Mercurial Hosting > traffic-intelligence
changeset 902:c69a8defe5c3
changed workflow of classify objects
author | Nicolas Saunier <nicolas.saunier@polymtl.ca> |
---|---|
date | Thu, 22 Jun 2017 16:57:34 -0400 |
parents | 753a081989e2 |
children | 81ee5aaf213d |
files | scripts/classify-objects.py scripts/display-trajectories.py scripts/extract-appearance-images.py |
diffstat | 3 files changed, 51 insertions(+), 27 deletions(-) [+] |
line wrap: on
line diff
diff -r 753a081989e2 -r c69a8defe5c3 scripts/classify-objects.py --- a/scripts/classify-objects.py Thu Jun 22 12:02:34 2017 -0400 +++ b/scripts/classify-objects.py Thu Jun 22 16:57:34 2017 -0400 @@ -20,7 +20,6 @@ args = parser.parse_args() params, videoFilename, databaseFilename, invHomography, intrinsicCameraMatrix, distortionCoefficients, undistortedImageMultiplication, undistort, firstFrameNum = storage.processVideoArguments(args) - classifierParams = storage.ClassifierParameters(params.classifierFilename) classifierParams.convertToFrames(params.videoFrameRate, 3.6) # conversion from km/h to m/frame @@ -66,18 +65,14 @@ sys.exit() objects = storage.loadTrajectoriesFromSqlite(databaseFilename, 'object', args.nObjects, withFeatures = True) -#features = storage.loadTrajectoriesFromSqlite(databaseFilename, 'feature') -intervals = [] -for obj in objects: - #obj.setFeatures(features) - intervals.append(obj.getTimeInterval()) -timeInterval = moving.TimeInterval.unionIntervals(intervals) +timeInterval = moving.TimeInterval.unionIntervals([obj.getTimeInterval() for obj in objects]) capture = cv2.VideoCapture(videoFilename) width = int(capture.get(cv2.cv.CV_CAP_PROP_FRAME_WIDTH)) height = int(capture.get(cv2.cv.CV_CAP_PROP_FRAME_HEIGHT)) pastObjects = [] +currentObjects = [] if undistort: # setup undistortion [map1, map2] = cvutils.computeUndistortMaps(width, height, undistortedImageMultiplication, intrinsicCameraMatrix, distortionCoefficients) if capture.isOpened(): @@ -92,26 +87,23 @@ if frameNum%50 == 0: print('frame number: {}'.format(frameNum)) if undistort: - img = cv2.remap(img, map1, map2, interpolation=cv2.INTER_LINEAR) - currentObjects = [] + img = cv2.remap(img, map1, map2, interpolation=cv2.INTER_LINEAR) for obj in objects: - inter = obj.getTimeInterval() - if inter.contains(frameNum): - if inter.first == frameNum: - obj.initClassifyUserTypeHoGSVM(speedAggregationFunc, pedBikeCarSVM, bikeCarSVM, classifierParams.maxPedestrianSpeed, classifierParams.maxCyclistSpeed, classifierParams.nFramesIgnoreAtEnds) - currentObjects.append(obj) - elif inter.last == frameNum: - obj.classifyUserTypeHoGSVM(minSpeedEquiprobable = classifierParams.minSpeedEquiprobable, speedProbabilities = speedProbabilities, maxPercentUnknown = classifierParams.maxPercentUnknown) - pastObjects.append(obj) - else: - obj.classifyUserTypeHoGSVMAtInstant(img, frameNum, invHomography, width, height, classifierParams.percentIncreaseCrop, classifierParams.percentIncreaseCrop, classifierParams.minNPixels, classifierParams.hogRescaleSize, classifierParams.hogNOrientations, classifierParams.hogNPixelsPerCell, classifierParams.hogNCellsPerBlock, classifierParams.hogBlockNorm) - currentObjects.append(obj) + if obj.getFirstInstant() == frameNum: + obj.initClassifyUserTypeHoGSVM(speedAggregationFunc, pedBikeCarSVM, bikeCarSVM, classifierParams.maxPedestrianSpeed, classifierParams.maxCyclistSpeed, classifierParams.nFramesIgnoreAtEnds) + currentObjects.append(obj) + objects.remove(obj) + + for obj in currentObjects: + if obj.getLastInstant() == frameNum: + obj.classifyUserTypeHoGSVM(minSpeedEquiprobable = classifierParams.minSpeedEquiprobable, speedProbabilities = speedProbabilities, maxPercentUnknown = classifierParams.maxPercentUnknown) + pastObjects.append(obj) + currentObjects.remove(obj) else: - currentObjects.append(obj) - objects = currentObjects + obj.classifyUserTypeHoGSVMAtInstant(img, frameNum, invHomography, width, height, classifierParams.percentIncreaseCrop, classifierParams.percentIncreaseCrop, classifierParams.minNPixels, classifierParams.hogRescaleSize, classifierParams.hogNOrientations, classifierParams.hogNPixelsPerCell, classifierParams.hogNCellsPerBlock, classifierParams.hogBlockNorm) frameNum += 1 - for obj in objects: + for obj in currentObjects: obj.classifyUserTypeHoGSVM(minSpeedEquiprobable = classifierParams.minSpeedEquiprobable, speedProbabilities = speedProbabilities, maxPercentUnknown = classifierParams.maxPercentUnknown) pastObjects.append(obj) print('Saving user types')
