Summary of the invention
In view of above content, be necessary to provide a kind of arrangement for detecting and method thereof, can not judge by accident with occurring again after the temporary transient disappearance of guaranteeing the object in the certain hour is different objects.
A kind of arrangement for detecting is used for detecting and record and appears at the interior object of a detection area, and described arrangement for detecting comprises:
One image input unit comprises:
One TOF video camera be used for to be taken the scene of described detection area and is obtained the distance between the every bit and TOF video camera in the scene; And
One image capture unit is used for obtaining the scene image signal of described TOF video camera shooting and the range information between the interior every bit of scene and the TOF video camera;
One threedimensional model is set up the unit, is used for setting up by the threedimensional model of the scene of detection area according to the range information between each point and the TOF video camera in the signal of video signal that obtains and the scene;
One object detecting unit is used for the object image of the threedimensional model of the scene that obtains is detected;
One feature extraction unit is used for the characteristic information according to the described object image of object image signal acquisition that receives;
One memory cell is used for storage and captures the characteristic information of each object image that arrives and the coding of corresponding each characteristic information;
One feature identifying unit, be used for to judge acquisition to the characteristic information of described object image the characteristic information with the characteristic information of the object image of described memory cell storage is identical, if judge that the characteristic information of acquisition is identical with a characteristic information of storage, described feature identifying unit sends the identical instruction of a result of determination; If the characteristic information of judgement acquisition and all characteristic informations of storage are all inequality, described feature identifying unit sends result of determination instruction inequality; And
One control unit, be used for after receiving the instruction inequality of described result of determination to acquisition to the characteristic information of object image give a new coding, and characteristic information one Preset Time of the new coding of storage and corresponding described new coding, and be used for after receiving the identical instruction of described result of determination with described acquisition to characteristic information substitute the identical characteristic information with it that is stored in the described memory cell, and keep original coding, control described memory cell simultaneously the characteristic information that substitutes is stored described Preset Time.
A kind of method for detecting may further comprise the steps:
A: by a TOF video camera distance parameter between the every bit and TOF video camera in the scene is taken and obtained to the scene of one detection area and obtain scene image signal that described TOF video camera takes and scene in range information between every bit and the TOF video camera;
B: set up by the threedimensional model of the scene of detection area according to the range information between each point in the signal of video signal that obtains and the scene and the TOF video camera;
C: the object image in the threedimensional model of the scene that obtains is detected, and output object signal of video signal;
D: the characteristic information of the described object image of object image signal acquisition that reception object signal of video signal and basis receive;
E: judge acquisition to the characteristic information of described object image whether identical with a characteristic information in the characteristic information of the object image of storage, if judge that the characteristic information of acquisition is identical with a characteristic information of storage, send the identical instruction of a result of determination, execution in step g; If the characteristic information of judgement acquisition and all characteristic informations of storage are all inequality, send result of determination instruction inequality, execution in step f;
F: to acquisition to the characteristic information of object image give a new coding, and store characteristic information one Preset Time of described new coding and corresponding described new coding; And
G: with described acquisition to characteristic information substitute the identical characteristic information with it of storage, and keep original coding, and control is stored described Preset Time to the characteristic information that substitutes.
Described arrangement for detecting and method obtain threedimensional model by scene image and the range information of described detection area that a TOF video camera is obtained, and detect to obtain the characteristic information of described object image to obtaining object image in the threedimensional model, encode by the characteristic information to described object image again, make same object image be endowed same coding, but and judge by the feature identifying unit whether whether its characteristic information that receives be same object with the object that a characteristic information identification of its storage is detected in a Preset Time section.
Embodiment
Below in conjunction with accompanying drawing and preferred embodiments the present invention is described in further detail:
Please refer to Fig. 1, arrangement for detecting 100 preferred embodiments of the present invention comprise an object detecting system 1 and a treatment system 2.Described object detecting system 1 is used for detecting and records the monitored object that appears at a detection area.Described treatment system 2 is connected to described object detecting system 1 so that the data in the described object detecting system 1 are handled.In the present embodiment, be illustrated as described monitored object with human-like, in other embodiments, described monitored object also can be other object that possesses special characteristic, as car plate etc.
Described object detecting system 1 comprises that an image input unit 11, an object detecting unit 12, a feature extraction unit 13, a memory cell 14, a control unit 15, a feature identifying unit 16 and a threedimensional model set up unit 17.
Please refer to Fig. 2, described image input unit 11 comprises a TOF(Time-of-flight) video camera 112 and an image capture unit such as image pick-up card 114.Described TOF video camera 112 is used for taking the scene of described detection area and obtains the distance between the every bit and TOF video camera 112 in the scene.Described image pick-up card 114 is used for obtaining scene image signal and the range information between the interior every bit of scene and the TOF video camera 112 of described TOF video camera 112 shootings and transfers to described threedimensional model sets up unit 17.
Described threedimensional model is set up unit 17 and is used for setting up by the threedimensional model of the scene of detection area according to the range information between each point and the TOF video camera 112 in the signal of video signal that obtains and the scene.
Described object detecting unit 12 is used for object image such as the human-like image of the threedimensional model of the scene that obtains are detected, and the human-like signal of video signal that will detect is sent to described feature extraction unit 13.
Described feature extraction unit 13 is used for capturing according to the human-like signal of video signal that receives the characteristic information of described human-like image, and described characteristic information comprises information such as the stature ratio of described human-like image and hair style.
Described memory cell 14 is used for storage and captures the characteristic information of each the human-like image that arrives and the coding of corresponding each characteristic information.
