|
@@ -1,154 +1,155 @@
|
|
|
-import cv2
|
|
|
-import argparse
|
|
|
-from ultralytics import YOLO
|
|
|
-import supervision as sv
|
|
|
-import numpy as np
|
|
|
-import requests
|
|
|
-import time
|
|
|
-import pandas as pd
|
|
|
-import os
|
|
|
-from ultralytics.yolo.utils.plotting import Annotator, colors, save_one_box
|
|
|
-
|
|
|
-ZONE_POLYGON = np.array([
|
|
|
- [0, 0],
|
|
|
- [0.3, 0],
|
|
|
- [0.3, 1],
|
|
|
- [0, 1]
|
|
|
-])
|
|
|
-
|
|
|
-
|
|
|
-def parse_arguments() -> argparse.Namespace:
|
|
|
- parser = argparse.ArgumentParser(description="YOLOv8 live")
|
|
|
- parser.add_argument(
|
|
|
- "--webcam-resolution",
|
|
|
- default=[800, 600],
|
|
|
- nargs=2,
|
|
|
- type=int
|
|
|
- )
|
|
|
- args = parser.parse_args()
|
|
|
- return args
|
|
|
-
|
|
|
-def main():
|
|
|
- args = parse_arguments()
|
|
|
- frame_width, frame_height = args.webcam_resolution
|
|
|
- # print(frame_width,frame_height)
|
|
|
- cap = cv2.VideoCapture(0)
|
|
|
- cap.set(cv2.CAP_PROP_FRAME_WIDTH, frame_width)
|
|
|
- cap.set(cv2.CAP_PROP_FRAME_HEIGHT, frame_height)
|
|
|
-
|
|
|
- # model = YOLO("yolov8n.pt")
|
|
|
- #使用模組
|
|
|
- model = YOLO("best.pt")
|
|
|
-
|
|
|
- #設定辨識方框參數
|
|
|
- box_annotator = sv.BoxAnnotator(
|
|
|
- thickness=2,
|
|
|
- text_thickness=2,
|
|
|
- text_scale=1
|
|
|
- )
|
|
|
-
|
|
|
- #設定區塊參數
|
|
|
- zone_polygon = (ZONE_POLYGON * np.array(args.webcam_resolution)).astype(int)
|
|
|
- zone = sv.PolygonZone(polygon=zone_polygon, frame_resolution_wh=tuple(args.webcam_resolution))
|
|
|
- zone_annotator = sv.PolygonZoneAnnotator(
|
|
|
- zone=zone,
|
|
|
- color=sv.Color.blue(),
|
|
|
- thickness=2,
|
|
|
- text_thickness=4,
|
|
|
- text_scale=2
|
|
|
- )
|
|
|
- # count 計算數量
|
|
|
- # FPS_count 計算FPS
|
|
|
- # start_time 用於計算FPS時間
|
|
|
- count = 0
|
|
|
- FPS_count = 0
|
|
|
- font = cv2.FONT_HERSHEY_SIMPLEX
|
|
|
- color = (255,0,0)
|
|
|
-
|
|
|
- start_time = time.time()
|
|
|
-
|
|
|
- while True:
|
|
|
- ret, frame = cap.read()
|
|
|
- result = model(frame, agnostic_nms=True,save_crop=False,save_conf=False)[0]
|
|
|
- detections = sv.Detections.from_yolov8(result)
|
|
|
-
|
|
|
- #辨識到的名稱與準確度
|
|
|
- labels = [
|
|
|
- f"{model.model.names[class_id]} {confidence:0.2f}"
|
|
|
- for _, confidence, class_id, _
|
|
|
- in detections
|
|
|
- ]
|
|
|
-
|
|
|
- #辨識名稱
|
|
