Add yolo_webcam_count.py

yolo_webcam_9_3.py

@@ -0,0 +1,255 @@
+
+from pyzbar import pyzbar
+from datetime import datetime
+import argparse
+import cv2
+import sys
+import time
+import os
+import sys
+import subprocess
+import signal
+import numpy as np
+from datetime import datetime
+
+
+#test0
+
+
+start2 = 0
+
+net = cv2.dnn.readNet("yolov4.weights", "cfg/yolov4.cfg")
+classes = []
+with open("cfg/coco.names", "r") as f:
+    classes = [line.strip() for line in f.readlines()]
+layer_names = net.getLayerNames()
+output_layers = [layer_names[i[0] - 1] for i in net.getUnconnectedOutLayers()]
+colors = np.random.uniform(0, 255, size=(len(classes), 3))
+
+
+#############
+count = 0
+attached = 0
+before = 0
+left_bound = 0
+right_bound = 0
+fishbowlsize = 0.25
+bowl_left = 0
+bowl_right = 0
+isDragging = False
+x0, y0, w, h = -1, -1, -1, -1
+blue, red, black = (255, 0, 0), (0, 0, 255), (0, 0, 0)
+meter_rate = 1
+left_frame = 0
+right_frame = 0
+##############333
+
+# vc0 = cv2.VideoCapture(0)
+vc = cv2.VideoCapture(2) # 0은 노트북 웹캠 2는 usb로 연결된 웹캠
+#vc = cv2.VideoCapture('/home/kjjgo35/PythonHome/darknet/test.mkv')
+
+
+global data
+data =  open('fish_logging.txt', 'a')
+
+def onMouse(event, x, y, flags, param):
+    global isDragging, x0, y0, img, left_bound, right_bound, height, name, meter_rate, bowl_left, bowl_right
+    if event == cv2.EVENT_LBUTTONDOWN:
+        isDragging = True
+        x0 = x
+        y0 = y
+    elif event == cv2.EVENT_MOUSEMOVE:
+        if isDragging:
+            img_draw = img.copy()
+            cv2.rectangle(img_draw, (x0, y0), (x, y), blue, 2)
+            cv2.imshow(name, img_draw)
+    elif event == cv2.EVENT_LBUTTONUP:
+        if isDragging:
+            isDragging = False
+            w = x - x0
+            h = y - y0
+            if w > 0 and h > 0:
+                img_draw = img.copy()
+                img_draw2 = img.copy()
+                cv2.rectangle(img_draw, (x0, y0), (x, y), red, 2)
+                cv2.imshow(name, img_draw)
+                roi = img_draw[y0:y0+h, x0:x0+w]
+                #cv2.imshow('cropped', roi)
+                cv2.imshow(name, img_draw)
+                #cv2.moveWindow('cropped', 0, 0)
+                #cv2.imwrite('./cropped.png', roi)
+                if name == 'mirror':
+                    left_bound = x0
+                    right_bound = x
+                    print(left_bound, right_bound)
+                    cv2.rectangle(img_draw2, (left_bound,1), (right_bound, int(height)),(0,0,255), 3)
+                elif name == 'bowl':
+                    bowl_left = x0
+                    bowl_right = x
+                    print(bowl_left, bowl_right)
+                    cv2.rectangle(img_draw2, (bowl_left,1), (bowl_right, int(height)),(0,0,255), 3)
+                    meter_rate = fishbowlsize /(bowl_right - bowl_left)
+                font = cv2.FONT_HERSHEY_PLAIN
+                cv2.putText(img_draw2, name, (x0, 40), font, 3, black, 3)
+                cv2.imshow(name, img_draw2)
+            else:
+                cv2.imshow(name, img)
+                print('drag should start from left-top side')
+
+
+def checkbound(label,xs, xe):
+
+     if(label != "zebra fish"):
+      print("other object")
+      return False
+
+     elif xs >= left_bound and xs <= right_bound:
+      print("Detect " + label)
+      data.write(str(datetime.utcnow().strftime('%Y-%m-%d %H:%M:%S.%f')) + " " + label + " " + str(count) + "\n" )
+      return True
+
+     else: return False
+
+def checkframe():
+    global left_frame,right_frame,bowl_left,bowl_right,left_bound,right_bound
+    bowl_size = bowl_right - bowl_left
+    left_frame = left_bound
+    right_frame = right_bound
+    if bowl_left < left_bound :
+        left_frame = left_bound - (bowl_size / 5)
+    if bowl_right > right_bound :
+        right_frame = right_bound + (bowl_size / 5)
+
+
+while True :
+    #time.sleep(0.1)
+    print("waiting...")
