Trying to save output file but it saves as directory , not file

[Stevew7777]

Hi

I'm trying to get YOLOv5 to save its processed output as a single file in a specific location that I want
e.g. /tmp/output/myoutput.jpg , but instead it saves it as a directory called /tmp/output/myoutput.jpg/ with an output file called myoutput.jpg

Not sure what I'm doing wrong?

Running YOLOv5 I use this command :

python /usr/src/app/detect.py --weights /usr/src/app/runs/train/exp7/weights/last.pt --source /tmp/input/input.jpg --name /tmp/output/myoutput.jpg --exist-ok

The config inside detect.py is :

def run(
weights=ROOT / 'yolov5s.pt', # model path or triton URL
source=ROOT / 'data/images', # file/dir/URL/glob/screen/0(webcam)
data=ROOT / 'data/coco128.yaml', # dataset.yaml path
imgsz=(640, 640), # inference size (height, width)
conf_thres=0.25, # confidence threshold
iou_thres=0.45, # NMS IOU threshold
max_det=1000, # maximum detections per image
device='', # cuda device, i.e. 0 or 0,1,2,3 or cpu
view_img=False, # show results
save_txt=False, # save results to *.txt
save_conf=False, # save confidences in --save-txt labels
save_crop=False, # save cropped prediction boxes
nosave=False, # do not save images/videos
classes=None, # filter by class: --class 0, or --class 0 2 3
agnostic_nms=False, # class-agnostic NMS
augment=False, # augmented inference
visualize=False, # visualize features
update=False, # update all models
project=ROOT / '/tmp/output/', # save results to project/name
name='', # save results to project/name
exist_ok=False, # existing project/name ok, do not increment
line_thickness=3, # bounding box thickness (pixels)
hide_labels=False, # hide labels
hide_conf=False, # hide confidences
half=False, # use FP16 half-precision inference
dnn=False, # use OpenCV DNN for ONNX inference
vid_stride=1, # video frame-rate stride

def parse_opt():
parser = argparse.ArgumentParser()
parser.add_argument('--weights', nargs='+', type=str, default=ROOT / 'yolov5s.pt', help='model path or triton URL')
parser.add_argument('--source', type=str, default=ROOT / 'data/images', help='file/dir/URL/glob/screen/0(webcam)')
parser.add_argument('--data', type=str, default=ROOT / 'data/coco128.yaml', help='(optional) dataset.yaml path')
parser.add_argument('--imgsz', '--img', '--img-size', nargs='+', type=int, default=[640], help='inference size h,w')
parser.add_argument('--conf-thres', type=float, default=0.25, help='confidence threshold')
parser.add_argument('--iou-thres', type=float, default=0.45, help='NMS IoU threshold')
parser.add_argument('--max-det', type=int, default=1000, help='maximum detections per image')
parser.add_argument('--device', default='', help='cuda device, i.e. 0 or 0,1,2,3 or cpu')
parser.add_argument('--view-img', action='store_true', help='show results')
parser.add_argument('--save-txt', action='store_true', help='save results to *.txt')
parser.add_argument('--save-conf', action='store_true', help='save confidences in --save-txt labels')
parser.add_argument('--save-crop', action='store_true', help='save cropped prediction boxes')
parser.add_argument('--nosave', action='store_true', help='do not save images/videos')
parser.add_argument('--classes', nargs='+', type=int, help='filter by class: --classes 0, or --classes 0 2 3')
parser.add_argument('--agnostic-nms', action='store_true', help='class-agnostic NMS')
parser.add_argument('--augment', action='store_true', help='augmented inference')
parser.add_argument('--visualize', action='store_true', help='visualize features')
parser.add_argument('--update', action='store_true', help='update all models')
parser.add_argument('--project', default=ROOT / '/tmp/docker_upload', help='save results to project/name')
parser.add_argument('--name', default='', help='save results to project/name')
parser.add_argument('--exist-ok', action='store_true', help='existing project/name ok, do not increment')
parser.add_argument('--line-thickness', default=3, type=int, help='bounding box thickness (pixels)')
parser.add_argument('--hide-labels', default=False, action='store_true', help='hide labels')
parser.add_argument('--hide-conf', default=False, action='store_true', help='hide confidences')
parser.add_argument('--half', action='store_true', help='use FP16 half-precision inference')
parser.add_argument('--dnn', action='store_true', help='use OpenCV DNN for ONNX inference')
parser.add_argument('--vid-stride', type=int, default=1, help='video frame-rate stride')
opt = parser.parse_args()
opt.imgsz *= 2 if len(opt.imgsz) == 1 else 1 # expand
print_args(vars(opt))
return opt

Any help appreciated. :-)

[Stevew7777]

Anyone? This cant be too hard?

