sort多目標跟蹤代碼解讀
算法整體流程：
1.讀取每一幀檢測的結果det。其中det.txt如下：第一個數代表幀號，第三個數~第六個數代表目標(x,y,w,h)，第七個數代表得分score,其它數據不詳。
1,-1,500,158,30.979,70.299,93.673,-3.70694,-7.16689,0
1,-1,246,218,40.258,91.355,69.358,-11.4773,-5.53043,0
1,-1,648,238,36.706,83.294,55.955,-8.82797,-12.7447,0
每一幀取出的數據爲：
x, y, w, h, s
500, 158, 30.979, 70.299, 93.673
246, 218, 40.258, 91.355, 69.358
648, 238, 36.706, 83.294, 55.955
2.sort循環更新
1）對於上一幀的跟蹤器，這一幀首先各做一次預測
2）將預測結果異常的跟蹤器存入to_del數組
3）屏蔽出現無效數值的跟蹤器，然後做了壓縮？
4）逆向刪除異常的跟蹤器
5）將檢測結果分配到跟蹤器上，這裏會得到三個list
跟蹤結果與檢測結果匹配上的、未匹配上的檢測結果、未匹配上的跟蹤器
6）對於匹配上的，將檢測結果更新跟蹤器；
對於未匹配上的檢測結果，認爲是新檢測出來的目標，給他送入新的跟蹤器，每天加一個新的跟蹤器，他的id+1。第一幀全部是未匹配到的。
7）獲取跟蹤器的結果
from __future__ import print_function
from numba import jit
import os.path
import numpy as np
import matplotlib.pyplot as plt
import matplotlib.patches as patches
from skimage import io
from sklearn.utils.linear_assignment_ import linear_assignment
import glob
import time
import argparse
from filterpy.kalman import KalmanFilter
@jit
def iou(bb_test,bb_gt):
"""
Computes IUO between two bboxes in the form [x1,y1,x2,y2]
"""
#大的左上角的點x
xx1 = np.maximum(bb_test[0], bb_gt[0])
# 大的左上角的點y
yy1 = np.maximum(bb_test[1], bb_gt[1])
# 小的右下角點x
xx2 = np.minimum(bb_test[2], bb_gt[2])
# 小的右下角點y
yy2 = np.minimum(bb_test[3], bb_gt[3])
#相交部分的區域
w = np.maximum(0., xx2 - xx1)
h = np.maximum(0., yy2 - yy1)
#相交區域的面積
wh = w * h
#iou, = a/(s1+s2-a),a是相交的面積，s1是第一個Box面積，s2是第二個box面積
o = wh / ((bb_test[2]-bb_test[0])*(bb_test[3]-bb_test[1])
+ (bb_gt[2]-bb_gt[0])*(bb_gt[3]-bb_gt[1]) - wh)
return(o)
def convert_bbox_to_z(bbox):
"""
Takes a bounding box in the form [x1,y1,x2,y2] and returns z in the form
[x,y,s,r] where x,y is the centre of the box and s is the scale/area and r is
the aspect ratio
"""
w = bbox[2]-bbox[0]
h = bbox[3]-bbox[1]
x = bbox[0]+w/2.
y = bbox[1]+h/2.
s = w*h #面積
r = w/float(h)
#轉成4行一列
return np.array([x,y,s,r]).reshape((4,1))
def convert_x_to_bbox(x,score=None):
"""
Takes a bounding box in the centre form [x,y,s,r] and returns it in the form
[x1,y1,x2,y2] where x1,y1 is the top left and x2,y2 is the bottom right
"""
w = np.sqrt(x[2]*x[3])
h = x[2]/w
if(score==None):
return np.array([x[0]-w/2.,x[1]-h/2.,x[0]+w/2.,x[1]+h/2.]).reshape((1,4))
else:
return np.array([x[0]-w/2.,x[1]-h/2.,x[0]+w/2.,x[1]+h/2.,score]).reshape((1,5))
class KalmanBoxTracker(object):
"""
This class represents the internel state of individual tracked objects observed as bbox.
"""
count = 0
def __init__(self,bbox):
"""
Initialises a tracker using initial bounding box.
"""
#define constant velocity model
self.kf = KalmanFilter(dim_x=7, dim_z=4)
self.kf.F = np.array([[1,0,0,0,1,0,0],[0,1,0,0,0,1,0],[0,0,1,0,0,0,1],[0,0,0,1,0,0,0], [0,0,0,0,1,0,0],[0,0,0,0,0,1,0],[0,0,0,0,0,0,1]])
self.kf.H = np.array([[1,0,0,0,0,0,0],[0,1,0,0,0,0,0],[0,0,1,0,0,0,0],[0,0,0,1,0,0,0]])
self.kf.R[2:,2:] *= 10.
self.kf.P[4:,4:] *= 1000. #give high uncertainty to the unobservable initial velocities
self.kf.P *= 10.
self.kf.Q[-1,-1] *= 0.01
self.kf.Q[4:,4:] *= 0.01
self.kf.x[:4] = convert_bbox_to_z(bbox)
self.time_since_update = 0
self.id = KalmanBoxTracker.count
KalmanBoxTracker.count += 1
self.history = []
self.hits = 0
self.hit_streak = 0
self.age = 0
def update(self,bbox):
"""
Updates the state vector with observed bbox.