+
+
+    if cv2.waitKey(1) & 0xFF == ord('q'):
+        break
+
+    if True:
+        print("webcam start!")
+        start = time.time()
+        chk = 0         
+        vc.set(10,0.5)
+
+        ret,frame = vc.read()
+        cv2.imshow("Video Window", frame)
+        cv2.imwrite('selectroi.jpg',frame)
+        img = cv2.imread('selectroi.jpg')
+        name = 'mirror'
+        cv2.imshow(name,img)
+        print("Set the mirror and press any key")
+        height, width, channels = img.shape
+        cv2.setMouseCallback(name,onMouse)
+        print(left_bound,',', right_bound)
+        cv2.waitKey()
+        cv2.destroyWindow(name)
+
+        name = 'bowl'
+        cv2.imshow(name,img)
+        print("Set the bowl and press any key")
+        cv2.setMouseCallback(name,onMouse)
+        print(left_bound,',', right_bound)
+        cv2.waitKey()
+        cv2.destroyWindow(name)
+
+        checkframe()
+        print("lf = " , left_frame, right_frame)
+
+        while time.time() - start <= 300 :
+
+            # ret0, frame0 = vc0.read()d
+            # cv2.imshow("Video Window0", frame0)  
+
+            ret, frame = vc.read()
+            cv2.imshow("Video Window", frame)
+            cv2.imwrite('image.jpg',frame)
+            image = cv2.imread('image.jpg')
+            cv2.imshow("Image", image)
+
+            # image = cv2.resize(image, None, fx=0.4, fy=0.4)
+            height, width, channels = image.shape
+            #left_bound = 1
+            #right_bound = int(width/6)
+
+
+            barcodes = pyzbar.decode(image)
+
+            blob = cv2.dnn.blobFromImage(image, 0.00392, (416, 416), (0, 0, 0), True, crop=False)
+            net.setInput(blob)
+            outs = net.forward(output_layers)
+
+            class_ids = []
+            confidences = []
+            boxes = []
+            for out in outs:
+             for detection in out:
+              scores = detection[5:]
+              class_id = np.argmax(scores)
+              confidence = scores[class_id]
+              if confidence > 0.5:
+               # Object detected
+               center_x = int(detection[0] * width)
+               center_y = int(detection[1] * height)
+               #if center_x < left_frame or center_x > right_frame :
+                #   continue
+               w = int(detection[2] * width)
+               h = int(detection[3] * height)
+               # 좌표
+               x = int(center_x - w / 2)
+               y = int(center_y - h / 2)
+               boxes.append([x, y, w, h])
+               confidences.append(float(confidence))
+               class_ids.append(class_id)
+
+            indexes = cv2.dnn.NMSBoxes(boxes, confidences, 0.5, 0.4)  
+            font = cv2.FONT_HERSHEY_PLAIN
+            attached = 0
+            for i in range(len(boxes)):
+             if i in indexes:
+              x, y, w, h = boxes[i]
+              x += w//20
+              y += h//20
+              w -= w//20
+              h -= h//20
+              if x + w <= left_bound :
+                distance = str(round((left_bound - x - w) * meter_rate,3))
+              elif x >= right_bound : 
+                distance = str(round((x - right_bound) * meter_rate, 3))
+              else :
+                distance = "0"
+              label = str(classes[class_ids[i]])
+              color = colors[i]
+              cv2.rectangle(image, (x, y), (x + w, y + h), color, 2)
+              cv2.putText(image, label, (x, y + 30), font, 3, color, 3)
+              cv2.putText(image, distance , (x, y - 30), font, 3, color, 3)
+              cv2.imshow("Image", image)
+              if checkbound(label,x,x+w):
+                attached += 1
+
+            if(attached > before):
+              count += attached - before
+              print(count)
+            before = attached
+
+            # change option   
+            cv2.rectangle(image, (left_bound,1), (right_bound, int(height)),(0,0,255), 3)
+            cv2.imshow("Image", image)
+
+
+
+            if cv2.waitKey(1) & 0xFF == ord('q'):
+                break
+
+        vc.release()
+        data.close()
+        cv2.destroyAllWindows()
+
+        if chk == 1:
+            print("access success")
+
+        else:
+            print("access failed")