Detection person on multiple cameras

[nakornsoft]

help me please! i need make detection person multiple cameras

how to do it?

[crazylazydboy]

Hi, Were you able run this repo on a video ??

I just want to run a simple video on this but.. after I updated this self.vdo = cv2.VideoCapture('MyVideo.avi') and it still ask me for VIDEO_PATH

[nakornsoft]

@crazylazydboy

for run code python demo_yolo3_deepsort.py demo.mp4

[crazylazydboy]

Thanks @nakhonsoft , but I went the other way and commented every code with VIDEO_PATH gave my link manually to self.vdo = cv2.VideoCapture('MyVideo.avi') it works anyway..

[nakornsoft]

@crazylazydboy

import os import cv2 import time import argparse import numpy as np

from YOLOv3 import YOLOv3 from deep_sort import DeepSort from util import COLORS_10, draw_bboxes

class Detector(object): def init(self, args): self.args = args if args.display: cv2.namedWindow("test", cv2.WINDOW_NORMAL) cv2.resizeWindow("test", args.display_width, args.display_height)

    self.vdo = cv2.VideoCapture()
    self.yolo3 = YOLOv3(args.yolo_cfg, args.yolo_weights, args.yolo_names, is_xywh=True, conf_thresh=args.conf_thresh, nms_thresh=args.nms_thresh)
    self.deepsort = DeepSort(args.deepsort_checkpoint)
    self.class_names = self.yolo3.class_names

def __enter__(self):    
    self.vdo.open("test.mp4")
    self.im_width = int(self.vdo.get(cv2.CAP_PROP_FRAME_WIDTH))
    self.im_height = int(self.vdo.get(cv2.CAP_PROP_FRAME_HEIGHT))

    if self.args.save_path:
        fourcc =  cv2.VideoWriter_fourcc(*'MJPG')
        self.output = cv2.VideoWriter(self.args.save_path, fourcc, 20, (self.im_width,self.im_height))

    assert self.vdo.isOpened()
    return self

def __exit__(self, exc_type, exc_value, exc_traceback):
    if exc_type:
        print(exc_type, exc_value, exc_traceback)

def detect(self):
    while self.vdo.grab(): 
        start = time.time()
        _, ori_im = self.vdo.retrieve()
        im = cv2.cvtColor(ori_im, cv2.COLOR_BGR2RGB)
        im = ori_im
        bbox_xcycwh, cls_conf, cls_ids = self.yolo3(im)
        if bbox_xcycwh is not None:
            # select class person
            mask = cls_ids==0

            bbox_xcycwh = bbox_xcycwh[mask]
            bbox_xcycwh[:,3:] *= 1.2

            cls_conf = cls_conf[mask]
            outputs = self.deepsort.update(bbox_xcycwh, cls_conf, im)
            if len(outputs) > 0:
                bbox_xyxy = outputs[:,:4]
                identities = outputs[:,-1]
                ori_im = draw_bboxes(ori_im, bbox_xyxy, identities)

        end = time.time()
        print("time: {}s, fps: {}".format(end-start, 1/(end-start)))

        if self.args.display:
            cv2.imshow("test", ori_im)
            cv2.waitKey(1)

        if self.args.save_path:
            self.output.write(ori_im)

def parse_args(): parser = argparse.ArgumentParser() parser.add_argument("--yolo_cfg", type=str, default="YOLOv3/cfg/yolo_v3.cfg") parser.add_argument("--yolo_weights", type=str, default="YOLOv3/yolov3.weights") parser.add_argument("--yolo_names", type=str, default="YOLOv3/cfg/coco.names") parser.add_argument("--conf_thresh", type=float, default=0.5) parser.add_argument("--nms_thresh", type=float, default=0.4) parser.add_argument("--deepsort_checkpoint", type=str, default="deep_sort/deep/checkpoint/ckpt.t7") parser.add_argument("--max_dist", type=float, default=0.2) parser.add_argument("--ignore_display", dest="display", action="store_false") parser.add_argument("--display_width", type=int, default=800) parser.add_argument("--display_height", type=int, default=600) parser.add_argument("--save_path", type=str, default="demo.avi") return parser.parse_args()

if name=="main": args = parse_args() with Detector(args) as det: det.detect()

---

Add code like this

def enter(self):
self.vdo.open("test.mp4") self.im_width = int(self.vdo.get(cv2.CAP_PROP_FRAME_WIDTH)) self.im_height = int(self.vdo.get(cv2.CAP_PROP_FRAME_HEIGHT))

---

delete code like this

parser.add_argument("VIDEO_PATH", type=str)

[crazylazydboy]

Hi, @nakhonsoft Have you tried using train.py.. ? I am trying to use it. But I get the following error.

Segmentation fault (core dumped)

Do you have any idea why ??

[nakornsoft]

no, so now i stoped project i developing face recognition

next year i will develop this project continue

i have project same amazon-go

[crazylazydboy]

Hi, Don't worry.. I found the error. The issue was in the matplotlib.pyplot version. I had the latest version which was 3.1.2 I uninstalled it and installed an older version 2.0.0 It works now..

[zubairahmed-ai]

Before I try this, whats the speed you're getting on single and multiple cameras/videos @crazylazydboy ?

[crazylazydboy]

@zubairahmed-ai Depends on your environment.. Basically, 7-9 FPS @ 1080p & 11-13 FPS @ 720p (running on a 1080Ti GPU)

[zubairahmed-ai]

Thanks for the reply, I am getting that kinda FPS from this so was looking to switch to PyTorch only if it gives me better FPS, I guess I should try changing that to MobileNetV3 and see what happens

[crazylazydboy]

@zubairahmed-ai Have a try. But I highly doubt, it will be better than this. You can also try reducing resolution below 720p, if you want more speed.

[nakornsoft]

https://github.com/layumi/Person_reID_baseline_pytorch/

[petalatuy]

@nakhonsoft did you get to run this on multiple cameras?