电赛E题视觉基础部分：
OpenMV固定于舵机云台前，屏幕画面置于镜头前，OpenMV将屏幕中数据信息，包括原点位置，激光点位置，A4靶纸边框位置利用串口传给主控芯片。（完整代码直接见第四部分）
一、OpenMV串口部分代码
初始化OpenMV串口3，利用数据包一次发送十个数据给主控，包括五个坐标点——激光点坐标，靶值四个角点坐标。当需要发送数据时，只要在程序中引用此函数即可。
import sensor, image, time
import ustruct
from pyb import UART,LED
uart = UART(3,9600)
uart.init(9600, bits=8, parity=None, stop=1) # 开启OpenMV串口3 波特率为9600，8位，一位停止位
def sending_data(rx,ry,cx,cy,cw,ch,xx,xy,xw,xh): # 定义发送数据包函数 包头位“FF”,包尾为“FE”
global uart; # 数据包中有十个数据
data = ustruct.pack("<bhhhhhhhhb", # 分别为激光点中心坐标以及靶纸corner坐标
0xFF,
int(rx),
int(ry),
int(cx),
int(cy),
int(cw),
int(ch),
int(xx),
int(xy),
int(xw),
int(xh),
0xFE)
uart.write(data);
二、识别激光点
利用OpenMV 自带的寻找色块的函数，寻找画面中最大激光点阈值的色块并读取中心坐标。3
2.1 阈值设置
首先进行阈值的选择、ROI的选择以及曝光时间选择。（所给的数据均为比赛时我用的数据，请自行调整）
#****************************阈值调整*******************************
area=(45, 2, 200, 198)
red=(88, 100, -16, 22, -128, 127)
sensor.reset()
sensor.set_pixformat(sensor.RGB565)
sensor.set_framesize(sensor.QVGA)
sensor.set_auto_whitebal(True)
sensor.set_auto_gain(False)
sensor.skip_frames(time = 2000)
sensor.set_vflip(True)
sensor.set_hmirror(True)
#****************************曝光时间*******************************
sensor.set_auto_exposure(False,4500)
clock = time.clock()2.2 色块识别
while(True)
clock.tick()
img = sensor.snapshot().lens_corr(strength = 1.6, zoom = 1.0)
red_point= img.find_blobs([red])
if red_point:
r= red_point[0]
img.draw_rectangle(r[0:4],color=(0,0,0)) # rect
rx=r[5]
ry=r[6]三、识别A4靶纸上电工胶带角点
3.1 方法一 ：直线交点法
用OpenMV自带的寻找直线的函数在ROI中找到所有的直线，然后通过判断所寻找的直线数是否为四进行下一步求交点。
#*****************************line_threshold**********************
line_threshold=2500
rect_threshold=10000
while(True):
clock.tick()
img = sensor.snapshot().lens_corr(strength = 1.6, zoom = 1.0)
red_point= img.find_blobs([red])
if red_point:
r= red_point[0]
img.draw_rectangle(r[0:4],color=(0,0,0)) # rect
rx=r[5]
ry=r[6]
#print(rx,ry)
#******************************************************************lines*****************
lines = img.find_lines(roi=area, threshold=line_threshold)
i=0
if len(lines)==4 :如果识别到四条直线，依次两两直线遍历，是否夹角大于一定值，判断直线是否相交，从而求出交点坐标 。
if len(lines)==4 :
theta0=lines[0].theta()
theta1=lines[1].theta()
theta2=lines[2].theta()
theta3=lines[3].theta()
for line in lines:
img.draw_line(line.line(),color=[255,0,0])
if abs(theta0-theta1)>45:
a[i]=calculate_intersection(lines[0],lines[1])[0]
i=i+1
a[i]=calculate_intersection(lines[0],lines[1])[1]
i=i+1
if abs(theta0-theta2)>45:
a[i]=calculate_intersection(lines[0],lines[2])[0]
i=i+1
a[i]=calculate_intersection(lines[0],lines[2])[1]
i=i+1
if abs(theta0-theta3)>45:
a[i]=calculate_intersection(lines[0],lines[3])[0]
i=i+1
a[i]=calculate_intersection(lines[0],lines[3])[1]
i=i+1
if abs(theta1-theta2)>45:
a[i]=calculate_intersection(lines[2],lines[1])[0]
i=i+1
a[i]=calculate_intersection(lines[2],lines[1])[1]
i=i+1
if abs(theta1-theta3)>45:
a[i]=calculate_intersection(lines[1],lines[3])[0]
i=i+1
a[i]=calculate_intersection(lines[1],lines[3])[1]
i=i+1
if abs(theta3-theta2)>45:
a[i]=calculate_intersection(lines[2],lines[3])[0]
i=i+1
a[i]=calculate_intersection(lines[2],lines[3])[1]
i=i+1其中求交点的坐标函数如下：
def calculate_intersection(line1, line2):
'''
计算两条线的交点
'''
a1 = line1.y2() - line1.y1()
b1 = line1.x1() - line1.x2()
c1 = line1.x2()*line1.y1() - line1.x1()*line1.y2()
a2 = line2.y2() - line2.y1()
b2 = line2.x1() - line2.x2()
c2 = line2.x2() * line2.y1() - line2.x1()*line2.y2()
if (a1 * b2 - a2 * b1) != 0 and (a2 * b1 - a1 * b2) != 0:
cross_x = int((b1*c2-b2*c1)/(a1*b2-a2*b1))
cross_y = int((c1*a2-c2*a1)/(a1*b2-a2*b1))
return (cross_x, cross_y)
return None3.2 方法二 : 矩形角点法
利用OpenMV自带的寻找矩形函数。但需要注意阈值以及ROI，可以设置识别矩形的长宽范围筛选出我们所需要的那个矩形，从而记录角点坐标。
while(True):
clock.tick()
img = sensor.snapshot()
# -----矩形框部分-----
# 在图像中寻找矩形
for r in img.find_rects(threshold = 10000):
# 判断矩形边长是否符合要求
if r.w() > 20 and r.h() > 20:
# 在屏幕上框出矩形
img.draw_rectangle(r.rect(), color = (255, 0, 0), scale = 4)
# 获取矩形角点位置
corner = r.corners()
# 在屏幕上圈出矩形角点
img.draw_circle(corner[0][0], corner[0][1], 5, color = (0, 0, 255), thickness = 2, fill = False)
img.draw_circle(corner[1][0], corner[1][1], 5, color = (0, 0, 255), thickness = 2, fill = False)
img.draw_circle(corner[2][0], corner[2][1], 5, color = (0, 0, 255), thickness = 2, fill = False)
img.draw_circle(corner[3][0], corner[3][1], 5, color = (0, 0, 255), thickness = 2, fill = False)
# 打印四个角点坐标, 角点1的数组是corner[0], 坐标就是(corner[0][0],corner[0][1])
# 角点检测输出的角点排序每次不一定一致，矩形左上的角点有可能是corner0,1,2,3其中一个
corner1_str = f"corner1 = ({corner[0][0]},{corner[0][1]})"
corner2_str = f"corner2 = ({corner[1][0]},{corner[1][1]})"
corner3_str = f"corner3 = ({corner[2][0]},{corner[2][1]})"
corner4_str = f"corner4 = ({corner[3][0]},{corner[3][1]})"
print(corner1_str + "\n" + corner2_str + "\n" + corner3_str + "\n" + corner4_str)四、完整程序
这里只给方法一的完整代码，方法二自行替换即可。
import sensor, image, time
import ustruct
from pyb import UART,LED
#****************************阈值调整*******************************
area=(45, 2, 200, 198)
red=(88, 100, -16, 22, -128, 127) #only red
sensor.reset()
sensor.set_pixformat(sensor.RGB565)
sensor.set_framesize(sensor.QVGA)
sensor.set_auto_whitebal(True)
sensor.set_auto_gain(False)
sensor.skip_frames(time = 2000)
sensor.set_vflip(True)