Described feature identifying unit 16 be used for judging acquisition to the characteristic information of described human-like image whether identical with the characteristic information of characteristic information of human-like images of storage in the described memory cell 14, when described feature extraction unit 13 captures the characteristic information of human-like image for the first time, do not store the characteristic information of human-like image in the described memory cell 14, described feature identifying unit 16 sends the different instruction of result of determination to described control unit 15.When storing the characteristic information of human-like image in the described memory cell 14, if described feature identifying unit 16 judges that the characteristic information of acquisition is identical with a characteristic information of storage, the people artificial same individual corresponding with the same characteristic features information of storage who then shows the characteristic information correspondence that acquisition is arrived, described feature identifying unit 16 sends the identical instruction of a result of determination to described control unit 15; If the characteristic information of judgement acquisition and all characteristic informations of storage are all inequality, the people people corresponding with the characteristic information of storage who then shows the characteristic information correspondence that acquisition is arrived is not same individual, and described feature identifying unit 16 sends result of determination instruction inequality to described control unit 15.
Described control unit 15 be used for after receiving the instruction inequality of described result of determination to acquisition to the characteristic information of human-like image give a new coding, and will described new coding and the characteristic information of the described new coding of correspondence be sent to described memory cell 14 and make 14 pairs of described characteristic informations of described memory cell store a Preset Time as 10 minutes.Described control unit 15 also be used for after receiving the identical instruction of described result of determination with described acquisition to characteristic information substitute the identical characteristic information with it that is stored in the described memory cell 14, and keep original coding, control 14 pairs of alternative characteristic informations of described memory cell simultaneously and store described Preset Time.So as long as a human-like image repeats in Preset Time in the zone that described object detecting system 1 is detected, described object detecting system 1 all can pick out described identical human-like image, and has avoided duplicate record.
Described treatment system 2 comprises a processing unit 22.Described processing unit 22 can obtain the coding that described memory cell 14 is stored, so that the coding that the user stores by 22 pairs of described memory cell 14 of described processing unit carries out subsequent analysis processing, as the quantity of statistical coding.
Described treatment system 2 also can comprise a coordinate storage unit 24, one target tracking unit 26 and a display unit 28, described feature extraction unit 13 is whenever given human-like image one coding, all send this coding and corresponding human-like time and the positional information that appears in the described detection area of encoding therewith to described treatment system 2, in the present embodiment, described human-like positional information is that the representation with coordinate sends described coordinate storage unit 24 to, when needs are followed the trail of human-like motion track, described target tracking unit 26 takes out the human-like coordinate information with same-code from described coordinate storage unit 24, and be presented on the described display unit 28 after respectively it being sorted with time sequencing, so that the user observes each human-like motion track, the monitoring objective that the user also can lock some as required continues to follow the trail of.
Please refer to Fig. 3, the preferred embodiments of method for detecting of the present invention is applied to the object detecting system 1 among Fig. 1, and described method for detecting may further comprise the steps:
Step S1: the distance parameter between the every bit and TOF video camera 112 in the scene is taken and obtained to the scene of the 112 pairs of described detection area of TOF video camera in the described image input unit 11, and the image pick-up card 114 in the described image input unit 11 obtains scene image signal and the range information between the interior every bit of scene and the TOF video camera 112 that described TOF video camera 112 takes and transfers to described threedimensional model sets up unit 17.
Step S2: described threedimensional model is set up unit 17 and is set up by the threedimensional model of the scene of detection area according to the range information between each point and the TOF video camera 112 in the signal of video signal that obtains and the scene.
Step S3: object image such as human-like image in the threedimensional model of 12 pairs of scenes that obtain of described object detecting unit are detected, and the human-like signal of video signal that will detect is sent to described feature extraction unit 13.
Step S4: the human-like signal of video signal that described feature extraction unit 13 bases receive captures the characteristic information of described human-like image, and described characteristic information comprises information such as described human-like image stature ratio and hair style.
Step S5: described feature identifying unit 16 judge acquisition to the characteristic information of described human-like image whether identical with a characteristic information in the characteristic information of human-like images of storage in the described memory cell 14, if judge that the characteristic information of acquisition is identical with a characteristic information of storage, described feature identifying unit sends the identical instruction of a result of determination to described control unit 15, execution in step S7; If judge that the characteristic information of acquisition is all different with all characteristic informations of storage, described feature identifying unit 16 sends result of determination instruction inequality to described control unit 15, execution in step S6.
Step S6: 15 pairs of acquisitions of described control unit to the characteristic information of human-like image give a new coding, and the characteristic information of described new coding and corresponding described coding sent to described memory cell 14 and make 14 pairs of described characteristic informations of described memory cell store described Preset Time, execution in step S8.
Step S7: described control unit 15 with described acquisition to characteristic information substitute the identical characteristic information with it of storage in the described memory cell 14, and keep original coding, and control 14 pairs of alternative characteristic informations of described memory cell and store described Preset Time.
Step S8: the coding that described treatment system 2 is obtained storage in the described memory cell 14 carries out subsequent treatment.
In step S8, the Unified coding that described treatment system 2 is stored by the described memory cell 14 of described processing unit 22 statistics is in the position in described zone that different time occurs.In step S6, described control unit 15 is also detected the positional information of described human-like signal of video signal in described detection area, and send the positional information in described detection area to described coordinate storage unit 24 with the representation of coordinate, after sorting with time sequencing, coordinate that described target tracking unit 26 is corresponding with each coding is presented on the described display unit 28.