Lt _{P (Lt, i)}Off-set value.

$\∀F_{P(\mathrm{Lt},i)}\∈D_{P\left(\mathrm{Lt}\right)},1\≤i\≤m,\mathrm{Lt}\∈{\left\{{\mathrm{Lt}}_i\right\}}_{i=1}^l,$

DV _(Lt，i)＝αDV _C(Lt，i)+βDV _S(Lt，i)+δDV _P(Lt，i)，α，β，δ≥0，α+β+δ＞0，

DV wherein _{C (Lt, i)}=f (C _{P (Lt, i)}, C _{S (Lt)}), DV _{S (Lt, i)}=g (S _{P (Lt, i)}, S _{S (Lt)}), (5)

DV _P(Lt，i)＝h(P _P(Lt，i)，P _S(Lt))，

Wherein, D _{P (Lt)}Be the set in candidate's laser spots/zone of type Lt, D _{P (Lt, i)}Be candidate laser spots/zone of i type Lt, m is the number in candidate's laser spots/zone of type Lt, DV _{(Lt, i)}The proper vector (eigenwert) in candidate's laser spots/zone that is i type Lt is with respect to the skew of the proper vector (eigenwert) in standard laser point/zone of type Lt, DV _{C (Lt, i)}, DV _{S (Lt, i)}DV _{P (Lt, i)}The proper vector (eigenwert) in candidate's laser spots/zone that is i type Lt respectively is with respect to color offset, size and the shape skew of the proper vector (eigenwert) in standard laser point/zone of type Lt, position and motion track skew, α, β, δ are respectively the weight factor of color offset, size and shape skew, position and motion track skew, f is the color offset computing function, g is size and shape calculations of offset function, and h is position and motion track calculations of offset function C _{P (Lt, i)}, S _{P (Lt, i)}, P _{P (Lt, i)}The proper vector F that represents candidate's laser spots/zone of i type Lt respectively _{P (Lt, i)}Color information, size and shape information, position and motion track information, C _{S (Lt)}, S _{S (Lt)}, P _{S (Lt)}The standard feature vector F that represents candidate's laser spots/zone of type Lt respectively _{S (Lt)}Color information, size and shape information, position and motion track information.

In step S340, with definite laser spots/zone, the laser spots/zone that possesses the minimal characteristic off-set value in all candidate laser spots/zones that are divided in the laser spots/area type as the type.

Particularly, based on formula (4) and (5),, for example can detect maximum laser spots/zones according to following strategy to each laser spots/area type.

$\∀\mathrm{Lt}\∈{\left\{{\mathrm{Lt}}_i\right\}}_{i=1}^l,\mathrm{if}{D}_{P\left(\mathrm{Lt}\right)}!=\mathrm{NULL},$

${\mathrm{DV}}_{(\mathrm{Lt},w)}=$

m≥1，1≤w≤m；

F _{P (Lt, w)}→ laser spots/zone.

Wherein, T _C, T _S, T _P, T represents color offset value threshold value, size and shape off-set value threshold value, position and motion track off-set value threshold value and total drift value threshold value respectively.

As shown in Equation (6), have only following candidate laser spots/zone just can be confirmed as laser spots/zone, DV _{C (Lt, i)}, DV _{S (Lt, i)}DV _{P (Lt, i)}Color offset, size and shape skew, position and motion track skew are smaller or equal to corresponding threshold T _C, T _S, T _P, and total drift DV _{(Lt, i)}Smaller or equal to the threshold value T of total drift value, and be through above-mentioned threshold value T _C, T _S, T _P, the candidate laser spots/zone of off-set value minimum in the candidate laser spots/zone after T filters.

But, need to prove, adopt the threshold value T of above-mentioned color offset, size and shape skew, position and motion track skew and total drift _C, T _S, T _P, it is optional and nonessential step that T filters.And, except adopting above-mentioned offset threshold T _C, T _S, T _P, outside T filters, also can adopt other threshold values to filter.For example, under the situation that laser spots/zone is divided into circle, square, five jiaos of types, can filter with one in luminance threshold, saturation degree threshold value, the RGB threshold value etc. or the multinomial candidate laser spots/zone of carrying out, think that promptly not meeting candidate laser spots/zone that threshold value requires is noise spot/zone.And the step in this filtration noise spot/zone is not to carry out after obtaining candidate's laser spots/zone, but all can carry out at any time as required.