|
- labels_name = [
|
|
|
- f"{model.model.names[class_id]}"
|
|
|
- for _, confidence, class_id, _
|
|
|
- in detections
|
|
|
- ]
|
|
|
- #辨識準確度
|
|
|
- labels_confidence = [
|
|
|
- f"{confidence:0.2f}"
|
|
|
- for _, confidence, class_id, _
|
|
|
- in detections
|
|
|
- ]
|
|
|
-
|
|
|
- #畫面顯示辨識框
|
|
|
- frame = box_annotator.annotate(
|
|
|
- scene=frame,
|
|
|
- detections=detections,
|
|
|
- labels=labels
|
|
|
- )
|
|
|
-
|
|
|
- #抓取辨識框的資料
|
|
|
- boxes_confidence = result.boxes.conf
|
|
|
- boxes_confidence = boxes_confidence * 100
|
|
|
- print(boxes_confidence)
|
|
|
-
|
|
|
- #顯示區塊內結果
|
|
|
- mask = zone.trigger(detections=detections)
|
|
|
-
|
|
|
- # 區塊內有辨識到會寫True,判斷等於'Fatwolf'時,判斷辨識框的準確度是否大於等於50,
|
|
|
- # 如果是就計算次數+1且把辨識框內的圖片擷取下來並記錄到txt。
|
|
|
- if mask.any() == True:
|
|
|
- if labels_name == 0:
|
|
|
- continue
|
|
|
- elif labels_name == ['Fatwolf']:
|
|
|
- print('fatwolf')
|
|
|
- #count += 1
|
|
|
- #x1,y1,x2,y2辨識框的座標
|
|
|
- x1 = result.boxes.xyxy[0][0]
|
|
|
- y1 = result.boxes.xyxy[0][1]
|
|
|
- x2 = result.boxes.xyxy[0][2]
|
|
|
- y2 = result.boxes.xyxy[0][3]
|
|
|
- if int(boxes_confidence) >= 50:
|
|
|
- count += 1
|
|
|
- roi2 = frame[int(y1) + 4:int(y2) - 2, int(x1) + 4:int(x2) - 2]
|
|
|
- now_time = time.strftime("%Y-%m-%d-%H-%M-%S", time.localtime(time.time()))
|
|
|
- save_pic_name = now_time+'_'+str(count) + '.jpg'
|
|
|
- print("存圖片:",save_pic_name, "可信度:",int(boxes_confidence))
|
|
|
- cv2.imwrite(save_pic_name, roi2)
|
|
|
- path = 'output.txt'
|
|
|
- with open(path, 'a') as f:
|
|
|
- f.write(now_time+'_'+str(count) + '.jpg'+' 可信度:'+str(boxes_confidence)+'\n')
|
|
|
- else:
|
|
|
- continue
|
|
|
- elif labels_name == ['Bottle']:
|
|
|
- print('Bottle')
|
|
|
- # 畫面添加區塊顯示
|
|
|
- frame = zone_annotator.annotate(scene=frame)
|
|
|
- # 即時顯示計算區塊內辨識到的次數
|
|
|
- cv2.putText(frame, 'Count: {}'.format(count), (10, 50), font, 1, color, 2, cv2.LINE_AA)
|
|
|
- # 計算FPS實際幀數並即時顯示
|
|
|
- FPS_count += 1
|
|
|
- fps = FPS_count / (time.time() - start_time)
|
|
|
- cv2.putText(frame, 'FPS: {:.2f}'.format(fps), (frame.shape[1]-200, 50), font, 1, color, 2, cv2.LINE_AA)
|
|
|
-
|
|
|
- cv2.imshow("yolov8", frame)
|
|
|
-
|
|
|
- key = cv2.waitKey(1)
|
|
|
- if key == ord('q'):
|
|
|
- break
|
|
|
-
|
|
|
-
|
|
|
-if __name__ == "__main__":
|
|
|
- main()
|
|
|
