RflySimhighschool/基础智能/e13.四旋翼无人机穿环实验/CrossRing3_vehicle2.py

# import required libraries
import time
import numpy as np
import cv2
import sys

# import RflySim APIs
import PX4MavCtrlV4 as PX4MavCtrl
import VisionCaptureApi
import UE4CtrlAPI
ue = UE4CtrlAPI.UE4CtrlAPI()

vis = VisionCaptureApi.VisionCaptureApi()
# VisionCaptureApi 中的配置函数
vis.jsonLoad(0,'Config2.json') # 使用共享内存方式，并加载Config2.json中的传感器配置文件
mav = PX4MavCtrl.PX4MavCtrler(2) #对应2号飞机端口

mav.InitMavLoop()
print("Simulation Start.")

isSuss = vis.sendReqToUE4() # 向RflySim3D发送取图请求，并验证
if not isSuss: # 如果请求取图失败，则退出
    sys.exit(0)
vis.startImgCap(True) # 开启取图，并启用共享内存图像转发，转发到填写的目录

# Send command to UE4 Window 1 to change resolution 
ue.sendUE4Cmd('r.setres 720x405w',0) # 设置UE4窗口分辨率，注意本窗口仅限于显示，取图分辨率在json中配置，本窗口设置越小，资源需求越少。
ue.sendUE4Cmd('t.MaxFPS 30',0) # 设置UE4最大刷新频率，同时也是取图频率
time.sleep(2)    


width = 720
height = 405
channel = 4

# define same functions for computaion
def angle_cos(p0, p1, p2):
    d1, d2 = (p0-p1).astype('float'), (p2-p1).astype('float')
    return abs( np.dot(d1, d2) / np.sqrt( np.dot(d1, d1)*np.dot(d2, d2) ) )

def diagonal_check(p):
    d1 = np.sqrt(np.dot(p[0]-p[2], p[0]-p[2]))
    d2 = np.sqrt(np.dot(p[1]-p[3], p[1]-p[3]))
    return abs(d1-d2)*1.0/d1

def saturationYawRate(yaw_rate):
    yr_bound = 20.0
    if yaw_rate > yr_bound:
        yaw_rate = yr_bound
    if yaw_rate < -yr_bound:
        yaw_rate = -yr_bound
    return yaw_rate

def taskChange(pos_x):
    if pos_x < 40:
        task = "range1"
    elif pos_x <70:
        task = "range2"
    elif pos_x < 130:
        task = "range3"
    elif pos_x < 140:
        task = "land"
    else:
        task = "finish"
    return task

def sat(inPwm,thres=1):
    outPwm= inPwm
    if inPwm>thres:
        outPwm = thres
    elif inPwm<-thres:
        outPwm = -thres
    return outPwm


# object detect function
def objectDetect(task):
    """According task to detect objects"""
    if vis.hasData[0]:
        img_bgr=vis.Img[0]
    else:
        return -1,-1,-1
    b,g,r = cv2.split(img_bgr)    
    img_edge = cv2.Canny(b, 50, 100)
    if task == "range1" or task == "range2":
        return circleDetect(img_bgr, img_edge, b)
    else:
        return squareDetect(img_bgr, img_edge)


# square detect for object detect function
def squareDetect(img_bgr, img_edge):
    """Detect Square with PolyDP and diagonal length"""
    # find contours
    squares = []
    cnts, hierarchy = cv2.findContours(img_edge, cv2.RETR_LIST, cv2.CHAIN_APPROX_SIMPLE)
    for cnt in cnts:
        cnt_len = cv2.arcLength(cnt, True)
        cnt = cv2.approxPolyDP(cnt, 0.05 * cnt_len, True)
        if len(cnt) == 4 and cv2.contourArea(cnt) > 2000 and cv2.isContourConvex(cnt):
            cnt = cnt.reshape(-1, 2)
            diag_delta = diagonal_check(cnt)
            if diag_delta < 0.2:
                squares.append(cnt)
    cv2.drawContours( img_bgr, squares, -1, (0, 255, 0), 3)
    cv2.imshow("img_bgr", img_bgr)

    cv2.waitKey(1)
    height, width, channel = img_bgr.shape
    if squares:
        return (squares[0][0][0]+squares[0][2][0])/2 - width/2, (squares[0][0][1]+squares[0][2][1])/2 - height/2, np.sqrt(np.dot(squares[0][0]-squares[0][2], squares[0][0]-squares[0][2]))
    else:
        return -1,-1,-1
    
# circle detect for object detect function
def circleDetect(img_bgr, img_edge, img_b):
    """Hough Circle detect"""
    circles = cv2.HoughCircles(img_b, cv2.HOUGH_GRADIENT, 1, 20, param1=100, param2=30, minRadius=0, maxRadius=0)
    if circles is not None:
        circles = np.uint16(np.around(circles))
        obj = circles[0, 0]
        cv2.circle(img_bgr, (obj[0], obj[1]), obj[2], (0,255,0), 2)
        cv2.circle(img_bgr, (obj[0], obj[1]), 2, (0,255,255), 3)
        cv2.imshow("img_bgr", img_bgr)
      
        cv2.waitKey(1)
        height, width, channel = img_bgr.shape
        return obj[0]-width/2, obj[1]-height/2, obj[2]
    else:
        return -1,-1,-1


# approaching Objective/ crossing rings algorithm
def approachObjective():
    # 0. parameters
    # some parameters that work:(0.03, -0.03, 1, 5.0); (0.06, -0.04, 1, 10.0)
    K_z = 0.004 * 640 / height
    K_yawrate = 0.005 * 480 / width
    task = "range1"
    # 1. start
    startAppTime= time.time()
    lastTime = time.time()
    # time interval of the timer
    timeInterval = 1/30.0 #here is 0.0333s (30Hz)
    while (task != "finish") & (task != "land"):
        lastTime = lastTime + timeInterval
        sleepTime = lastTime - time.time()
        if sleepTime > 0:
            time.sleep(sleepTime) # sleep until the desired clock
        else:
            lastTime = time.time()
        # The following code will be executed 30Hz (0.0333s)
        
        # 1.1. detect object and get error with center of image
        ex, ey, r = objectDetect(task)
        # 1.2. where is the drone
        pos_x = mav.uavPosNED[0]
        #print(time.time())
        # 1.3. update task
        task = taskChange(pos_x)
        # 1.4. attack
        if ex != -1:
            vx = sat(time.time()-startAppTime,3)
            vy = 0
            vz = K_z * ey
            yawrate = K_yawrate * ex
            mav.SendVelFRD(vx, vy, vz, yawrate)
            
    lastTime = time.time()  
    while task == "land":
        lastTime = lastTime + timeInterval
        sleepTime = lastTime - time.time()
        if sleepTime > 0:
            time.sleep(sleepTime) # sleep until the desired clock
        else:
            lastTime = time.time()
        
        pos_x = mav.uavPosNED[0]
        task = taskChange(pos_x)
        mav.SendVelFRD(0, 0, 1, 0)
        
        


print("Enter Offboard mode.")
time.sleep(5)    
mav.initOffboard()
time.sleep(0.5)
mav.SendMavArm(True) # Arm the drone
mav.SendPosNED(5, 0, -5, 0) # Fly to target position 5,0，-5    

# Show CV image and set the position
if vis.hasData[0]:           
    img_bgr=vis.Img[0]
    cv2.imshow("img_bgr", img_bgr)
    cv2.waitKey(1)
    #time.sleep(0.5)

time.sleep(5)

# start crossing ring task    
approachObjective()