Utilize the processing shown in the formula (6), each off-set value and/or satisfy in all luminous point/zones of threshold setting, the candidate laser spots/zone with minimal characteristic off-set value is detected as laser spots/zone.In other words, has feature F _{P (Lt, w)}Candidate laser spots/zone be detected as the laser spots/zone of Lt type.

As previously mentioned, alternatively, to each laser spots/area type Lt, can be as shown in Equation (7) detected laser spots/region D _{P (Lt)}Characteristic attribute value F _{P (Lt, w)}Be incorporated into the characteristic attribute value f in standard laser point/zone _{S (Lt)}In.

$\∀\mathrm{Lt}\∈{\left\{{\mathrm{Lt}}_i\right\}}_{i=1}^l\mathrm{if}{D}_{P\left(\mathrm{Lt}\right)}!=\mathrm{NULL},$

F′ _S(Lt)＝k(F _P(Lt，w)，F _S(Lt))，1≤w≤m (7)

Wherein k is the proper vector pooled function.About the form of proper vector pooled function K, for color attribute value or size shape property value, proper vector pooled function K can be F _{P (Lt, w)}And F _{S (Lt)}Weighted mean, as (F _{P (Lt, w)}+ F _{S (Lt)})/2, formula in the given example (8) etc. hereinafter.

For position motion track attribute, the proper vector pooled function can be for being updated to detected laser spots/zone position information in the position motion track community set in standard laser point/zone as latest position, can consider simultaneously in the set laser spots/the zone position information of old position remove.

In the laser spots/zone of next frame image was detected, the standard laser point after the renewal/provincial characteristics value was estimated F ' _{S (Lt)}Alternative former eigenwert F _{S (Lt)}, be used for laser spots/zone and detect.

Above-mentionedly can from a two field picture, detect polytype laser spots/zone, widen the range of application of this laser spots/method for detecting area with reference to figure 3, laser spots/method for detecting area shown in Figure 4.

Following content is a specific implementation example.

As shown in Figure 6, laser spots/area type is defined as { Red, Green}.To each Red (redness) or Green (green) laser spots/area type, n=25 two field picture is regarded as primary data (primary data number of image frames by repeatedly experimental check obtain), and wherein each two field picture comprises and only comprise a Red or Green class laser spots/zone (noiseless point).

This realizes that example uses HSV model and RGB model, and the attribute that proper vector F comprises is as follows:

C={avgH, avgS, avgV, RGPercent), and S={pixelNum}, P={}, wherein

--avgH, avgS and avgV represent average " hue (colourity) " of luminous point/all pixels of zone,

" saturation (saturation degree) " and " value (intensity) " value (HSV model);

--RGPercent represent Red laser spots/area type Red pixel quantity number percent or

The Green pixel quantity number percent (RGB model) of Green laser spots/area type;

--pixelNum represents the total pixel number amount in luminous point/zone.

To each eigenvalue calculation mean value of primary data, the mean value assignment of gained is given the eigenwert in defined standard laser point/zone in the formula (1).In other words, each the α value in the formula (3) is 1/25 by composing, and the P feature is not considered at this.

The described off-set value computing function of formula (5) is achieved as follows

$f:{\mathrm{DV}}_C=2*\frac{|\mathrm{avgH}-\mathrm{avg}{H}_S|}{{\mathrm{avgH}}_S}+10*\frac{|\mathrm{avgS}-\mathrm{avg}{S}_S|}{{\mathrm{avgS}}_S}+$

$30*\frac{|\mathrm{avgV}-\mathrm{avg}{V}_S|}{{\mathrm{avgV}}_S}+5*\frac{|\mathrm{RGPercent}-\mathrm{RGP}{\mathrm{ercent}}_S|}{{\mathrm{RGPercent}}_S}$

$g:{\mathrm{DV}}_S=\mathrm{pixelNum}>{\mathrm{pixelNum}}_S?0:\frac{({\mathrm{pixelNum}}_S-\mathrm{pixelNum})}{{\mathrm{pixelNum}}_S}$

h：DV _P＝0

DV＝DV _C+10*DV _S

This shows that α, β and δ value described in the formula (5) correspondingly are made as 1,10 and 0 respectively.

Characteristic attribute threshold value described in the formula (6) is set as some concrete numerical value equally, and for example, pixel count off-set value threshold value is set as 0.5.

After detecting laser spots/zone, corresponding characteristic attribute value can be incorporated in standard laser point/provincial characteristics property value.(7) described proper vector value pooled function realizes in this example as follows in the formula.