diff -r 753a081989e2 -r c69a8defe5c3 scripts/display-trajectories.py --- a/scripts/display-trajectories.py Thu Jun 22 12:02:34 2017 -0400 +++ b/scripts/display-trajectories.py Thu Jun 22 16:57:34 2017 -0400 @@ -9,8 +9,8 @@ parser = argparse.ArgumentParser(description='The program displays feature or object trajectories overlaid over the video frames.', epilog = 'Either the configuration filename or the other parameters (at least video and database filenames) need to be provided.') parser.add_argument('--cfg', dest = 'configFilename', help = 'name of the configuration file') -parser.add_argument('-d', dest = 'databaseFilename', help = 'name of the Sqlite database file') -parser.add_argument('-i', dest = 'videoFilename', help = 'name of the video file') +parser.add_argument('-d', dest = 'databaseFilename', help = 'name of the Sqlite database file (overrides the configuration file)') +parser.add_argument('-i', dest = 'videoFilename', help = 'name of the video file (overrides the configuration file)') parser.add_argument('-t', dest = 'trajectoryType', help = 'type of trajectories to display', choices = ['feature', 'object'], default = 'feature') parser.add_argument('-o', dest = 'homographyFilename', help = 'name of the image to world homography file') parser.add_argument('--intrinsic', dest = 'intrinsicCameraMatrixFilename', help = 'name of the intrinsic camera file')
diff -r 753a081989e2 -r c69a8defe5c3 scripts/extract-appearance-images.py --- a/scripts/extract-appearance-images.py Thu Jun 22 12:02:34 2017 -0400 +++ b/scripts/extract-appearance-images.py Thu Jun 22 16:57:34 2017 -0400 @@ -9,14 +9,46 @@ parser = argparse.ArgumentParser(description='The program extracts labeled image patches to train the HoG-SVM classifier, and optionnally speed information') parser.add_argument('--cfg', dest = 'configFilename', help = 'name of the configuration file', required = True) +parser.add_argument('-d', dest = 'databaseFilename', help = 'name of the Sqlite database file (overrides the configuration file)') +parser.add_argument('-i', dest = 'videoFilename', help = 'name of the video file (overrides the configuration file)') +parser.add_argument('--gt', dest = 'classificationAnnotationFilename', help = 'name of the file containing the correct classes (user types)', required = True) +parser.add_argument('-s', dest = 'nFramesStep', help = 'number of frames between each saved patch', default = 50, type = int) +parser.add_argument('-n', dest = 'nObjects', help = 'number of objects to use to extract patches from', type = int, default = None) +parser.add_argument('--compute-speed-distributions', dest = 'computeSpeedDistribution', help = 'computes the distribution of the road users of each type and fits parameters to each', action = 'store_true') + #parser.add_argument('-d', dest = 'directoryName', help = 'parent directory name for the directories containing the samples for the different road users', required = True) args = parser.parse_args() -params = storage.ProcessParameters(args.configFilename) +params, videoFilename, databaseFilename, invHomography, intrinsicCameraMatrix, distortionCoefficients, undistortedImageMultiplication, undistort, firstFrameNum = storage.processVideoArguments(args) classifierParams = storage.ClassifierParameters(params.classifierFilename) -# need all info as for classification (image info) +objects = storage.loadTrajectoriesFromSqlite(databaseFilename, 'object', args.nObjects, withFeatures = True) +timeInterval = moving.TimeInterval.unionIntervals([obj.getTimeInterval() for obj in objects]) + +capture = cv2.VideoCapture(videoFilename) +width = int(capture.get(cv2.cv.CV_CAP_PROP_FRAME_WIDTH)) +height = int(capture.get(cv2.cv.CV_CAP_PROP_FRAME_HEIGHT)) + +if undistort: # setup undistortion + [map1, map2] = cvutils.computeUndistortMaps(width, height, undistortedImageMultiplication, intrinsicCameraMatrix, distortionCoefficients) +if capture.isOpened(): + ret = True + frameNum = timeInterval.first + capture.set(cv2.cv.CV_CAP_PROP_POS_FRAMES, frameNum) + lastFrameNum = timeInterval.last + while ret and frameNum <= lastFrameNum: + ret, img = capture.read() + if ret: + if frameNum%50 == 0: + print('frame number: {}'.format(frameNum)) + if undistort: + img = cv2.remap(img, map1, map2, interpolation=cv2.INTER_LINEAR) + + + frameNum += 1 + + # todo speed info: distributions AND min speed equiprobable