Mercurial > hg > nsaunier > traffic-intelligence

comparison scripts/learn-motion-patterns.py @ 907:9fd7b18f75b4

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
re arranged motion pattern learning
author Nicolas Saunier <nicolas.saunier@polymtl.ca>
date Fri, 23 Jun 2017 23:50:02 -0400
parents 8e8ec4ece66e
children b297525b2cbf
comparison
equal deleted inserted replaced
906:a57e6fbcd8e3 907:9fd7b18f75b4
16 parser.add_argument('-e', dest = 'epsilon', help = 'distance for the similarity of trajectory points', type = float, required = True) 16 parser.add_argument('-e', dest = 'epsilon', help = 'distance for the similarity of trajectory points', type = float, required = True)
17 parser.add_argument('--metric', dest = 'metric', help = 'metric for the similarity of trajectory points', default = 'cityblock') # default is manhattan distance 17 parser.add_argument('--metric', dest = 'metric', help = 'metric for the similarity of trajectory points', default = 'cityblock') # default is manhattan distance
18 parser.add_argument('-s', dest = 'minSimilarity', help = 'minimum similarity to put a trajectory in a cluster', type = float, required = True) 18 parser.add_argument('-s', dest = 'minSimilarity', help = 'minimum similarity to put a trajectory in a cluster', type = float, required = True)
19 parser.add_argument('-c', dest = 'minClusterSize', help = 'minimum cluster size', type = int, default = None) 19 parser.add_argument('-c', dest = 'minClusterSize', help = 'minimum cluster size', type = int, default = None)
20 parser.add_argument('--random', dest = 'randomInitialization', help = 'random initialization of clustering algorithm', action = 'store_true') 20 parser.add_argument('--random', dest = 'randomInitialization', help = 'random initialization of clustering algorithm', action = 'store_true')
21 parser.add_argument('--subsample', dest = 'positionSubsamplingRate', help = 'rate of position subsampling (1 every n positions)', type = int, default = None) 21 parser.add_argument('--subsample', dest = 'positionSubsamplingRate', help = 'rate of position subsampling (1 every n positions)', type = int)
22 parser.add_argument('--display', dest = 'display', help = 'display trajectories', action = 'store_true') 22 parser.add_argument('--display', dest = 'display', help = 'display trajectories', action = 'store_true')
23 parser.add_argument('--save-similarities', dest = 'saveSimilarities', help = 'save computed similarities (in addition to prototypes)', action = 'store_true') 23 parser.add_argument('--save-similarities', dest = 'saveSimilarities', help = 'save computed similarities (in addition to prototypes)', action = 'store_true')
24 #parser.add_argument('--save-matches', dest = 'saveMatches', help = 'save the matched prototype information', action = 'store_true')
24 25
25 args = parser.parse_args() 26 args = parser.parse_args()
26 27
27 # TODO parameters (random init?) and what to learn from: objects, features, longest features from objects 28 # use cases
29 # 1. learn proto from one file, save in same or another (with traj)
30 # 2. load proto, load objects, update proto, save proto
31 # 3. assign objects from one db to proto
32 # 4. load objects from several files, save in another
33
28 # TODO add possibility to cluter with velocities 34 # TODO add possibility to cluter with velocities
35 # TODO add possibility to start with saved prototypes so that one can incrementally learn from several databases
36 # save prototypes with database name, add option to keep trajectory along: if saved in same db, no need
37 # load proto must load the movingobject
38 # save the objects that match the prototypes
39 # write an assignment function for objects
29 40
30 trajectoryType = args.trajectoryType 41 trajectoryType = args.trajectoryType
31 prototypeType = args.trajectoryType 42 prototypeType = args.trajectoryType
32 if args.trajectoryType == 'objectfeatures': 43 if args.trajectoryType == 'objectfeatures':
33 trajectoryType = 'object' 44 trajectoryType = 'object'
47 lcss = utils.LCSS(metric = args.metric, epsilon = args.epsilon) 58 lcss = utils.LCSS(metric = args.metric, epsilon = args.epsilon)
48 nTrajectories = len(trajectories) 59 nTrajectories = len(trajectories)
49 60
50 similarities = -np.ones((nTrajectories, nTrajectories)) 61 similarities = -np.ones((nTrajectories, nTrajectories))
51 62
52 prototypeIndices, labels = ml.prototypeCluster(trajectories, similarities, args.minSimilarity, lambda x,y : lcss.computeNormalized(x, y), args.minClusterSize, args.randomInitialization) # this line can be called again without reinitializing similarities 63 prototypeIndices, labels = ml.prototypeCluster(trajectories, similarities, args.minSimilarity, lambda x,y : lcss.computeNormalized(x, y), args.minClusterSize, args.randomInitialization, True, None) # this line can be called again without reinitializing similarities
53 64
54 clusterSizes = ml.computeClusterSizes(labels, prototypeIndices, -1) 65 clusterSizes = ml.computeClusterSizes(labels, prototypeIndices, -1)
55 print(clusterSizes) 66 print(clusterSizes)
56 67
57 storage.savePrototypesToSqlite(args.databaseFilename, [objects[i].getNum() for i in prototypeIndices], prototypeType, [clusterSizes[i] for i in prototypeIndices]) # if saving filenames, add for example [objects[i].dbFilename for i in prototypeIndices] 68 storage.savePrototypesToSqlite(args.databaseFilename, [objects[i].getNum() for i in prototypeIndices], prototypeType, [clusterSizes[i] for i in prototypeIndices]) # if saving filenames, add for example [objects[i].dbFilename for i in prototypeIndices]
58 69
59 if args.saveSimilarities: 70 if args.saveSimilarities:
60 np.savetxt(utils.removeExtension(args.databaseFilename)+'-prototype-similarities.txt.gz', similarities, '%.4f') 71 np.savetxt(utils.removeExtension(args.databaseFilename)+'-prototype-similarities.txt.gz', similarities, '%.4f')
72
73 # if args.saveMatches:
74 # out = storage.openCheck(utils.removeExtension(args.databaseFilename)+'prototypes-matches.csv', 'w')
75 # for o in ojbects:
76 # out.write('')
61 77
62 if args.display: 78 if args.display:
63 from matplotlib.pyplot import figure, show, axis 79 from matplotlib.pyplot import figure, show, axis
64 figure() 80 figure()
65 for i,o in enumerate(objects): 81 for i,o in enumerate(objects):
70 o.plot(utils.colors[labels[i]]) 86 o.plot(utils.colors[labels[i]])
71 for i in prototypeIndices: 87 for i in prototypeIndices:
72 objects[i].plot(utils.colors[i]+'o') 88 objects[i].plot(utils.colors[i]+'o')
73 axis('equal') 89 axis('equal')
74 show() 90 show()
75
76 # TODO store the prototypes trajectories, add option so store similarities (the most expensive stuff) with limited accuracy