Mercurial > hg > nsaunier > traffic-intelligence

comparison scripts/dltrack.py @ 1236:100fe098abe9

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
progress on classification
author Nicolas Saunier <nicolas.saunier@polymtl.ca>
date Tue, 19 Sep 2023 17:04:30 -0400
parents 855abc69fa99
children 31a441efca6c
comparison
equal deleted inserted replaced
1235:855abc69fa99 1236:100fe098abe9
1 #! /usr/bin/env python3 1 #! /usr/bin/env python3
2 # from https://docs.ultralytics.com/modes/track/ 2 # from https://docs.ultralytics.com/modes/track/
3 import sys, argparse 3 import sys, argparse
4 from copy import copy 4 from copy import copy
5 from collections import Counter
5 from ultralytics import YOLO 6 from ultralytics import YOLO
7 from torch import cat
8 from torchvision import ops
6 import cv2 9 import cv2
7 10
8 from trafficintelligence import cvutils, moving, storage, utils 11 from trafficintelligence import cvutils, moving, storage, utils
9 12
10 parser = argparse.ArgumentParser(description='The program tracks objects following the ultralytics yolo executable.')#, epilog = 'Either the configuration filename or the other parameters (at least video and database filenames) need to be provided.') 13 parser = argparse.ArgumentParser(description='The program tracks objects following the ultralytics yolo executable.')#, epilog = 'Either the configuration filename or the other parameters (at least video and database filenames) need to be provided.')
13 parser.add_argument('-m', dest = 'detectorFilename', help = 'name of the detection model file', required = True) 16 parser.add_argument('-m', dest = 'detectorFilename', help = 'name of the detection model file', required = True)
14 parser.add_argument('-t', dest = 'trackerFilename', help = 'name of the tracker file', required = True) 17 parser.add_argument('-t', dest = 'trackerFilename', help = 'name of the tracker file', required = True)
15 parser.add_argument('--display', dest = 'display', help = 'show the results (careful with long videos, risk of running out of memory)', action = 'store_true') 18 parser.add_argument('--display', dest = 'display', help = 'show the results (careful with long videos, risk of running out of memory)', action = 'store_true')
16 parser.add_argument('-f', dest = 'firstFrameNum', help = 'number of first frame number to process', type = int, default = 0) 19 parser.add_argument('-f', dest = 'firstFrameNum', help = 'number of first frame number to process', type = int, default = 0)
17 parser.add_argument('-l', dest = 'lastFrameNum', help = 'number of last frame number to process', type = int, default = float('Inf')) 20 parser.add_argument('-l', dest = 'lastFrameNum', help = 'number of last frame number to process', type = int, default = float('Inf'))
21 parser.add_argument('--bike-pct', dest = 'bikeProportion', help = 'percent of time a person classified as bike or motorbike to be classified as cyclist', type = float, default = 0.2)
18 args = parser.parse_args() 22 args = parser.parse_args()
19 23
20 # required functionality? 24 # required functionality?
21 # # filename of the video to process (can be images, eg image%04d.png) 25 # # filename of the video to process (can be images, eg image%04d.png)
22 # video-filename = laurier.avi 26 # video-filename = laurier.avi
64 model = YOLO(args.detectorFilename, ) # seg yolov8x-seg.pt 68 model = YOLO(args.detectorFilename, ) # seg yolov8x-seg.pt
65 # seg could be used on cropped image... if can be loaded and kept in memory 69 # seg could be used on cropped image... if can be loaded and kept in memory
66 # model = YOLO('/home/nicolas/Research/Data/classification-models/yolo_nas_l.pt ') # AttributeError: 'YoloNAS_L' object has no attribute 'get' 70 # model = YOLO('/home/nicolas/Research/Data/classification-models/yolo_nas_l.pt ') # AttributeError: 'YoloNAS_L' object has no attribute 'get'
67 71
68 # Track with the model 72 # Track with the model
69 #results = model.track(source=args.videoFilename, tracker="/home/nicolas/Research/Data/classification-models/bytetrack.yaml", classes=list(moving.cocoTypeNames.keys()), show=True) # , save_txt=True
70 if args.display: 73 if args.display:
71 windowName = 'frame' 74 windowName = 'frame'
72 cv2.namedWindow(windowName, cv2.WINDOW_NORMAL) 75 cv2.namedWindow(windowName, cv2.WINDOW_NORMAL)
73 76
74 capture = cv2.VideoCapture(args.videoFilename) 77 capture = cv2.VideoCapture(args.videoFilename)
85 results = model.track(frame, tracker=args.trackerFilename, classes=list(moving.cocoTypeNames.keys()), persist=True) 88 results = model.track(frame, tracker=args.trackerFilename, classes=list(moving.cocoTypeNames.keys()), persist=True)
86 # create object with user type and list of 3 features (bottom ones and middle) + projection 89 # create object with user type and list of 3 features (bottom ones and middle) + projection
87 while capture.isOpened() and success and frameNum <= lastFrameNum: 90 while capture.isOpened() and success and frameNum <= lastFrameNum:
88 #for frameNum, result in enumerate(results): 91 #for frameNum, result in enumerate(results):