$k:\∀V\∈\{\mathrm{avgH},\mathrm{avgS},\mathrm{avgV},\mathrm{RGPercent},\mathrm{pixelNum}\},$

$V_S=\frac{(n-1)*V_S+V}{n},n=25---\left(8\right)$

This realizes example through repeatedly experiment and substantive test data detection and evaluation and test, and check and evaluation result show that compare with art methods, this method has promoted laser spots/regional verification and measurement ratio.

Fig. 8 shows the schematic block diagram of laser spots/regional detection device 800 according to an embodiment of the invention.This laser spots pick-up unit 800 can comprise: standard laser point/provincial characteristics value is obtained parts 810, is used to obtain the eigenwert in standard laser point/zone; Candidate's laser spots/regional production part 820 is used for from laser spots/zone of input picture generation candidate; Off-set value calculating unit 830, the eigenwert in laser spots/zone that is used to calculate each candidate is with respect to the off-set value of the eigenwert in standard laser point/zone; Parts 840 are determined in laser spots/zone, are used to determine to possess the candidate's laser spots/zone of minimal characteristic off-set value as laser spots/zone.

Described laser spots/regional detection device 800 can also comprise standard laser point/provincial characteristics value updating component, is used for upgrading based on the characteristic information in determined laser spots/zone the eigenwert in standard laser point/zone.

Candidate's laser spots/regional production part can filter to obtain laser spots/zone of candidate two field picture with the predetermined characteristic threshold value.

Described laser spots/regional detection device 800 can also the pre-defined predetermined number that does not overlap each other laser spots/area type, wherein, standard laser point/provincial characteristics value is obtained the feature that parts 810 can obtain each standard laser point/zone of this predetermined number type respectively; Candidate's laser spots/regional production part 820 can utilize with this predetermined number laser spots/area type in every type of characteristic threshold value that is associated, produce the laser spots/zone be divided into the candidate in predetermined number laser spots/area type; Off-set value calculating unit 830 can calculate this candidate laser spots/zone eigenwert with respect to laser spots/zone of this candidate under the off-set value of eigenwert in standard laser point/zone of the corresponding type of type; Laser spots/zone determines that parts 840 can be with the definite laser spots/zone as the type, the candidate's laser spots/zone that possesses the minimal characteristic off-set value that is divided in laser spots/zone of all candidates in laser spots/area type.

The eigenwert that candidate's laser spots/regional production part 820 obtains standard laser point/zone can comprise at least a in the following operation: the fixed value that the eigenwert in standard laser point/zone is defined as being scheduled to; To every kind of laser spots/area type, the several leading two field picture that comprises respective type laser spots/zone behind the described laser spots/method for detecting area of startup is considered as primary data, obtain the eigenwert in the standard laser point/zone of the type based on this primary data training; From training the eigenwert in every type the standard laser point/zone that obtains based on a large amount of primary datas at every kind of laser spots/area type in advance, obtain the eigenwert in the standard laser point/zone of the type.

Described predetermined characteristic threshold value can comprise one or more in each color attribute threshold value.

Described feature can comprise with the next item down or multinomial: color; Size and shape; And position and motion track.

The eigenwert that off-set value calculating unit 830 calculates laser spots/zone of each candidate can comprise with respect to the off-set value of the eigenwert in standard laser point/zone: the color character value calculating color character off-set value that the color character value that comprises based on candidate laser spots/zone and standard laser point/zone have been estimated; Size that comprises based on candidate laser spots/zone and shape facility value and size and shape facility value that the standard laser point/zone has been estimated are calculated size and shape facility off-set value; Position that comprises based on candidate laser spots/zone and motion track eigenwert and position and motion track eigenvalue calculation position and motion track feature off-set value that the standard laser point/zone has been estimated; For color, size and shape, position and motion track feature off-set value are provided with weighting factor respectively; The feature off-set value in candidate's laser spots/zone is following three value sums: the color character off-set value multiply by the respective weight factor of setting, size and shape facility off-set value multiply by the respective weight factor of setting, and position and motion track feature off-set value multiply by the respective weight factor of setting.

If any during candidate laser spots/zone meets the following conditions then can be considered as noise spot/zone with it, and abandon: the predetermined characteristic property value does not meet predetermined value or preset range; The color character off-set value is greater than predetermined value; Size and shape facility off-set value are greater than predetermined value; Position and motion track feature off-set value are greater than predetermined value; The feature off-set value in this candidate's laser spots/zone is greater than particular value.

