yolov5 deepsort pedestrian/vehicle (detection + counting + tracking + ranging + speed measurement)

Function introduction

• Realized the separate counting of the local area's exit/entry.
• Display the detection category and the number of IDs.
• The default is south/north direction detection, if you want to detect different positions and directions, you need to modify it
• Click to run in count_car/traffic.py
• Default detection categories: pedestrian, bicycle, car, motorcycle, bus, truck, boat.
• The detection class can be modified in the objdetector.py file.

code run

$ git clone 追踪代码

Therefore, the repo contains files such as weights and mp4. If the speed of git clone is slow, you can directly download the zip

enter directory

$ cd unbox_yolov5_deepsort_counting

Create a python virtual environment

$ python3 -m venv venv

Activate the virtual environment

$ source venv/bin/activate

upgrade pip

$ python -m pip install --upgrade pip

Install pytorch

According to your operating system, installation tool and CUDA version, find the corresponding installation command at https://pytorch.org/get-started/locally/. My environment is ubuntu 18.04.5, pip, CUDA 11.0. ![Insert picture description here](https://img-blog.csdnimg.cn/77d60150764741b0bafa0927e1042a46.png

$ pip install torch==1.7.1+cu110 torchvision==0.8.2+cu110 torchaudio===0.7.2 -f https://download.pytorch.org/whl/torch_stable.html

install package

$ pip install -r requirements.txt

In the demo.py file, set the video file path to be detected, the default is'./video/test.mp4'

capture = cv2.VideoCapture(‘./video/test.mp4’)

run the program

python count.py

demo code

detector = Detector()

    # 打开视频
    capture = cv2.VideoCapture(VIDEO_PATH)

    while True:
        # 读取每帧图片
        _, im = capture.read()
        if im is None:
            break

        # 缩小尺寸
        im = cv2.resize(im, (width//2, height//2))

        list_bboxs = []
        # 更新跟踪器
        output_image_frame, list_bboxs = objtracker.update(detector, im)
        # 输出图片
        output_image_frame = cv2.add(output_image_frame, color_polygons_image)

        if len(list_bboxs) > 0:
            # ----------------------判断撞线----------------------
            for item_bbox in list_bboxs:
                x1, y1, x2, y2, _, track_id = item_bbox
                # 撞线检测点，(x1，y1)，y方向偏移比例 0.0~1.0
                y1_offset = int(y1 + ((y2 - y1) * 0.6))
                # 撞线的点
                y = y1_offset
                x = x1
                if polygon_mask_blue_and_yellow[y, x] == 1:
                    # 如果撞 蓝polygon
                    if track_id not in list_overlapping_blue_polygon:
                        list_overlapping_blue_polygon.append(track_id)
                    # 判断 黄polygon list里是否有此 track_id
                    # 有此track_id，则认为是 UP (上行)方向
                    if track_id in list_overlapping_yellow_polygon:
                        # 上行+1
                        up_count += 1
                        print('up count:', up_count, ', up id:', list_overlapping_yellow_polygon)
                        # 删除 黄polygon list 中的此id
                        list_overlapping_yellow_polygon.remove(track_id)

                elif polygon_mask_blue_and_yellow[y, x] == 2:
                    # 如果撞 黄polygon
                    if track_id not in list_overlapping_yellow_polygon:
                        list_overlapping_yellow_polygon.append(track_id)
                    # 判断 蓝polygon list 里是否有此 track_id
                    # 有此 track_id，则 认为是 DOWN（下行）方向
                    if track_id in list_overlapping_blue_polygon:
                        # 下行+1
                        down_count += 1
                        print('down count:', down_count, ', down id:', list_overlapping_blue_polygon)
                        # 删除 蓝polygon list 中的此id
                        list_overlapping_blue_polygon.remove(track_id)
            # ----------------------清除无用id----------------------
            list_overlapping_all = list_overlapping_yellow_polygon + list_overlapping_blue_polygon
            for id1 in list_overlapping_all:
                is_found = False
                for _, _, _, _, _, bbox_id in list_bboxs:
                    if bbox_id == id1:
                        is_found = True
                if not is_found:
                    # 如果没找到，删除id
                    if id1 in list_overlapping_yellow_polygon:
                        list_overlapping_yellow_polygon.remove(id1)

                    if id1 in list_overlapping_blue_polygon:
                        list_overlapping_blue_polygon.remove(id1)
            list_overlapping_all.clear()
            # 清空list
            list_bboxs.clear()
        else:
            # 如果图像中没有任何的bbox，则清空list
            list_overlapping_blue_polygon.clear()
            list_overlapping_yellow_polygon.clear()
            
        # 输出计数信息
        text_draw = 'DOWN: ' + str(down_count) + \
                    ' , UP: ' + str(up_count)
        output_image_frame = cv2.putText(img=output_image_frame, text=text_draw,
                                         org=draw_text_postion,
                                         fontFace=font_draw_number,
                                         fontScale=0.75, color=(0, 0, 255), thickness=2)
        cv2.imshow('Counting Demo', output_image_frame)
        cv2.waitKey(1)

    capture.release()
    cv2.destroyAllWindows()

Result display

Various tracking, ranging, attitude estimation, target detection, counting, and speed measurement functions have been realized, welcome to communicate!
See the homepage for more projects!