-
Notifications
You must be signed in to change notification settings - Fork 0
/
multiObjectTracking.m
259 lines (214 loc) · 7.82 KB
/
multiObjectTracking.m
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
% VIDEO INFO - 11 FPS
function multiObjectTracking()
%Setup vlFeat
warning off;
%JUAN PC
%run ('vlfeat\toolbox\mex\mexw64\vl_version.mexw64')
%run('C:\MATLAB\ImageCateogrisation\vlfeat\toolbox\vl_setup');
%ADRIAN PC
run('C:\Users\Admin\Documents\MATLAB\Extra\vlfeat\toolbox\vl_setup');
%Load Masks
ATTARD_MASK = single(rgb2gray(imread('attardmask.jpg')))./255;
%Load PreTrained Data
load('positiveHistograms.mat');
load('negativeHistograms.mat');
load('cTotalDescriptors.mat');
%Setup Detector to use
detector = 'SIFT'; %SIFT/SURF/ESURF
%Setup Training Data
trainingData = [positiveHistograms(1:end,:);negativeHistograms(1:end,:)];
trainingLabels = [ones(size(positiveHistograms,1),1);zeros(size(negativeHistograms,1),1)];
%Train KNN Model
KNNTrainedModel = fitcknn(trainingData, trainingLabels);
%Setup Video Paths
ATTARD_VIDS_PATH = 'C:\Users\Admin\Documents\GIT\skylinewebcamcrawler\videos\Attard, Mdina Road from Citroen Showroom\';
if ~exist('output.json')
fileID = fopen('output.json','w');
fprintf(fileID,'{\n');
fclose(fileID);
end
while(true)
direc = dir(strcat(ATTARD_VIDS_PATH,'*.mp4'));
currFile = strcat(ATTARD_VIDS_PATH,direc(1).name)
tic;
[carSpeed,carCount] = trackVideo(currFile,ATTARD_MASK,KNNTrainedModel,cTotalDescriptors, detector);
elapsedTime = toc
if(carSpeed <= 35) && (carCount >=15)
isTrafficResult = 'true'
else
isTrafficResult = 'false'
end
fileID = fopen('output.json','a');
toOutput = strcat('"', char(datetime('now')),'":[');
toOutput = strcat(toOutput,'{','"LOCATION": "ATTARD", "TRAFFIC":"',isTrafficResult, '","SPEED":"', num2str(carSpeed), '","COUNT":"', num2str(carCount),'"}');
toOutput = strcat(toOutput,'],\n');
fprintf(fileID,toOutput);
fclose(fileID);
currentFileContents = fileread('output.json');
currentFileContents = currentFileContents(1:size(currentFileContents,2)-2); %Remove las newline and comma
file2 = fopen('resultsToUpload.json','w');
toOutput = strcat(currentFileContents,'\n}\n');
fprintf(file2,toOutput);
fclose(fileID);
%Setup FTP
ftpObject = ftp('ftp.nnjconstruction.com:21','lifex@nnjconstruction.com','Lifex..2016');
mput(ftpObject,'resultsToUpload.json');
close(ftpObject);
videosToDelete = ceil((elapsedTime ./ 60) ./ 2);
for i = 1:videosToDelete
delete(strcat(ATTARD_VIDS_PATH,direc(i).name));
end
end
end
function [carSpeed,carCount] = trackVideo(videoPath,videoMask,KNNTrainedModel,cTotalDescriptors, detector)
videoInfo = VideoReader(videoPath);
videoDur = videoInfo.Duration;
videoFPS = videoInfo.FrameRate;
lastFrame = round(videoDur*videoFPS);
%Setup System
obj = setupSystemObjects(videoPath);
opticFlow = opticalFlowLK('NoiseThreshold',0.009);
% Detect moving objects, and track them across video frames.
frameCount = 0;
mag = [];
count = [];
tenFG = 0;
elevenFG = 0;
%figure;
while ~isDone(obj.reader)
if frameCount == tenFG && frameCount ~= elevenFG
tenFG = tenFG + (videoFPS-1);
%Read next frame and apply mask
frame = readFrame(obj, videoMask, true);
%Transform image into blob image
blobImg = createBlobImage(frame, obj, true);
flow = estimateFlow(opticFlow,blobImg);
elseif frameCount == tenFG && frameCount == elevenFG
tenFG = tenFG + (videoFPS-1);
end
%Every 11 Frames ie every second calculate speed and car count
if frameCount == elevenFG
elevenFG = elevenFG + videoFPS;
%Read next frame and apply mask
frame = readFrame(obj, videoMask, true);
%Transform image into blob image
blobImg = createBlobImage(frame, obj, true);
%detect velocity
flow = estimateFlow(opticFlow,blobImg);
Vx = nansum(nansum(flow.Vx));
Vy = nansum(nansum(flow.Vy));
mag(end+1) = sqrt(Vx^2 + Vy^2);
count(end+1) = getNumberOfCars(KNNTrainedModel,frame,9, cTotalDescriptors, detector);
%fprintf( 'Car Speed: %d\n', mag(end));
%fprintf( 'Car Count: %d\n\n', count(end));
%imshow(frame);
elseif frameCount ~= tenFG
%Read next frame and apply mask
frame = readFrame(obj, videoMask, false);
%Transform image into blob image
blobImg = createBlobImage(frame, obj, false);
end
frameCount = frameCount +1;
%At the end of the 2 mins (11 frames per second *120 seconds) = 1320 frames
%if frameCount == 1320
if frameCount == lastFrame
magMedian = median(mag);
countMedian = round(median(count));
if (countMedian ~= 0)
magMedian = (magMedian / countMedian);
else
magMedian = 0;
end
carSpeed = magMedian;
carCount = countMedian;
return
end
end
end
function obj = setupSystemObjects(vidPath)
% Create a video file reader.
obj.reader = vision.VideoFileReader(vidPath);
% Create System objects for foreground detection and blob analysis
% The foreground detector is used to segment moving objects from
% the background. It outputs a binary mask, where the pixel value
% of 1 corresponds to the foreground and the value of 0 corresponds
% to the background.
obj.detector = vision.ForegroundDetector('NumGaussians', 3, ...
'NumTrainingFrames', 40, 'MinimumBackgroundRatio', 0.7);
end
%reads a frame and applies a mask to remove extra detail
function frame = readFrame(obj, mask, check)
frame = obj.reader.step();
if(check)
frame(:,:,1) = frame(:, :, 1).*mask;
frame(:,:,2) = frame(:, :, 2).*mask;
frame(:,:,3) = frame(:, :, 3).*mask;
end
end
%returns a blob image (black/white)
function mask = createBlobImage(frame,obj, check)
% Detect foreground.
mask = obj.detector.step(frame);
if(check)
% Apply morphological operations to remove noise and fill in holes.
mask = imopen(mask, strel('rectangle', [3,3]));
mask = imclose(mask, strel('rectangle', [15, 15]));
mask = imfill(mask, 'holes');
end
end
%Returns the number of cars in an image using a pretrained knn model
function result = getNumberOfCars(trainingModel, frame, parts, cTotalDescriptors, detector)
if ndims(frame) == 3
frame = single(rgb2gray(frame));
else
frame = single(frame);
end
step1 = floor(size(frame,1)./parts);
step2 = floor(size(frame,2)./parts);
imggrid = [];
for i = 1:step1:size(frame,1)-step1
for j = 1:step2:size(frame,2)-step2
tmp = frame(i:i+step1-1,j:j+step2-1);
imggrid = [imggrid;getSingleBagOfWords(tmp, cTotalDescriptors, detector)];
end
end
pred = predict(trainingModel, imggrid);
result = sum(pred);
end
%Returns a bag of words histogram of an image using the given bins
function bag = getSingleBagOfWords(im, bins, detector)
[~,desc] = runDetector(im,detector);
bag = zeros(1,size(bins,2));
for j = 1:size(desc,2)
dist = [];
for k = 1:size(bins,2)
dist(k) = norm(double(desc(:,j))-bins(:,k));
end
[~,bin] = min(dist);
bag(bin) = bag(bin)+1;
end
end
function [f,d] = runDetector(image, detector)
if(strcmp(detector,'SIFT'))
if ndims(image) == 3
image = single(rgb2gray(image));
else
image = single(image);
end
[f,d] = vl_sift(image);
elseif ((strcmp(detector,'SURF')) || (strcmp(detector,'ESURF')))
if ndims(image) == 3
image = rgb2gray(image); %Do not convert to single when using SURF
end
points = detectSURFFeatures(image);
if(strcmp(detector,'SURF'))
[d, f] = extractFeatures(image, points);
elseif (strcmp(detector,'ESURF'))
[d, f] = extractFeatures(image, points, 'SURFSize',128);
end
d = d';
end
d = normc(single(d));
d(d > 0.2) = 0.2;
d = normc(d);
end