+import cv2
|
|
|
+import argparse
|
|
|
+from ultralytics import YOLO
|
|
|
+import supervision as sv
|
|
|
+import numpy as np
|
|
|
+import requests
|
|
|
+import time
|
|
|
+import pandas as pd
|
|
|
+import os
|
|
|
+from ultralytics.yolo.utils.plotting import Annotator, colors, save_one_box
|
|
|
+
|
|
|
+ZONE_POLYGON = np.array([
|
|
|
+ [0, 0],
|
|
|
+ [0.3, 0],
|
|
|
+ [0.3, 1],
|
|
|
+ [0, 1]
|
|
|
+])
|
|
|
+
|
|
|
+
|
|
|
+def parse_arguments() -> argparse.Namespace:
|
|
|
+ parser = argparse.ArgumentParser(description="YOLOv8 live")
|
|
|
+ parser.add_argument(
|
|
|
+ "--webcam-resolution",
|
|
|
+ default=[800, 600],
|
|
|
+ nargs=2,
|
|
|
+ type=int
|
|
|
+ )
|
|
|
+ args = parser.parse_args()
|
|
|
+ return args
|
|
|
+
|
|
|
+def main():
|
|
|
+ args = parse_arguments()
|
|
|
+ frame_width, frame_height = args.webcam_resolution
|
|
|
+ # print(frame_width,frame_height)
|
|
|
+ cap = cv2.VideoCapture(0)
|
|
|
+ cap.set(cv2.CAP_PROP_FRAME_WIDTH, frame_width)
|
|
|
+ cap.set(cv2.CAP_PROP_FRAME_HEIGHT, frame_height)
|
|
|
+
|
|
|
+ # model = YOLO("yolov8n.pt")
|
|
|
+ #使用模組
|
|
|
+ model = YOLO("best.pt")
|
|
|
+
|
|
|
+ #設定辨識方框參數
|
|
|
+ box_annotator = sv.BoxAnnotator(
|
|
|
+ thickness=2,
|
|
|
+ text_thickness=2,
|
|
|
+ text_scale=1
|
|
|
+ )
|
|
|
+
|
|
|
+ #設定區塊參數
|
|
|
+ zone_polygon = (ZONE_POLYGON * np.array(args.webcam_resolution)).astype(int)
|
|
|
+ zone = sv.PolygonZone(polygon=zone_polygon, frame_resolution_wh=tuple(args.webcam_resolution))
|
|
|
+ zone_annotator = sv.PolygonZoneAnnotator(
|
|
|
+ zone=zone,
|
|
|
+ color=sv.Color.blue(),
|
|
|
+ thickness=2,
|
|
|
+ text_thickness=4,
|
|
|
+ text_scale=2
|
|
|
+ )
|
|
|
+ # count 計算數量
|
|
|
+ # FPS_count 計算FPS
|
|
|
+ # start_time 用於計算FPS時間
|
|
|
+ count = 0
|
|
|
+ FPS_count = 0
|
|
|
+ font = cv2.FONT_HERSHEY_SIMPLEX
|
|
|
+ color = (255,0,0)
|
|
|
+
|
|
|
+ start_time = time.time()
|
|
|
+
|
|
|
+ while True:
|
|
|
+ ret, frame = cap.read()
|
|
|
+ result = model(frame, agnostic_nms=True,save_crop=False,save_conf=False)[0]
|
|
|
+ detections = sv.Detections.from_yolov8(result)
|
|
|
+
|
|
|
+ #辨識到的名稱與準確度
|
|
|
+ labels = [
|
|
|
+ f"{model.model.names[class_id]} {confidence:0.2f}"
|
|
|
+ for _, confidence, class_id, _
|
|
|
+ in detections
|
|
|
+ ]
|
|
|
+
|
|
|
+ #辨識名稱
|
|
|
+ labels_name = [
|
|
|