Need to prove, aspect hardware device, this method also can use specific camera head or extra filter consequently to reduce some environmental interference points, thereby it is more accurate effective that laser spots/zone is detected, for example, the adjustable camera of aperture can be used to reduce two field picture brightness and improve contrast.

Fig. 9 can put into practice exemplary computer system of the present invention according to an embodiment of the invention.

As shown in Figure 9, a computer system 10 that detects at laser spots/zone can comprise main frame 11, keyboard 16, monitor 17, printer 18, floppy disk 19, network insertion device 20 and hard disk drive 21.Main frame 11 comprises data bus 12, random access memory (RAM) 13, ROM (read-only memory) (ROM) 14, center processor 15 and peripheral bus 23.

According to the instruction of from random access memory 13, receiving, the reception and the processing of center processor 15 control input data, and output to monitor 17 or other peripheral hardwares.In the present embodiment, a function of center processor 15 is to obtain the eigenwert in standard laser point/zone, handle the image of input, produce laser spots/zone of candidate, the eigenwert in calculated candidate laser spots/zone determines to possess the candidate's laser spots/zone of minimal characteristic off-set value as laser spots/zone with respect to the off-set value of the eigenwert in standard laser point/zone.Detected laser spots/zone can be used for other application programs.

Center processor 15 visits random access memory 13 and ROM (read-only memory) 14 by data bus 12.Random access memory 13 as a read-write internal memory, is used as the workspace and the variable data memory block of each process by center processor 15, also can be used for storing input picture as the image from camera 22, the laser region of detection, laser spots etc.ROM (read-only memory) 14 is stored for example laser spots/region detector, and other application programs etc.

Peripheral bus 23 is used to visit the input that links to each other with computing machine 11, peripherals such as output and storage.In the present embodiment, these peripherals comprise monitor 17, printer 18, floppy disk 19, network insertion device 20, hard disk drive 21, picture pick-up device 22.Monitor 17 comes the data and the image of display centre processor 15 outputs by peripheral bus 23.It can be grid type display device such as CRT or LCD display.Printer 18 arrives paper or the media similar with paper to the data of center processor input with image print.Floppy disk 19 and hard disk drive 21 are used for storing input picture as the image from camera 22, the laser spots/zone of detection.By floppy disk 19, image can transmit between various computing machine system.Hard disk drive 21 storage spaces are bigger, and access speed is faster.Other memory device such as flash memory, also can be used for image for computer system 10 visits.Picture pick-up device 22 can be used for taking the screen picture that contains laser spots, handles for center processor 15, thus the laser region in the detected image.Computer system 10 sends data and the data that receive from other computer systems by network insertion device 20 on network.The user gives computer system 10 by keyboard 16 input instructions.

Ultimate principle of the present invention has below been described in conjunction with specific embodiments, but, it is to be noted, for those of ordinary skill in the art, can understand the whole or any steps or the parts of method and apparatus of the present invention, can be in the network of any calculation element (comprising processor, storage medium etc.) or calculation element, realized that with hardware, firmware, software or their combination this is that those of ordinary skills use their basic programming skill just can realize under the situation of having read explanation of the present invention.

Therefore, purpose of the present invention can also realize by program of operation or batch processing on any calculation element.Described calculation element can be known fexible unit.Therefore, purpose of the present invention also can be only by providing the program product that comprises the program code of realizing described method or device to realize.That is to say that such program product also constitutes the present invention, and the storage medium that stores such program product also constitutes the present invention.Obviously, described storage medium can be any storage medium that is developed in any known storage medium or future.

It is pointed out that also that in apparatus and method of the present invention obviously, each parts or each step can decompose and/or reconfigure.These decomposition and/or reconfigure and to be considered as equivalents of the present invention.And, carry out the step of above-mentioned series of processes and can order following the instructions naturally carry out in chronological order, but do not need necessarily to carry out according to time sequencing.Some step can walk abreast or carry out independently of one another, for example, is independence between the generation step in the obtaining step of standard laser point/provincial characteristics value and candidate's laser spots/zone, can walk abreast or carries out with any sequencing.

Above-mentioned embodiment does not constitute limiting the scope of the invention.Those skilled in the art should be understood that, depend on designing requirement and other factors, and various modifications, combination, sub-portfolio and alternative can take place.Any modification of being done within the spirit and principles in the present invention, be equal to and replace and improvement etc., all should be included within the protection domain of the present invention.

