Mercurial > hg > nsaunier > traffic-intelligence

comparison scripts/classify-objects.py @ 683:f3a0b652b590 dev

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
added function to display the speed distributions and optimize a little going through objects
author Nicolas Saunier <nicolas.saunier@polymtl.ca>
date Fri, 05 Jun 2015 11:04:06 +0200
parents fbe29be25501
children 6d2ece4aed7c
comparison
equal deleted inserted replaced
682:fbe29be25501 683:f3a0b652b590
10 10
11 # TODO add mode detection live 11 # TODO add mode detection live
12 12
13 parser = argparse.ArgumentParser(description='The program processes indicators for all pairs of road users in the scene') 13 parser = argparse.ArgumentParser(description='The program processes indicators for all pairs of road users in the scene')
14 parser.add_argument('--cfg', dest = 'configFilename', help = 'name of the configuration file', required = True) 14 parser.add_argument('--cfg', dest = 'configFilename', help = 'name of the configuration file', required = True)
15 parser.add_argument('--plot-speed-distributions', dest = 'plotSpeedDistribution', help = 'simply plots the distributions used for each user type', action = 'store_true')
16 parser.add_argument('--max-speed-distribution-plot', dest = 'maxSpeedDistributionPlot', help = 'if plotting the user distributions, the maximum speed to display', type = float, default = 50.)
15 #parser.add_argument('-u', dest = 'undistort', help = 'undistort the video (because features have been extracted that way)', action = 'store_true') 17 #parser.add_argument('-u', dest = 'undistort', help = 'undistort the video (because features have been extracted that way)', action = 'store_true')
16 #parser.add_argument('-f', dest = 'firstFrameNum', help = 'number of first frame number to display', type = int) 18 #parser.add_argument('-f', dest = 'firstFrameNum', help = 'number of first frame number to display', type = int)
17 #parser.add_argument('--last-frame', dest = 'lastFrameNum', help = 'number of last frame number to save (for image saving, no display is made)', type = int) 19 #parser.add_argument('--last-frame', dest = 'lastFrameNum', help = 'number of last frame number to save (for image saving, no display is made)', type = int)
18 # parser.add_argument('--min-speed-equiprobable', dest = 'minSpeedEquiprobable', help = 'speed value below which all classes are equiprobable (distributions give odd values there) (km/h)', type = float, default = 3.33) 20 # parser.add_argument('--min-speed-equiprobable', dest = 'minSpeedEquiprobable', help = 'speed value below which all classes are equiprobable (distributions give odd values there) (km/h)', type = float, default = 3.33)
19 # parser.add_argument('--speed-aggregation', dest = 'speedAggregationMethod', help = 'method to aggregate road user speed', type = str, choices = ['median', 'mean', 'quantile'], default = 'median') 21 # parser.add_argument('--speed-aggregation', dest = 'speedAggregationMethod', help = 'method to aggregate road user speed', type = str, choices = ['median', 'mean', 'quantile'], default = 'median')
21 23
22 args = parser.parse_args() 24 args = parser.parse_args()
23 params = storage.ProcessParameters(args.configFilename) 25 params = storage.ProcessParameters(args.configFilename)
24 26
25 params.convertToFrames(3.6) 27 params.convertToFrames(3.6)
26 invHomography = np.linalg.inv(params.homography) 28 if params.homography is not None:
29 invHomography = np.linalg.inv(params.homography)
27 30
28 if params.speedAggregationMethod == 'median': 31 if params.speedAggregationMethod == 'median':
29 speedAggregationFunc = np.median 32 speedAggregationFunc = np.median
30 elif params.speedAggregationMethod == 'mean': 33 elif params.speedAggregationMethod == 'mean':
31 speedAggregationFunc = np.mean 34 speedAggregationFunc = np.mean
32 elif params.speedAggregationMethod == 'quantile': 35 elif params.speedAggregationMethod == 'quantile':
33 speedAggregationFunc = lambda speeds: np.percentile(speeds, args.speedAggregationQuantile) 36 speedAggregationFunc = lambda speeds: np.percentile(speeds, args.speedAggregationQuantile)
34 else: 37 else:
35 print('Unknown speed aggregation method: {}. Exiting'.format(params.speedAggregationMethod)) 38 print('Unknown speed aggregation method: {}. Exiting'.format(params.speedAggregationMethod))
36 from sys import exit 39 sys.exit()
37 exit()
38 40
39 pedBikeCarSVM = ml.SVM() 41 pedBikeCarSVM = ml.SVM()
40 pedBikeCarSVM.load(params.pedBikeCarSVMFilename) 42 pedBikeCarSVM.load(params.pedBikeCarSVMFilename)
41 bikeCarSVM = ml.SVM() 43 bikeCarSVM = ml.SVM()
42 bikeCarSVM.load(params.bikeCarSVMFilename) 44 bikeCarSVM.load(params.bikeCarSVMFilename)
44 # log logistic for ped and bik otherwise ((pedBeta/pedAlfa)*((sMean/pedAlfa)**(pedBeta-1)))/((1+(sMean/pedAlfa)**pedBeta)**2.) 46 # log logistic for ped and bik otherwise ((pedBeta/pedAlfa)*((sMean/pedAlfa)**(pedBeta-1)))/((1+(sMean/pedAlfa)**pedBeta)**2.)