[OJJJN]

maybe try this, you need to improve some code:

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

model = 'best.onnx'
img_w = 640
img_h = 640
classes_file = 'classes.txt'

def class_names():
classes = []
with open(classes_file, 'r') as file:
for line in file:
name = line.strip('\n')
classes.append(name)
return classes

width_frame = 640
net = cv2.dnn.readNetFromONNX(model)
classes = class_names()

Dahua CCTV camera settings

rtsp_url = '' # Replace with your Dahua CCTV camera RTSP URL

cap = cv2.VideoCapture(rtsp_url)

total_slots = 15
cars_count = 0
empty_slots = 0
last_detection_time = datetime.datetime.now()

Specify the directory path to save the image results

output_directory = r'D:\Database'

while True:
current_time = datetime.datetime.now()
time_difference = current_time - last_detection_time

if time_difference.total_seconds() >= (15):
    detection_start_time = datetime.datetime.now()
    detection_end_time = detection_start_time + datetime.timedelta(seconds=5)
    last_detection_time = detection_end_time

    while datetime.datetime.now() <= detection_end_time:
        ret, frame = cap.read()

        if not ret:
            print("Failed to capture frame")
            break

        height = int(frame.shape[0] * (width_frame / frame.shape[1]))
        dim = (width_frame, height)
        img = cv2.resize(frame, dim, interpolation=cv2.INTER_AREA)

        blob = cv2.dnn.blobFromImage(img, 1/255, (img_w, img_h), swapRB=True, mean=(0, 0, 0), crop=False)
        net.setInput(blob)
        t1 = time.time()
        outputs = net.forward(net.getUnconnectedOutLayersNames())
        t2 = time.time()
        out = outputs[0]
        n_detections = out.shape[1]
        height, width = img.shape[:2]
        x_scale = width / img_w
        y_scale = height / img_h
        conf_threshold = 0.7
        score_threshold = 0.5
        nms_threshold = 0.5

        class_ids = []
        score = []
        boxes = []

        for i in range(n_detections):
            detect = out[0][i]
            confidence = detect[4]
            if confidence >= conf_threshold:
                class_score = detect[5:]
                class_id = np.argmax(class_score)
                if class_score[class_id] > score_threshold:
                    score.append(confidence)
                    class_ids.append(class_id)
                    x, y, w, h = detect[0], detect[1], detect[2], detect[3]
                    left = int((x - w/2) * x_scale)
                    top = int((y - h/2) * y_scale)
                    width = int(w * x_scale)
                    height = int(h * y_scale)
                    box = np.array([left, top, width, height])
                    boxes.append(box)

        indices = cv2.dnn.NMSBoxes(boxes, np.array(score), conf_threshold, nms_threshold)

        cars_count = len(indices)
        empty_slots = total_slots - cars_count

        for i in indices:
            box = boxes[i]
            left = box[0]
            top = box[1]
            width = box[2]
            height = box[3]

            left, top, width, height = boxes[i]

            cv2.rectangle(img, (left, top), (left + width, top + height), (0, 0, 255), 2)
            label = "{}".format(classes[class_ids[i]])
            text_size = cv2.getTextSize(label, cv2.FONT_HERSHEY_SIMPLEX, 0.5, 1)
            dim, baseline = text_size[0], text_size[1]
            cv2.rectangle(img, (left, top - 20), (left + dim[0], top + dim[1] + baseline - 20), (0, 0, 0), cv2.FILLED)
            cv2.putText(img, label, (left, top + dim[1] - 20), cv2.FONT_HERSHEY_SIMPLEX, 0.5, (255, 255, 0), 1, cv2.LINE_AA)

        # Display car count and empty slots on theframe
        cv2.putText(img, f"Cars Count: {cars_count}", (10, 30), cv2.FONT_HERSHEY_SIMPLEX, 0.8, (255, 255, 255), 2)[0]
        cv2.putText(img, f"Empty Slots: {empty_slots}", (10, 70), cv2.FONT_HERSHEY_SIMPLEX, 0.8, (255, 255, 255), 2)[0]

        # Save the image results
        filename = f"result_{datetime.datetime.now().strftime('%Y%m%d_%H%M%S')}.jpg"
        output_path = os.path.join(output_directory, filename)
        cv2.imwrite(output_path, img)

        # Display the captured frame
        cv2.imshow("Object Detection", img)

        # Check for 'q' key press to exit
        if cv2.waitKey(1) & 0xFF == ord('q'):
            break

else:
    ret, frame = cap.read()

    if not ret:
        print("Failed to capture frame")
        break

    # Display the captured frame
    cv2.imshow("Camera Feed", frame)

    # Check for 'q' key press to exit
    if cv2.waitKey(1) & 0xFF == ord('q'):
        break

Release the video capture and destroy all windows

cap.release()
cv2.destroyAllWindows()

this is using opencv and onnx file, you can try to change the onnx required or something. also this code saving an image with date and time name, so you know when the object got detect.

[OJJJN]

this code will run every 15 sec, so it not consume so many cpu or gpu