Mercurial > hg > nsaunier > traffic-intelligence

comparison scripts/classify-objects.py @ 1242:4cd8ace3552f

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
major update for classification, allowing the use of neural network classification
author Nicolas Saunier <nicolas.saunier@polymtl.ca>
date Wed, 07 Feb 2024 11:43:03 -0500
parents ab4c72b9475c
children 88eedf79f16a
comparison
equal deleted inserted replaced
1241:ab4c72b9475c 1242:4cd8ace3552f
50 50
51 # log logistic for ped and bik otherwise ((pedBeta/pedAlfa)*((sMean/pedAlfa)**(pedBeta-1)))/((1+(sMean/pedAlfa)**pedBeta)**2.) 51 # log logistic for ped and bik otherwise ((pedBeta/pedAlfa)*((sMean/pedAlfa)**(pedBeta-1)))/((1+(sMean/pedAlfa)**pedBeta)**2.)
52 carNorm = norm(classifierParams.meanVehicleSpeed, classifierParams.stdVehicleSpeed) 52 carNorm = norm(classifierParams.meanVehicleSpeed, classifierParams.stdVehicleSpeed)
53 pedNorm = norm(classifierParams.meanPedestrianSpeed, classifierParams.stdPedestrianSpeed) 53 pedNorm = norm(classifierParams.meanPedestrianSpeed, classifierParams.stdPedestrianSpeed)
54 # numpy lognorm shape, loc, scale: shape for numpy is scale (std of the normal) and scale for numpy is exp(location) (loc=mean of the normal) 54 # numpy lognorm shape, loc, scale: shape for numpy is scale (std of the normal) and scale for numpy is exp(location) (loc=mean of the normal)
55 bicLogNorm = lognorm(classifierParams.scaleCyclistSpeed, loc = 0., scale = np.exp(classifierParams.locationCyclistSpeed)) 55 cycLogNorm = lognorm(classifierParams.scaleCyclistSpeed, loc = 0., scale = np.exp(classifierParams.locationCyclistSpeed))
56 speedProbabilities = {moving.userTypeNames[1]: lambda s: carNorm.pdf(s), 56 speedProbabilities = {moving.userType2Num['car']: lambda s: carNorm.pdf(s),
57 moving.userTypeNames[2]: lambda s: pedNorm.pdf(s), 57 moving.userType2Num['pedestrian']: lambda s: pedNorm.pdf(s),
58 moving.userTypeNames[4]: lambda s: bicLogNorm.pdf(s)} 58 moving.userType2Num['cyclist']: lambda s: cycLogNorm.pdf(s)}
59 if useYolo: # consider other user types
60 for i in [3, 5, 6, 7]:
61 speedProbabilities[i] = speedProbabilities[moving.userType2Num['car']]
59 62
60 if args.plotSpeedDistribution: 63 if args.plotSpeedDistribution:
61 import matplotlib.pyplot as plt 64 import matplotlib.pyplot as plt
62 plt.figure() 65 plt.figure()
63 speeds = np.arange(0.1, args.maxSpeedDistributionPlot, 0.1) 66 speeds = np.arange(0.1, args.maxSpeedDistributionPlot, 0.1)
104 if frameNum%50 == 0: 107 if frameNum%50 == 0:
105 print('frame number: {}'.format(frameNum)) 108 print('frame number: {}'.format(frameNum))
106 #if undistort: 109 #if undistort:
107 # img = cv2.remap(img, map1, map2, interpolation=cv2.INTER_LINEAR) 110 # img = cv2.remap(img, map1, map2, interpolation=cv2.INTER_LINEAR)
108 if useYolo: 111 if useYolo:
109 results = yolo.predict(img, classes=list(moving.cocoTypeNames.keys()), verbose=False) 112 results = yolo.predict(img, conf = classifierParams.confidence, classes=list(moving.cocoTypeNames.keys()), verbose=False)
110 113
111 for obj in objects[:]: 114 for obj in objects[:]:
112 if obj.getFirstInstant() <= frameNum: # if images are skipped 115 if obj.getFirstInstant() <= frameNum: # if images are skipped