+ f"{model.model.names[class_id]}"
|
|
|
+ for _, confidence, class_id, _
|
|
|
+ in detections
|
|
|
+ ]
|
|
|
+ #辨識準確度
|
|
|
+ labels_confidence = [
|
|
|
+ f"{confidence:0.2f}"
|
|
|
+ for _, confidence, class_id, _
|
|
|
+ in detections
|
|
|
+ ]
|
|
|
+
|
|
|
+ #畫面顯示辨識框
|
|
|
+ frame = box_annotator.annotate(
|
|
|
+ scene=frame,
|
|
|
+ detections=detections,
|
|
|
+ labels=labels
|
|
|
+ )
|
|
|
+
|
|
|
+ #抓取辨識框的資料
|
|
|
+ boxes_confidence = result.boxes.conf
|
|
|
+ boxes_confidence = boxes_confidence * 100
|
|
|
+ print(boxes_confidence)
|
|
|
+
|
|
|
+ #顯示區塊內結果
|
|
|
+ mask = zone.trigger(detections=detections)
|
|
|
+
|
|
|
+ # 區塊內有辨識到會寫True,判斷等於'Fatwolf'時,判斷辨識框的準確度是否大於等於50,
|
|
|
+ # 如果是就計算次數+1且把辨識框內的圖片擷取下來並記錄到txt。
|
|
|
+ if mask.any() == True:
|
|
|
+ if labels_name == 0:
|
|
|
+ continue
|
|
|
+ elif labels_name == ['Fatwolf']:
|
|
|
+ print('fatwolf')
|
|
|
+ #count += 1
|
|
|
+ #x1,y1,x2,y2辨識框的座標
|
|
|
+ x1 = result.boxes.xyxy[0][0]
|
|
|
+ y1 = result.boxes.xyxy[0][1]
|
|
|
+ x2 = result.boxes.xyxy[0][2]
|
|
|
+ y2 = result.boxes.xyxy[0][3]
|
|
|
+ if int(boxes_confidence) >= 50:
|
|
|
+ count += 1
|
|
|
+ roi2 = frame[int(y1) + 4:int(y2) - 2, int(x1) + 4:int(x2) - 2]
|
|
|
+ now_time = time.strftime("%Y-%m-%d-%H-%M-%S", time.localtime(time.time()))
|
|
|
+ save_pic_name = now_time+'_'+str(count) + '.jpg'
|
|
|
+ print("存圖片:",save_pic_name, "可信度:",int(boxes_confidence))
|
|
|
+ cv2.imwrite(save_pic_name, roi2)
|
|
|
+ path = 'output.txt'
|
|
|
+ with open(path, 'a') as f:
|
|
|
+ f.write(now_time+'_'+str(count) + '.jpg'+' 可信度:'+str(int(boxes_confidence))+'\n')
|
|
|
+ f.write('數量:'+str(count)+'\n')
|
|
|
+ else:
|
|
|
+ continue
|
|
|
+ elif labels_name == ['Bottle']:
|
|
|
+ print('Bottle')
|
|
|
+ # 畫面添加區塊顯示
|
|
|
+ frame = zone_annotator.annotate(scene=frame)
|
|
|
+ # 即時顯示計算區塊內辨識到的次數
|
|
|
+ cv2.putText(frame, 'Count: {}'.format(count), (10, 50), font, 1, color, 2, cv2.LINE_AA)
|
|
|
+ # 計算FPS實際幀數並即時顯示
|
|
|
+ FPS_count += 1
|
|
|
+ fps = FPS_count / (time.time() - start_time)
|
|
|
+ cv2.putText(frame, 'FPS: {:.2f}'.format(fps), (frame.shape[1]-200, 50), font, 1, color, 2, cv2.LINE_AA)
|
|
|
+
|
|
|
+ cv2.imshow("yolov8", frame)
|
|
|
+
|
|
|
+ key = cv2.waitKey(1)
|
|
|
+ if key == ord('q'):
|
|
|
+ break
|
|
|
+
|
|
|
+
|
|
|
+if __name__ == "__main__":
|
|
|
+ main()
|