从前景提取grabcut中检测物体的运动

问题描述

我是 opencv python 的新手

现在，我正在对从前景提取（使用抓取）中获得的对象进行运动检测（在网络摄像头中）。我已经从抓取中获得了一个对象，但我不知道如何编写代码来检测该对象的移动并在网络摄像头屏幕中显示该移动。

非常感谢您提前

解决方法

这里是使用运动跟踪进行对象检测的代码，但您需要下载一些外部文件，如 mobilenet 模型、CentroidTracker

import datetime
import imutils
import numpy as np
import csv 
# from centroidtracker import CentroidTracker
from pyimagesearch.centroidtracker import CentroidTracker
protopath = "mobilenet_ss/MobileNetSSD_deploy.prototxt"
modelpath = "mobilenet_ss/MobileNetSSD_deploy.caffemodel"


detector = cv2.dnn.readNetFromCaffe(prototxt=protopath,caffeModel=modelpath)
# detector.setPreferableBackend(cv2.dnn.DNN_BACKEND_INFERENCE_ENGINE)
detector.setPreferableTarget(cv2.dnn.DNN_TARGET_CPU)

fields=['new_id_detected',"total_person_count"]
filename = "person_records.csv"
  
outputlist=[]        

CLASSES = ["background","aeroplane","bicycle","bird","boat","bottle","bus","car","cat","chair","cow","diningtable","dog","horse","motorbike","person","pottedplant","sheep","sofa","train","tvmonitor"]

tracker = CentroidTracker(maxDisappeared=80,maxDistance=90)


def non_max_suppression_fast(boxes,overlapThresh):
    try:
        if len(boxes) == 0:
            return []

        if boxes.dtype.kind == "i":
            boxes = boxes.astype("float")

        pick = []

        x1 = boxes[:,0]
        y1 = boxes[:,1]
        x2 = boxes[:,2]
        y2 = boxes[:,3]

        area = (x2 - x1 + 1) * (y2 - y1 + 1)
        idxs = np.argsort(y2)

        while len(idxs) > 0:
            last = len(idxs) - 1
            i = idxs[last]
            pick.append(i)

            xx1 = np.maximum(x1[i],x1[idxs[:last]])
            yy1 = np.maximum(y1[i],y1[idxs[:last]])
            xx2 = np.minimum(x2[i],x2[idxs[:last]])
            yy2 = np.minimum(y2[i],y2[idxs[:last]])

            w = np.maximum(0,xx2 - xx1 + 1)
            h = np.maximum(0,yy2 - yy1 + 1)

            overlap = (w * h) / area[idxs[:last]]

            idxs = np.delete(idxs,np.concatenate(([last],np.where(overlap > overlapThresh)[0])))

        return boxes[pick].astype("int")
    except Exception as e:
        print("Exception occurred in non_max_suppression : {}".format(e))


def main():
    cap = cv2.VideoCapture('project_video.mp4')

    fourcc = cv2.VideoWriter_fourcc('m','p','4','v')
    out = cv2.VideoWriter("output/output.mp4",fourcc,5.0,(600,337))


    fps_start_time = datetime.datetime.now()
    fps = 0
    total_frames = 0
    lpc_count = 0
    opc_count = 0
    object_id_list = []

    
    # dtime = dict()
    # dwell_time = dict()

    while True:
        ret,frame = cap.read()
        if not ret:
            break
        frame = imutils.resize(frame,width=600)
        total_frames = total_frames + 1

        (H,W) = frame.shape[:2]
        #print("h,w",H,W)
        blob = cv2.dnn.blobFromImage(frame,0.007843,(W,H),127.5)

        detector.setInput(blob)
        person_detections = detector.forward()
        rects = []
        for i in np.arange(0,person_detections.shape[2]):
            confidence = person_detections[0,i,2]
            if confidence > 0.5:
                idx = int(person_detections[0,1])

               

                person_box = person_detections[0,3:7] * np.array([W,W,H])
                (startX,startY,endX,endY) = person_box.astype("int")
                rects.append(person_box)

        boundingboxes = np.array(rects)
        boundingboxes = boundingboxes.astype(int)
        rects = non_max_suppression_fast(boundingboxes,0.3)

        objects = tracker.update(rects)
        for (objectId,bbox) in objects.items():
            x1,y1,x2,y2 = bbox
            x1 = int(x1)
            y1 = int(y1)
            x2 = int(x2)
            y2 = int(y2)

            cv2.rectangle(frame,(x1,y1),(x2,y2),(0,255),2)
            text = "ID: {}".format(objectId)
            cv2.putText(frame,text,y1-5),cv2.FONT_HERSHEY_COMPLEX_SMALL,1,1)

            if objectId not in object_id_list:
                object_id_list.append(objectId)
                # dtime[objectId] = datetime.datetime.now()
                # dwell_time[objectId] = 0
            # else:
            #     curr_time = datetime.datetime.now()
            #     old_time = dtime[objectId]
            #     time_diff = curr_time - old_time
            #     dtime[objectId] = datetime.datetime.now()
            #     sec = time_diff.total_seconds()
            #     dwell_time[objectId] += sec

            # text = "{}|{}".format(objectId,int(dwell_time[objectId]))
            # cv2.putText(frame,1)
        fps_end_time = datetime.datetime.now()
        time_diff = fps_end_time - fps_start_time
        if time_diff.seconds == 0:
            fps = 0.0
        else:
            fps = (total_frames / time_diff.seconds)

        fps_text = "FPS: {:.2f}".format(fps)
        
        cv2.putText(frame,fps_text,(5,30),1)

        lpc_count = len(objects)
        opc_count = len(object_id_list)

        lpc_txt = "LPC: {}".format(lpc_count)
        opc_txt = "OPC: {}".format(opc_count)
        # writing to csv file  
        outputlist.append([lpc_count,opc_count])
        
            
        cv2.putText(frame,lpc_txt,60),1)
        cv2.putText(frame,opc_txt,90),1)

        out.write(frame)
        cv2.imshow("Application",frame)
        key = cv2.waitKey(1)
        if key == ord('q'):
            break
        
    cv2.destroyAllWindows()
    with open(filename,'w') as csvfile:  
            # creating a csv writer object  
            csvwriter = csv.writer(csvfile)   
            csvwriter.writerow(fields)
            csvwriter.writerows(outputlist)

main()```