"""
self.time_since_update = 0
self.history = []
self.hits += 1
self.hit_streak += 1
self.kf.update(convert_bbox_to_z(bbox))
def predict(self):
"""
Advances the state vector and returns the predicted bounding box estimate.
"""
if((self.kf.x[6]+self.kf.x[2])<=0):
self.kf.x[6] *= 0.0
self.kf.predict()
self.age += 1
if(self.time_since_update>0):
self.hit_streak = 0
self.time_since_update += 1
self.history.append(convert_x_to_bbox(self.kf.x))
return self.history[-1]
def get_state(self):
"""
Returns the current bounding box estimate.
"""
return convert_x_to_bbox(self.kf.x)
def associate_detections_to_trackers(detections,trackers,iou_threshold = 0.3):
"""
將檢測結果指定給跟蹤目標
返回匹配到、檢測未匹配到、跟蹤未匹配到
"""
#如果跟蹤器的目標個數爲0
if(len(trackers)==0):
return np.empty((0,2),dtype=int), np.arange(len(detections)), np.empty((0,5),dtype=int)
#iou矩陣，檢測的目標框和跟蹤的目標框的iou
iou_matrix = np.zeros((len(detections),len(trackers)),dtype=np.float32)
#對於每個檢測的結果，計算它和每個跟蹤器結果的Iou
for d,det in enumerate(detections):
for t,trk in enumerate(trackers):
iou_matrix[d,t] = iou(det,trk)
#這裏指的是利用匈牙利算法匹配跟蹤和檢測的iou
matched_indices = linear_assignment(-iou_matrix)
print("matched_indices: ", matched_indices)
'''
matched_indices:
[[0 1]
[1 0]
[2 2]]
'''
unmatched_detections = []
#對於沒有匹配到的檢測結果，將它存放在unmatched_detections
for d,det in enumerate(detections):
if(d not in matched_indices[:,0]):#matched_indices[:,0]代表匹配的檢測器，matched_indices[:,1]匹配的跟蹤器
unmatched_detections.append(d)#將沒匹配到檢測編號存起來
unmatched_trackers = []
for t,trk in enumerate(trackers):
if(t not in matched_indices[:,1]):
unmatched_trackers.append(t)
#低IOU匹配濾除
matches = []
for m in matched_indices:
#如果iou小於閾值的話，將檢測和跟蹤分別存入未匹配檢測、未匹配跟蹤器中
if(iou_matrix[m[0],m[1]]<iou_threshold):
unmatched_detections.append(m[0])
unmatched_trackers.append(m[1])
else:
matches.append(m.reshape(1,2))#轉換成1行2列[檢測器， 跟蹤器]
if(len(matches)==0):
matches = np.empty((0,2),dtype=int)
else:
matches = np.concatenate(matches,axis=0)
return matches, np.array(unmatched_detections), np.array(unmatched_trackers)
class Sort(object):
def __init__(self,max_age=1,min_hits=3):
"""
Sets key parameters for SORT
"""
self.max_age = max_age
self.min_hits = min_hits
self.trackers = []
self.frame_count = 0
def update(self,dets):
"""
Params:
dets是一組數組(x1,y1,x2,y2,score)
dets - a numpy array of detections in the format [[x1,y1,x2,y2,score],[x1,y1,x2,y2,score],...]
Requires: this method must be called once for each frame even with empty detections.
Returns the a similar array, where the last column is the object ID.
返回類似的數組，其中最後一列是對象ID。
NOTE: The number of objects returned may differ from the number of detections provided.
"""
self.frame_count += 1
#從現有的跟蹤器獲取預測的位置
#上一幀目標個數self.trackers
trks = np.zeros((len(self.trackers),5))#len(self.trackers)初始爲0
print("trks: ", trks)
to_del = []
ret = []
for t,trk in enumerate(trks):
pos = self.trackers[t].predict()[0]#對於上一幀的目標，這一幀進行預測
print("pos: ", pos)
trk[:] = [pos[0], pos[1], pos[2], pos[3], 0]
#np.isnan()判斷是否爲空，np.any數組中只有有一個爲true，則返回true
if(np.any(np.isnan(pos))):
to_del.append(t)#存放跟蹤座標數據爲空的數據
#numpy.ma.masked_invaid屏蔽出現無效值的數組, numpy.ma.compress_rows壓縮包含掩碼值2-D數組的整行。
trks = np.ma.compress_rows(np.ma.masked_invalid(trks))
for t in reversed(to_del):#逆向刪除異常的目標
self.trackers.pop(t)
#將檢測結果指定給跟蹤器
matched, unmatched_dets, unmatched_trks = associate_detections_to_trackers(dets,trks)
# 用指定的檢測器，更新匹配到的跟蹤器
for t, trk in enumerate(self.trackers):
#匈牙利算法沒匹配到以及匹配到但是iou低於閾值的跟蹤器
if(t not in unmatched_trks):#如果t是匹配到的目標
d = matched[np.where(matched[:,1]==t)[0],0]#匹配的跟蹤器的編號與t相等，檢測器的id
trk.update(dets[d,:][0])#利用檢測器的結果更新卡爾曼
#對於沒有匹配到的檢測結果，初始化一個新的跟蹤器
for i in unmatched_dets:
trk = KalmanBoxTracker(dets[i,:])
self.trackers.append(trk)
i = len(self.trackers)
for trk in reversed(self.trackers):
d = trk.get_state()[0]
#匹配到的時候hit_streak會加1
if((trk.time_since_update < 1) and (trk.hit_streak >= self.min_hits or self.frame_count <= self.min_hits)):
ret.append(np.concatenate((d,[trk.id+1])).reshape(1,-1)) # +1 as MOT benchmark requires positive
i -= 1
#remove dead tracklet
if(trk.time_since_update > self.max_age):
self.trackers.pop(i)
if(len(ret)>0):
return np.concatenate(ret)
return np.empty((0,5))
def parse_args():
"""Parse input arguments."""