sensor.set_hmirror(True)
#****************************曝光时间*******************************
sensor.set_auto_exposure(False,4500)
clock = time.clock()
#*****************************line_threshold**********************
line_threshold=2500
rect_threshold=10000
uart = UART(3,9600)
uart.init(9600, bits=8, parity=None, stop=1)
def calculate_intersection(line1, line2):
'''
计算两条线的交点
'''
a1 = line1.y2() - line1.y1()
b1 = line1.x1() - line1.x2()
c1 = line1.x2()*line1.y1() - line1.x1()*line1.y2()
a2 = line2.y2() - line2.y1()
b2 = line2.x1() - line2.x2()
c2 = line2.x2() * line2.y1() - line2.x1()*line2.y2()
if (a1 * b2 - a2 * b1) != 0 and (a2 * b1 - a1 * b2) != 0:
cross_x = int((b1*c2-b2*c1)/(a1*b2-a2*b1))
cross_y = int((c1*a2-c2*a1)/(a1*b2-a2*b1))
return (cross_x, cross_y)
return None
def sending_data(rx,ry,cx,c):
global uart;
data = ustruct.pack("<bhhhhhhhhb",
0xFF,
int(rx),
int(ry),
int(cx),
int(cy),
int(cw),
int(ch),
int(xx),
int(xy),
int(xw),
int(xh),
0xFE)
uart.write(data);
cx=0
cy=0
rx=0
ry=0
a=[0,0,0,0,0,0,0,0,0,0]
while(True):
clock.tick()
img = sensor.snapshot().lens_corr(strength = 1.6, zoom = 1.0)
red_point= img.find_blobs([red])
if red_point:
r= red_point[0]
img.draw_rectangle(r[0:4],color=(0,0,0)) # rect
rx=r[5]
ry=r[6]
#print(rx,ry)
#******************************************************************lines*****************
lines = img.find_lines(roi=area, threshold=line_threshold)
i=0
if len(lines)==4 :
theta0=lines[0].theta()
theta1=lines[1].theta()
theta2=lines[2].theta()
theta3=lines[3].theta()
for line in lines:
img.draw_line(line.line(),color=[255,0,0])
if abs(theta0-theta1)>45:
a[i]=calculate_intersection(lines[0],lines[1])[0]
i=i+1
a[i]=calculate_intersection(lines[0],lines[1])[1]
i=i+1
if abs(theta0-theta2)>45:
a[i]=calculate_intersection(lines[0],lines[2])[0]
i=i+1
a[i]=calculate_intersection(lines[0],lines[2])[1]
i=i+1
if abs(theta0-theta3)>45:
a[i]=calculate_intersection(lines[0],lines[3])[0]
i=i+1
a[i]=calculate_intersection(lines[0],lines[3])[1]
i=i+1
if abs(theta1-theta2)>45:
a[i]=calculate_intersection(lines[2],lines[1])[0]
i=i+1
a[i]=calculate_intersection(lines[2],lines[1])[1]
i=i+1
if abs(theta1-theta3)>45:
a[i]=calculate_intersection(lines[1],lines[3])[0]
i=i+1
a[i]=calculate_intersection(lines[1],lines[3])[1]
i=i+1
if abs(theta3-theta2)>45:
a[i]=calculate_intersection(lines[2],lines[3])[0]
i=i+1
a[i]=calculate_intersection(lines[2],lines[3])[1]
i=i+1
print(a)
FH = bytearray([0xFF,rx,ry,a[0],a[1],a[2],a[3],a[4],a[5],a[6],a[7],0xFE])
uart.write(FH)
print(rx,ry)
print(clock.fps())
823
被折叠的 条评论
为什么被折叠?
到【灌水乐园】发言
