add m

xuepengqiang

2020-05-26 bb5cb224c9abe4216aaa49a8287b06d9f05dab60

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
3  
€~Â]ã@sVdd
lm
Z

ddlZddlmZ
ddlmZ
ddlmZ
ddlm    Z    
Gdd    „d    ƒZ
dS)
é)
Úabsolute_importNé
)
Ú kalman_filter)
Úlinear_assignment)
Úiou_matching)
ÚTrackc@s:eZ
dZd
Zddd„
Zdd„Zd    d
„Zdd„Zd d„ZdS)ÚTrackeraà
    This is the multi-target tracker.
 
    Parameters
    ----------
    metric : nn_matching.NearestNeighborDistanceMetric
        A distance metric for measurement-to-track association.
    max_age : int
        Maximum number of missed misses before a track is deleted.
    n_init : int
        Number of consecutive detections before the track is confirmed. The
        track state is set to `Deleted` if a miss occurs within the first
        `n_init` frames.
 
    Attributes
    ----------
    metric : nn_matching.NearestNeighborDistanceMetric
        The distance metric used for measurement to track association.
    max_age : int
        Maximum number of missed misses before a track is deleted.
    n_init : int
        Number of frames that a track remains in initialization phase.
    kf : kalman_filter.KalmanFilter
        A Kalman filter to filter target trajectories in image space.
    tracks : List[Track]
        The list of active tracks at the current time step.
 
    çffffffæ?é,
écCs2|
|_||_
||_||_tjƒ|_g|_d
|_dS)Nr)    ÚmetricÚmax_iou_distanceÚmax_ageÚn_initrZKalmanFilterÚkfÚtracksÚ_next_id)Úselfrr rr©rúHD:\project\07_tracking_Re-id\04_s_yolo_tracking\lib\deep_sort\tracker.pyÚ__init__(s






zTracker.__init__c
Cs x|jD]}
|
j
|jƒ

qWd
S)z’Propagate track state distributions one time step forward.
 
        This function should be called once every time step, before `update`.
        N)rÚpredictr)rÚtrackrrrr3s
zTracker.predictc
s
|j|
ƒ
\}}}x(|D] \}}|j
|j|j|
|ƒ
qWx|D]}|j
|jƒ
q@Wx|D]}|j|
|ƒ

q\Wd
d„|j
Dƒ
|_
dd„|j
Dƒ
}gg}}    x@|j
D]6‰ˆjƒs´q¦|ˆj7}|    ‡f
dd„ˆjDƒ
7}    gˆ_q¦W|jj    t
j|ƒ
t
j|    ƒ
|ƒ
dS)zÕPerform measurement update and track management.
 
        Parameters
        ----------
        detections : List[deep_sort.detection.Detection]
            A list of detections at the current time step.
 
        c
Ssg|]}
|
jƒs|
‘qSr)
Z
is_deleted)Ú.0Ú
trrrú
<listcomp>Psz"Tracker.update.<locals>.<listcomp>c
Ssg|]}
|
jƒr|
j
‘qSr)Úis_confirmedÚtrack_id)rrrrrrSsc
sg|]
}
ˆj‘qSr)
r)rÚ
_)
rrrrYsN)Ú_matchrÚupdaterZmark_missedÚ_initiate_trackrÚfeaturesrZpartial_fitÚnpÚasarray)
rÚ
detectionsÚmatchesÚunmatched_tracksÚunmatched_detectionsZ    track_idxZ detection_idxZactive_targetsr"Útargetsr)
rrr ;s&

 
 
 
 
 


zTracker.updatec sÀ‡f
d
d„}dd„tˆj
ƒ
Dƒ
}dd„tˆj
ƒ
Dƒ
}tj|ˆjjˆjˆj
|
|ƒ\}}}|‡f
dd„|Dƒ
}‡f
dd„|Dƒ
}tjtj    ˆj
ˆj
|
||ƒ\}    }
}||    }tt||
ƒ
ƒ
}|||fS)NcsXtj
‡f
d
d„|Dƒ
ƒ
}tj
‡
f
dd„|Dƒ
ƒ
}ˆjj||ƒ}tjˆj|ˆ
ˆ||ƒ}|S)Nc
sg|]}
ˆ|
j‘qSr)
Úfeature)rÚ
i)
Údetsrrrbsz8Tracker._match.<locals>.gated_metric.<locals>.<listcomp>c
sg|]}
ˆ|
j‘qSr)
r)rr+)
rrrrcs)r#ÚarrayrZdistancerZgate_cost_matrixr)rr,Z track_indicesZdetection_indicesr"r)Zcost_matrix)
r)r,rrÚgated_metricas





z$Tracker._match.<locals>.gated_metricc
Ssg|]\}
}|jƒr|
‘qSr)
r)rr+rrrrrmsz"Tracker._match.<locals>.<listcomp>c
Ssg|]\}
}|jƒs|
‘qSr)
r)rr+rrrrrosc
s g|]}
ˆj|
j
dkr|
‘qS)
r)rÚtime_since_update)rÚ
k)
rrrr{s
c
s g|]}
ˆj|
j
dkr|
‘qS)
r)rr/)rr0)
rrrr~s
) Ú    enumeraterrZmatching_cascaderZmatching_thresholdrZmin_cost_matchingrZiou_costr ÚlistÚset) rr%r.Zconfirmed_tracksZunconfirmed_tracksZ    matches_aZunmatched_tracks_ar(Ziou_track_candidatesZ    matches_bZunmatched_tracks_br&r'r)
rrr^s






zTracker._matchcCsH|jj
|
jƒƒ
\}}|jjt|||j|j|j|
j    ƒƒ

|jd
7_dS)Nr)
rZinitiateZto_xyahrÚappendrrrrr*)rZ    detectionÚmeanZ
covariancerrrr!‰s
 

zTracker._initiate_trackN)r    r
r)    Ú__name__Ú
__module__Ú__qualname__Ú__doc__rrr rr!rrrrr
s
#+r)Ú
__future__rÚnumpyr#ÚrrrrrrrrrrÚ<module>s