45 speedProbabilities = {'car': lambda s: norm(params.meanVehicleSpeed, params.stdVehicleSpeed).pdf(s), 47 speedProbabilities = {'car': lambda s: norm(params.meanVehicleSpeed, params.stdVehicleSpeed).pdf(s),
46 'pedestrian': lambda s: norm(params.meanPedestrianSpeed, params.stdPedestrianSpeed).pdf(s), 48 'pedestrian': lambda s: norm(params.meanPedestrianSpeed, params.stdPedestrianSpeed).pdf(s),
47 'bicycle': lambda s: lognorm(params.scaleCyclistSpeed, loc = 0., scale = np.exp(params.locationCyclistSpeed)).pdf(s)} # lognorm shape, loc, scale 49 'bicycle': lambda s: lognorm(params.scaleCyclistSpeed, loc = 0., scale = np.exp(params.locationCyclistSpeed)).pdf(s)} # lognorm shape, loc, scale
48 50
49 def plotSpeedProbabilities(): 51 if args.plotSpeedDistribution:
52 import matplotlib.pyplot as plt
53 plt.figure()
50 for k in speedProbabilities: 54 for k in speedProbabilities:
51 plot(arange(0.1, 50, 0.1), [speedProbabilities[k](s/3.6/25) for s in arange(0.1, 50, 0.1)]) 55 plt.plot(np.arange(0.1, args.maxSpeedDistributionPlot, 0.1), [speedProbabilities[k](s/3.6/25) for s in np.arange(0.1, args.maxSpeedDistributionPlot, 0.1)], label = k)
56 plt.xlabel('Speed (km/h)')
57 plt.ylabel('Probability')
58 plt.legend()
59 plt.title('Probability Density Function')
60 plt.show()
61 sys.exit()
52 62
53 objects = storage.loadTrajectoriesFromSqlite(params.databaseFilename, 'object') 63 objects = storage.loadTrajectoriesFromSqlite(params.databaseFilename, 'object')
54 features = storage.loadTrajectoriesFromSqlite(params.databaseFilename, 'feature') 64 features = storage.loadTrajectoriesFromSqlite(params.databaseFilename, 'feature')
55 intervals = [] 65 intervals = []
56 for obj in objects: 66 for obj in objects:
73 while ret and frameNum <= lastFrameNum: 83 while ret and frameNum <= lastFrameNum:
74 ret, img = capture.read() 84 ret, img = capture.read()
75 if ret: 85 if ret:
76 if frameNum%50 == 0: 86 if frameNum%50 == 0:
77 print('frame number: {}'.format(frameNum)) 87 print('frame number: {}'.format(frameNum))
88 currentObjects = []
89 for obj in objects:
90 if obj.getLastFrameNum() < frameNum:
91 obj.classifyUserTypeHoGSVM(minSpeedEquiprobable = params.minSpeedEquiprobable, speedProbabilities = speedProbabilities)
92 else:
93 currentObjects.append(obj)
94 objects = currentObjects
78 if params.undistort: 95 if params.undistort:
79 img = cv2.remap(img, map1, map2, interpolation=cv2.INTER_LINEAR) 96 img = cv2.remap(img, map1, map2, interpolation=cv2.INTER_LINEAR)
80 for obj in objects: 97 for obj in objects:
81 if obj.existsAtInstant(frameNum): 98 if obj.existsAtInstant(frameNum):
82 if obj.getFirstInstant() == frameNum: 99 if obj.getFirstInstant() == frameNum:
83 print 'first frame for obj {}'.format(obj.getNum())
84 obj.initClassifyUserTypeHoGSVM(speedAggregationFunc, pedBikeCarSVM, bikeCarSVM, params.maxPedestrianSpeed, params.maxCyclistSpeed, params.nFramesIgnoreAtEnds) 100 obj.initClassifyUserTypeHoGSVM(speedAggregationFunc, pedBikeCarSVM, bikeCarSVM, params.maxPedestrianSpeed, params.maxCyclistSpeed, params.nFramesIgnoreAtEnds)
85 obj.classifyUserTypeHoGSVMAtInstant(img, frameNum, invHomography, width, height, 0.2, 0.2, 800) # px, py, pixelThreshold 101 obj.classifyUserTypeHoGSVMAtInstant(img, frameNum, invHomography, width, height, 0.2, 0.2, 800) # px, py, pixelThreshold
86 frameNum += 1 102 frameNum += 1
87 103
88 for obj in objects: 104 for obj in objects: