there's no color algorithm here,just read result from csv file?

Hello sir, I want to use this code in the purpose of counting cars, Pedestrians, trucks, buses, bikes moving in the junction, so that camera video, I want the separate the based on the direction and the total count of them, can you help me out.

Thanks in advance.

how to change the position of ROI Line, and now it is horizontal, what should I do.. to make it Vertical for the same video input?

Hey. The sample video output on the project demo page is long with a lot of cars. However, the one provided in the video download link is very short. Can you post the full video which is used for detection please? @ahmetozlu

Line 38 of vehicle_detection_main.py checks for tf.version < '1.4.0' and throws an error for tensorflow v1.10+

I guess because numerically 1.10.0 is < 1.4.0

I tried running this program on raspberry pi 3 but was not able to run it.
Can anyone please help me.

Hi @ahmetozlu this code sometime detect(count) vehicle and sometime it doesn't count the vehicle
suggest some changes or any link which will help me and if you solve this issue please share the code
Thanks

Hello,

if i try to start your code with
python vehicle_detection_main.py i get this error code

python vehicle_detection_main.py File "vehicle_detection_main.py", line 93 ret, frame = cap.read() ^ TabError: inconsistent use of tabs and spaces in indentation

Heave you an idee why this happend ?

And can i use your code with my Webcam if i change the line
cap = cv2.VideoCapture('sub-1504619634606.mp4') to cap = cv2.VideoCapture(0) ?

Hi,
Can u provide me a link for published paper of Vehicle speed prediction for this project which you have done? and also provide me the original speed of the vehicles in the video for accuracy and comparison study with predicted speed?
I need to show false positive and false negative for my project for vehicle speed prediction. So please help me and kindly reply for this.

if ((bottom)< 250):
scale = 1

elif ((bottom > 250) and (bottom < 320)):
	scale = 2
else:
	isInROI = False

How did u decide the numbers 250 and 320? I think they are based on line y coordinate. But how did it get to 250 and 320?
Thanks

How do I cancel the color prediction

sometimes the result is NAN. why so?

I already tested the code. Although, I'm not getting any error, there is not any detected vehicle in the output. I tried object detection API only (not your code package) to see what the problem is; many vehicles were detected for the same input and same model (ssd-mobilenet). After some investigations, I found out that the variable 'is_vehicle_detected" in 'predict_speed' function is replaced by 1. I thought it should be increased one be one, not to be substituted by a constant. Also the variable ‘bottom-bottom_position_of_detected_vehicle’ is not initialized.
Moreover, in 'visualization_utils.py', the function '
draw_bounding_box_on_image' is supposed to return the output image, however, it updates the is_vehicle_detected variable as well as csv file.
I'm not sure on how to fix this. My python version 3.6. Did you get the demo output in

Hi! I have integrated the object Counting API for the model I have trained on my own custom dataset. After integrating the vehicle counting/object counting api, the customs classes (bikes , cars , bus , rickshaws) are being detected but they are not being counted separately. On video, I am getting total count of all vehicles only. Can you please tell me how to get commulative counts of each vehicle type as they pass ROI ? I urgently need help in this regard. Thanks in advance.

As per the snippet from the documentation, visualize_boxes_and_labels_on_image_array function accepts arguments which are completely different from your own arguments. Also it returns the image as you can see below, then how are you supposed to collect the 'counter' and 'csv line' from that fucntion?

 Returns:

    uint8 numpy array with shape (img_height, img_width, 3) with overlaid boxes

I modified the code as per the arguments but now I am not able to find the count when the object crosses ROI. Please help.
MODIFIED SNIPPET

counter = len(boxes)
                print("Cureent running counter: ",counter)

                # Visualization of the results of a detection.

                input_frame = vis_util.visualize_boxes_and_labels_on_image_array(

                    input_frame,

                    np.squeeze(boxes),

                    np.squeeze(classes).astype(np.int32),

                    np.squeeze(scores),

                    category_index,

                    use_normalized_coordinates=True,

                    line_thickness=4)

How can i be able to get the vehicle in and vehicle out count?

hi

I want to detect stop sign alone, which has the id: 13

So how to change that in util function?

in the below example images,

1. first white car detected
   2.second not car not detected
   3.again third car detected ..
   i want to count all the vehicle.

Thanks for your explanation, can you please explain how you are calculating the speed in this block ?
Because I can't understand some of the transformations like scale_real_length, scale_real_time_passed, speed = speed/6*40 and scale = 1
please elaborate a bit.

temp, counter, csv_line = vis_util.visualize_boxes_and_labels_on_image_array(cap.get(1),
image_np,
np.squeeze(boxes),
np.squeeze(classes).astype(np.int32),
np.squeeze(scores),
category_index,
use_normalized_coordinates=True,
line_thickness=4)
This is some part of my code. But i am getting the following error.

KeyError Traceback (most recent call last)
in ()
38 category_index,
39 use_normalized_coordinates=True,
---> 40 line_thickness=4)
41 print("-*-GELEN COUNTER: " + str(counter))
42 temp_counter = temp_counter + counter

~/video-analytics/models/object_detection/utils/visualization_utils.py in visualize_boxes_and_labels_on_image_array(image, boxes, classes, scores, category_index, instance_masks, keypoints, use_normalized_coordinates, max_boxes_to_draw, min_score_thresh, agnostic_mode, line_thickness)
443 box = tuple(boxes[i].tolist())
444 if instance_masks is not None:
--> 445 box_to_instance_masks_map[box] = instance_masks[i]
446 if keypoints is not None:
447 box_to_keypoints_map[box].extend(keypoints[i])

KeyError: 0

Hello Sir
Is this source code estimate speed of multiple vehicles?

Thanks and Regards
Ganesh Koushik

Hi ,
1.I need to run this source code Windows OS.can you please send me the procedure for run this source code to windows
2.Is it possible open Visual Studio 2015?

RESTART: C:\tensorflow\models-master\research\vehicle_counting_tensorflow\vehicle_detection_main.py
Traceback (most recent call last):
File "C:\tensorflow\models-master\research\vehicle_counting_tensorflow\vehicle_detection_main.py", line 27, in
from utils import visualization_utils as vis_util
File "C:\tensorflow\models-master\research\vehicle_counting_tensorflow\utils\visualization_utils.py", line 30, in
from image_utils import image_saver
ImportError: cannot import name 'image_saver'

=====

.

RESTART: C:\tensorflow\models-master\research\vehicle_counting_tensorflow\utils\speed_and_direction_prediction_module\speed_prediction.py
Traceback (most recent call last):
File "C:\tensorflow\models-master\research\vehicle_counting_tensorflow\utils\speed_and_direction_prediction_module\speed_prediction.py", line 1, in
from utils.image_utils import image_saver
ModuleNotFoundError: No module named 'utils'

=======
utils is folder inside but while complie it error like above.
How to fix this? Please help

Hi. I want to run my own video. What to change and how to change in the code so as to run my own vehicle traffic video.
Thanks in advance!!

I want to use this program with another traffic video but can't find the code where the coordinates of roi line is defined. I tried changin the variable in visualization file that was previously set to 200 but it didn't work. Can you please help me with this?

What does "bottom" and "bottom_position_of_detected_vehicle" variable mean? I get so confused. Does bottom refer to the bottom coordinates of the current object then the bottom_position_of_detected_vehicle is the previous bottom position of the previous vehicle?

Can you brief me on what I have to do, so make the script compatible with CUDA 10.

Hi every one,
This works fine for single camera, but how i can implement multiple cameras for vehicles detection? . i googled a lot but couldn't find the practical approach for it.

can any one can help me, please!!!

what is the difference between these parameters, as

1. cumulative_object_counting_x_axis only does people counting?
2. cumulative_object_counting_y_axis only does vehicle counting?
3. object_counting detects all objects..?

If, yes what is meaning of parameters we are passing here..?

How do we count number of vehicles when they pass ROI area? Guess in current code bottom coordinates ymax is checked with 200 - but in subsequent frames this will be same since video moves fast - how is duplicate count eliminated? Thanks

if You need help let me know.Otherwise mark this issue as question.