89 result = results[0] 92 result = results[0]
90 print(frameNum, len(result.boxes)) 93 print(frameNum, len(result.boxes), 'objects')
91 for box in result.boxes: 94 for box in result.boxes:
92 #print(box.cls, box.id, box.xyxy) 95 #print(box.cls, box.id, box.xyxy)
93 if box.id is not None: # None are objects with low confidence 96 if box.id is not None: # None are objects with low confidence
94 num = int(box.id) 97 num = int(box.id.item())
95 xyxy = box.xyxy[0].tolist() 98 #xyxy = box.xyxy[0].tolist()
96 if num in currentObjects: 99 if num in currentObjects:
97 currentObjects[num].timeInterval.last = frameNum 100 currentObjects[num].timeInterval.last = frameNum
98 currentObjects[num].userTypes.append(moving.coco2Types[int(box.cls)]) 101 currentObjects[num].bboxes[frameNum] = copy(box.xyxy)
99 currentObjects[num].features[0].tmpPositions[frameNum] = moving.Point(xyxy[0],xyxy[1]) 102 currentObjects[num].userTypes.append(moving.coco2Types[int(box.cls.item())])
100 currentObjects[num].features[1].tmpPositions[frameNum] = moving.Point(xyxy[2],xyxy[3]) 103 currentObjects[num].features[0].tmpPositions[frameNum] = moving.Point(box.xyxy[0,0].item(), box.xyxy[0,1].item())
101 #features[0].getPositions().addPositionXY(xyxy[0],xyxy[1]) 104 currentObjects[num].features[1].tmpPositions[frameNum] = moving.Point(box.xyxy[0,2].item(), box.xyxy[0,3].item())
102 #features[1].getPositions().addPositionXY(xyxy[2],xyxy[3])
103 else: 105 else:
104 inter = moving.TimeInterval(frameNum,frameNum) 106 inter = moving.TimeInterval(frameNum,frameNum)
105 currentObjects[num] = moving.MovingObject(num, inter) 107 currentObjects[num] = moving.MovingObject(num, inter)
106 currentObjects[num].userTypes = [moving.coco2Types[int(box.cls)]] 108 currentObjects[num].bboxes = {frameNum: copy(box.xyxy)}
109 currentObjects[num].userTypes = [moving.coco2Types[int(box.cls.item())]]
107 currentObjects[num].features = [moving.MovingObject(featureNum), moving.MovingObject(featureNum+1)] 110 currentObjects[num].features = [moving.MovingObject(featureNum), moving.MovingObject(featureNum+1)]
108 currentObjects[num].featureNumbers = [featureNum, featureNum+1] 111 currentObjects[num].featureNumbers = [featureNum, featureNum+1]
109 currentObjects[num].features[0].tmpPositions = {frameNum: moving.Point(xyxy[0],xyxy[1])} 112 currentObjects[num].features[0].tmpPositions = {frameNum: moving.Point(box.xyxy[0,0].item(), box.xyxy[0,1].item())}
110 currentObjects[num].features[1].tmpPositions = {frameNum: moving.Point(xyxy[2],xyxy[3])} 113 currentObjects[num].features[1].tmpPositions = {frameNum: moving.Point(box.xyxy[0,2].item(), box.xyxy[0,3].item())}
111 featureNum += 2 114 featureNum += 2
112 if args.display: 115 if args.display:
113 cvutils.cvImshow(windowName, result.plot()) # original image in orig_img 116 cvutils.cvImshow(windowName, result.plot()) # original image in orig_img
114 key = cv2.waitKey() 117 key = cv2.waitKey()
115 if cvutils.quitKey(key): 118 if cvutils.quitKey(key):
116 break 119 break
117 frameNum += 1 120 frameNum += 1
118 success, frame = capture.read() 121 success, frame = capture.read()
119 results = model.track(frame, persist=True) 122 results = model.track(frame, persist=True)
120 123
121 # interpolate and generate velocity before saving 124 # classification
122 for num, obj in currentObjects.items(): 125 for num, obj in currentObjects.items():
123 obj.setUserType(utils.mostCommon(obj.userTypes)) 126 #obj.setUserType(utils.mostCommon(obj.userTypes)) # improve? mix with speed?
127 userTypeStats = Counter(obj.userTypes)
128 if (4 in userTypeStats or (3 in userTypeStats and 4 in userTypeStats and userTypeStats[3]<=userTypeStats[4])) and userTypeStats[3]+userTypeStats[4] > args.bikeProportion*userTypeStats.total(): # 3 is motorcycle and 4 is cyclist (verif if not turning all motorbike into cyclists)
129 obj.setUserType(4)
130 else:
131 obj.setUserType(userTypeStats.most_common()[0][0])
132
133 # merge bikes and people
134 #Construire graphe bipartite vélo/moto personne
135 #Lien = somme des iou / longueur track vélo
136 #Algo Hongrois
137 #Verif overlap piéton vélo : si long, changement mode (trouver exemples)
138
139 # for all cyclists and motorbikes
140
141 # interpolate and generate velocity (?) before saving
142 for num, obj in currentObjects.items():
124 obj.features[0].timeInterval = copy(obj.getTimeInterval()) 143 obj.features[0].timeInterval = copy(obj.getTimeInterval())
125 obj.features[1].timeInterval = copy(obj.getTimeInterval()) 144 obj.features[1].timeInterval = copy(obj.getTimeInterval())
126 if obj.length() != len(obj.features[0].tmpPositions): # interpolate 145 if obj.length() != len(obj.features[0].tmpPositions): # interpolate
127 obj.features[0].positions = moving.Trajectory.fromPointDict(obj.features[0].tmpPositions) 146 obj.features[0].positions = moving.Trajectory.fromPointDict(obj.features[0].tmpPositions)
128 obj.features[1].positions = moving.Trajectory.fromPointDict(obj.features[1].tmpPositions) 147 obj.features[1].positions = moving.Trajectory.fromPointDict(obj.features[1].tmpPositions)