Claims (10)

1. laser spots/method for detecting area may further comprise the steps:

Obtain the eigenwert in standard laser point/zone;

From input picture, produce laser spots/zone of candidate;

The eigenwert in laser spots/zone of calculating each candidate is with respect to the off-set value of the eigenwert in standard laser point/zone; And

Determine to possess the candidate's laser spots/zone of minimal characteristic off-set value as laser spots/zone.

2. according to the laser spots/method for detecting area of claim 1, also comprise:

Upgrade the eigenwert in standard laser point/zone based on the characteristic information in determined laser spots/zone.

3. according to the method for claim 1, the described laser spots/zone that produces the candidate from input picture comprises with the predetermined characteristic threshold value filters to obtain laser spots/zone of candidate two field picture.

4. according to the laser spots/method for detecting area of claim 1, also comprise, the pre-defined predetermined number that does not overlap each other laser spots/area type, wherein:

The described eigenwert of obtaining standard laser point/zone comprises, obtains the eigenwert in each standard laser point/zone of this predetermined number type respectively;

The described laser spots zone that produces the candidate from input picture comprises, utilize with this predetermined number laser spots/area type in every type of characteristic threshold value that is associated, generation is divided into the laser spots/zone of the candidate in predetermined number laser spots/area type;

The eigenwert in laser spots/zone of described each candidate of calculating comprises with respect to the off-set value of the eigenwert in standard laser point/zone, calculate this candidate laser spots/zone eigenwert with respect to laser spots/zone of this candidate under the off-set value of eigenwert in standard laser point/zone of the corresponding type of type;

Describedly determine that the laser spots/zone that possesses the minimal characteristic off-set value comprises as laser spots/zone, determine laser spots/zone as the type with being divided into the laser spots/zone that possesses the minimal characteristic off-set value in laser spots/zone of all candidates in laser spots/area type.

5. according to the laser spots/method for detecting area of claim 1, the wherein said eigenwert of obtaining standard laser point/zone comprises at least a in the following operation:

The fixed value that the eigenwert in standard laser point/zone is defined as being scheduled to;

To every kind of laser spots/area type, the several leading two field picture that comprises respective type laser spots/zone behind the described laser spots/method for detecting area of startup is considered as primary data, obtain the eigenwert in the standard laser point/zone of the type based on this primary data training;

From training the eigenwert in every type the standard laser point/zone that obtains based on a large amount of primary datas at every kind of laser spots/area type in advance, obtain the eigenwert in the standard laser point/zone of the type.

6. according to the method for claim 3, wherein said predetermined characteristic threshold value comprises one or more in each color attribute threshold value.

7. according to the method for claim 1, described feature comprises with the next item down or multinomial:

Color;

Size and shape; And

Position and motion track.

8. according to the method for claim 1, the eigenwert in laser spots/zone of described each candidate of calculating comprises with respect to the off-set value of the eigenwert in standard laser point/zone:

The color character value that the color character value that comprises based on candidate laser spots/zone and standard laser point/zone have been estimated is calculated the color character off-set value;

Size that comprises based on candidate laser spots/zone and shape facility value and size and shape facility value that the standard laser point/zone has been estimated are calculated size and shape facility off-set value;

Position that comprises based on candidate laser spots/zone and motion track eigenwert and position and motion track eigenvalue calculation position and motion track feature off-set value that the standard laser point/zone has been estimated;

For color, size and shape, position and motion track feature off-set value are provided with weighting factor respectively;

The feature off-set value in candidate's laser spots/zone is following three value sums: the color character off-set value multiply by the respective weight factor of setting, size and shape facility off-set value multiply by the respective weight factor of setting, and position and motion track feature off-set value multiply by the respective weight factor of setting.

9. method according to Claim 8, if candidate laser spots/zone any in meeting the following conditions then will be regarded as noise spot/zone, and be dropped:

The predetermined characteristic property value does not meet predetermined value or preset range;

The color character off-set value is greater than predetermined value;

Size and shape facility off-set value are greater than predetermined value;

Position and motion track feature off-set value are greater than predetermined value;

The feature off-set value in this candidate's laser spots/zone is greater than particular value.

10. laser spots/regional detection device comprises:

Standard laser point/provincial characteristics value is obtained parts, is used to obtain the eigenwert in standard laser point/zone;

Candidate's laser spots/regional production part is used for from laser spots/zone of input picture generation candidate;

The off-set value calculating unit, the eigenwert in laser spots/zone that is used to calculate each candidate is with respect to the off-set value of the eigenwert in standard laser point/zone;

Parts are determined in laser spots/zone, are used to determine to possess the candidate's laser spots/zone of minimal characteristic off-set value as laser spots/zone.

Images