parser = argparse.ArgumentParser(description='SORT demo')
parser.add_argument('--display', dest='display', help='Display online tracker output (slow) [False]',action='store_true')
args = parser.parse_args()
return args
if __name__ == '__main__':
#所有的視頻序列
sequences = ['PETS09-S2L1','TUD-Campus','TUD-Stadtmitte','ETH-Bahnhof','ETH-Sunnyday','ETH-Pedcross2','KITTI-13','KITTI-17','ADL-Rundle-6','ADL-Rundle-8','Venice-2']
args = parse_args()#獲取參數
display = args.display#是否顯示跟蹤過程
phase = 'train'
total_time = 0.0#總共耗時
total_frames = 0#視頻幀總數
colours = np.random.rand(32,3) #32x3的隨機矩陣，用於顯示用
if(display):
#if not os.path.exists('mot_benchmark'):
#print('\n\tERROR: mot_benchmark link not found!\n\n Create a symbolic link to the MOT benchmark\n (https://motchallenge.net/data/2D_MOT_2015/#download). E.g.:\n\n $ ln -s /path/to/MOT2015_challenge/2DMOT2015 mot_benchmark\n\n')
#exit()
plt.ion()#打開交互模式
fig = plt.figure() #圖片1
#如果沒有output文件夾，則創建一個
if not os.path.exists('output'):
os.makedirs('output')
for seq in sequences:
mot_tracker = Sort() #創建一個跟蹤器
print(seq)
#讀取檢測文本
seq_dets = np.loadtxt('data/%s/det/det.txt'%(seq),delimiter=',') #load detections
with open('output/%s.txt'%(seq),'w') as out_file:
print("Processing %s."%(seq))
#seq_dets[:,0].max()))視頻幀數最大值
for frame in range(int(seq_dets[:,0].max())):
frame += 1 #detection and frame numbers begin at 1
dets = seq_dets[seq_dets[:,0]==frame,2:7]#每一幀的數據，第3到第6個數據
dets[:,2:4] += dets[:,0:2] #獲取右下角的座標=左上角座標+寬和高
total_frames += 1
if(display):
ax1 = fig.add_subplot(111, aspect='equal')#1x1第一個子圖
fn = 'data/%s/img1/%06d.jpg' % (seq, frame)#圖像數據
#fn = 'mot_benchmark/%s/%s/img1/%06d.jpg'%(phase,seq,frame)
im =io.imread(fn)#讀取每一幀圖像
ax1.imshow(im)#顯示圖像
plt.title(seq+' Tracked Targets')
start_time = time.time()
#print(dets)
trackers = mot_tracker.update(dets)#利用檢測的結果更新跟蹤器,返回一個5個數的數組
#print(trackers)
cycle_time = time.time() - start_time
total_time += cycle_time
for d in trackers:
print('%d,%d,%.2f,%.2f,%.2f,%.2f,1,-1,-1,-1'%(frame,d[4],d[0],d[1],d[2]-d[0],d[3]-d[1]),file=out_file)
if(display):
d = d.astype(np.int32)#轉換成整形
ax1.add_patch(patches.Rectangle((d[0],d[1]),d[2]-d[0],d[3]-d[1],fill=False,lw=3,ec=colours[d[4]%32,:]))
#ax1.set_adjustable('box-forced')
if(display):
fig.canvas.flush_events()
plt.draw()
ax1.cla()
print("Total Tracking took: %.3f for %d frames or %.1f FPS"%(total_time,total_frames,total_frames/total_time))
if(display):
print("Note: to get real runtime results run without the option: --display")
暫時未理解部分：
1.matched_indices = linear_assignment(-iou_matrix)
2.trks = np.ma.compress_rows(np.ma.masked_invalid(trks))
缺點：
1.這個算法直接利用已標註號的檢測的結果，作爲算法中的一部分。
2.待補充。