113 obj.initClassifyUserTypeHoGSVM(speedAggregationFunc, pedBikeCarSVM, bikeCarSVM, classifierParams.maxPedestrianSpeed, classifierParams.maxCyclistSpeed, classifierParams.nFramesIgnoreAtEnds, invHomography, intrinsicCameraMatrix, distortionCoefficients) 116 obj.initClassifyUserTypeHoGSVM(speedAggregationFunc, pedBikeCarSVM, bikeCarSVM, classifierParams.maxPedestrianSpeed, classifierParams.maxCyclistSpeed, classifierParams.nFramesIgnoreAtEnds, invHomography, intrinsicCameraMatrix, distortionCoefficients)
114 currentObjects.append(obj) 117 currentObjects.append(obj)
115 objects.remove(obj) 118 objects.remove(obj)
116 119
117 for obj in currentObjects[:]: 120 for obj in currentObjects[:]:
118 if obj.getLastInstant() <= frameNum: 121 if obj.getLastInstant() <= frameNum:
119 obj.classifyUserTypeHoGSVM(minSpeedEquiprobable = classifierParams.minSpeedEquiprobable, speedProbabilities = speedProbabilities, maxPercentUnknown = classifierParams.maxPercentUnknown) 122 obj.classifyUserTypeHoGSVM(classifierParams.minSpeedEquiprobable, speedProbabilities, classifierParams.maxPercentUnknown)
120 pastObjects.append(obj) 123 pastObjects.append(obj)
121 currentObjects.remove(obj) 124 currentObjects.remove(obj)
122 else: 125 else:
123 if useYolo: 126 if useYolo:
124 # if one feature falls in bike, it's a bike 127 # if one feature falls in bike, it's a bike
125 # could one count all hits in various objects, or one takes majority at the instant? 128 # could one count all hits in various objects, or one takes majority at the instant?
126 # obj.classifyUserTypeYoloAtInstant(img, frameNum, width, height, classifierParams.percentIncreaseCrop, classifierParams.percentIncreaseCrop, results[0].boxes) 129 #print('yolo', len(results[0].boxes))
127 pass 130 obj.classifyUserTypeYoloAtInstant(frameNum, results[0].boxes)
128 else: 131 else:
129 obj.classifyUserTypeHoGSVMAtInstant(img, frameNum, width, height, classifierParams.percentIncreaseCrop, classifierParams.percentIncreaseCrop, classifierParams.minNPixels, classifierParams.hogRescaleSize, classifierParams.hogNOrientations, classifierParams.hogNPixelsPerCell, classifierParams.hogNCellsPerBlock, classifierParams.hogBlockNorm) 132 obj.classifyUserTypeHoGSVMAtInstant(img, frameNum, width, height, classifierParams.percentIncreaseCrop, classifierParams.percentIncreaseCrop, classifierParams.minNPixels, classifierParams.hogRescaleSize, classifierParams.hogNOrientations, classifierParams.hogNPixelsPerCell, classifierParams.hogNCellsPerBlock, classifierParams.hogBlockNorm)
130 if args.verbose: 133 if args.verbose:
131 print('obj {}@{}: {}'.format(obj.getNum(), frameNum, moving.userTypeNames[obj.userTypes[frameNum]])) 134 print('obj {}@{}: {}'.format(obj.getNum(), frameNum, moving.userTypeNames[obj.userTypes[frameNum]]))
132 frameNum += 1 135 frameNum += 1
133 136
134 for obj in currentObjects: 137 for obj in currentObjects:
135 obj.classifyUserTypeHoGSVM(minSpeedEquiprobable = classifierParams.minSpeedEquiprobable, speedProbabilities = speedProbabilities, maxPercentUnknown = classifierParams.maxPercentUnknown) 138 obj.classifyUserTypeHoGSVM(classifierParams.minSpeedEquiprobable, speedProbabilities, classifierParams.maxPercentUnknown)
136 pastObjects.append(obj) 139 pastObjects.append(obj)
137 print('Saving user types') 140 print('Saving user types')
138 storage.setRoadUserTypes(databaseFilename, pastObjects) 141 storage.setRoadUserTypes(databaseFilename, pastObjects)