Code is working fine but not getting any output.

Hi sir its me again :)
can you teach me how to get size of object and how to add new vehicle like motorcycle and big truck.
thank you for your email respons

why tracking algorithm is not required?
I am to count the number of people entering a vehicle. How can I provide the ROI line so that people leaving or entering past that line will be counted?

thank you

Hi,
Thank you so much for providing this repo.
I'm running your code in a virtual machine which I only have access to the terminal, so I decided to save each frame instead of displaying it. I get this error when running the code
error while decoding MB 4 6, bytestream -9
The code stops after processing first frame of video file. I'd be thankful if you kindly help on that.

Thanks,

Bahareh

File "D:/ObjectDetect_Project/vehicle_counting_tensorflow/vehicle_detection.py", line 19, in
from utils import label_map_util

ImportError: cannot import name 'label_map_util'

how can i solve this import error ?

I have a trained model of inception v3, retrained on the class of Vehicles. I want to use that instead of the by default setting here. How can i do that ?

i tried it on faster_rcnn_resnet101_coco_2018_01_28 and faster_rcnn_inception_v2_coco_2018_01_28 but both of them results in slow output and not able to detect speed but the already present model in this project works fine.Please help me @ahmetozlu

Hi, how can i set the coordinates for a new video.
And my output isn't coming for a different video

Hi guys
we know that this model (ssd_mobilenet_v1_coco_2017_11_17) is general object detection problem which changes we perform to detect only vehicle like in this project
Thanks

I am having trouble running the code without fixing identation which is a mix of tab and space usages

Dear sir,
I'm having a question.
Which of the following is more efficient?and why?

Applying the detection-tracking part on sequential images OR a video?

If we apply the pipeline on series of images, the detected bounding box needs to be saved within running of pipeline on each frame, however when running on video, it seems that it performs better by utilizing the functions in moviepy library. What do you think?

I am trying to use my own video which is 30 frames per second and resolution is of 1920*1080. I tried to change the ROI value from 200 to different values but it didnt work.

Can you please guide me on the same??

I would also like to know where to update the frames per second value.

Looking forward to hearing from you.

When I run the code then it gives me the error....

Traceback (most recent call last):

File "", line 1, in
runfile('C:/Users/mark/CarSpeedDetection/vehicle_detection_main.py', wdir='C:/Users/mark/CarSpeedDetection')

File "C:\Users\Administrator\Anaconda3\lib\site-packages\spyder\utils\site\sitecustomize.py", line 705, in runfile
execfile(filename, namespace)

File "C:\Users\Administrator\Anaconda3\lib\site-packages\spyder\utils\site\sitecustomize.py", line 102, in execfile
exec(compile(f.read(), filename, 'exec'), namespace)

File "C:/Users/mark/CarSpeedDetection/vehicle_detection_main.py", line 124, in
object_detection_function()

File "C:/Users/mark/CarSpeedDetection/vehicle_detection_main.py", line 91, in object_detection_function
np.squeeze(scores), category_index, use_normalized_coordinates=True, line_thickness=4)

File "C:\Users\mark\CarSpeedDetection\utils\visualization_utils.py", line 304, in visualize_boxes_and_labels_on_image_array
use_normalized_coordinates=use_normalized_coordinates)

File "C:\Users\mark\CarSpeedDetection\utils\visualization_utils.py", line 74, in draw_bounding_box_on_image_array
display_str_list, use_normalized_coordinates)

File "C:\Users\mark\CarSpeedDetection\utils\visualization_utils.py", line 103, in draw_bounding_box_on_image
current_path + "/utils/color_recognition_module/" + "test.data")

File "./utils\color_recognition_module\knn_classifier.py", line 64, in main
neighbors = kNearestNeighbors(training_feature_vector, test_feature_vector[x], k)

File "./utils\color_recognition_module\knn_classifier.py", line 24, in kNearestNeighbors
neighbors.append(distances[x][0])

IndexError: list index out of range