树莓派4B + ESC (Hobbywing) + SKYwalker 2212 无刷电机

测试步骤

0. 搭建电路
   
1. 校准电调
2. 驱动电机ESC.py 测试代码:

# This program will let you test your ESC and brushless motor.
# Make sure your battery is not connected if you are going to calibrate it at first.
# Since you are testing your motor, I hope you don't have your propeller attached to it otherwise you are in trouble my friend...?
# This program is made by AGT @instructable.com. DO NOT REPUBLISH THIS PROGRAM... actually the program itself is harmful                                             pssst Its not, its safe.

import os     #importing os library so as to communicate with the system
import time   #importing time library to make Rpi wait because its too impatient 
os.system ("sudo pigpiod") #Launching GPIO library
time.sleep(1) # As i said it is too impatient and so if this delay is removed you will get an error
import pigpio #importing GPIO library

ESC=4  #Connect the ESC in this GPIO pin 

pi = pigpio.pi();
pi.set_servo_pulsewidth(ESC, 0) 

max_value = 2000 #change this if your ESC's max value is different or leave it be
min_value = 700  #change this if your ESC's min value is different or leave it be
print "For first time launch, select calibrate"
print "Type the exact word for the function you want"
print "calibrate OR manual OR control OR arm OR stop"

def manual_drive(): #You will use this function to program your ESC if required
    print "You have selected manual option so give a value between 0 and you max value"    
    while True:
        inp = raw_input()
        if inp == "stop":
            stop()
            break
        elif inp == "control":
            control()
            break
        elif inp == "arm":
            arm()
            break   
        else:
            pi.set_servo_pulsewidth(ESC,inp)

def calibrate():   #This is the auto calibration procedure of a normal ESC
    pi.set_servo_pulsewidth(ESC, 0)
    print("Disconnect the battery and press Enter")
    inp = raw_input()
    if inp == '':
        pi.set_servo_pulsewidth(ESC, max_value)
        print("Connect the battery NOW.. you will here two beeps, then wait for a gradual falling tone then press Enter")
        inp = raw_input()
        if inp == '':            
            pi.set_servo_pulsewidth(ESC, min_value)
            print "Wierd eh! Special tone"
            time.sleep(7)
            print "Wait for it ...."
            time.sleep (5)
            print "Im working on it, DONT WORRY JUST WAIT....."
            pi.set_servo_pulsewidth(ESC, 0)
            time.sleep(2)
            print "Arming ESC now..."
            pi.set_servo_pulsewidth(ESC, min_value)
            time.sleep(1)
            print "See.... uhhhhh"
            control() # You can change this to any other function you want

def control(): 
    print "I'm Starting the motor, I hope its calibrated and armed, if not restart by giving 'x'"
    time.sleep(1)
    speed = 1500    # change your speed if you want to.... it should be between 700 - 2000
    print "Controls - a to decrease speed & d to increase speed OR q to decrease a lot of speed & e to increase a lot of speed"
    while True:
        pi.set_servo_pulsewidth(ESC, speed)
        inp = raw_input()

        if inp == "q":
            speed -= 100    # decrementing the speed like hell
            print "speed = %d" % speed
        elif inp == "e":    
            speed += 100    # incrementing the speed like hell
            print "speed = %d" % speed
        elif inp == "d":
            speed += 10     # incrementing the speed 
            print "speed = %d" % speed
        elif inp == "a":
            speed -= 10     # decrementing the speed
            print "speed = %d" % speed
        elif inp == "stop":
            stop()          #going for the stop function
            break
        elif inp == "manual":
            manual_drive()
            break
        elif inp == "arm":
            arm()
            break   
        else:
            print "WHAT DID I SAID!! Press a,q,d or e"

def arm(): #This is the arming procedure of an ESC 
    print "Connect the battery and press Enter"
    inp = raw_input()    
    if inp == '':
        pi.set_servo_pulsewidth(ESC, 0)
        time.sleep(1)
        pi.set_servo_pulsewidth(ESC, max_value)
        time.sleep(1)
        pi.set_servo_pulsewidth(ESC, min_value)
        time.sleep(1)
        control() 

def stop(): #This will stop every action your Pi is performing for ESC ofcourse.
    pi.set_servo_pulsewidth(ESC, 0)
    pi.stop()

#This is the start of the program actually, to start the function it needs to be initialized before calling... stupid python.    
inp = raw_input()
if inp == "manual":
    manual_drive()
elif inp == "calibrate":
    calibrate()
elif inp == "arm":
    arm()
elif inp == "control":
    control()
elif inp == "stop":
    stop()
else :
    print "Thank You for not following the things I'm saying... now you gotta restart the program STUPID!!"

起因

今天 rockets 问我, 是否能够将二维码里面的色块用 0 和 1 表示出来.

分析

我脑海里分析了一下,不知道是不是可以用 opencv 实现,就打开电脑尝试了一下,似乎是可以的.
思路, 首先通过 qrcode 生成一个图片,然后保存一下,通过opencv 导入图片,然后再判断 10x10 大小的块中是否全是白色,如果是白色,就写一个 1
,如果是黑色,就写个 0. 然后遍历整个图片的 shape.

Talk is cheap, show me the code

import qrcode
import cv2
import numpy as np

url = 'https://www.dfrobot.com/product-2480.html'
file_name = 'qrcode.png'

def generate_image(url):
    qr = qrcode.QRCode()
    qr.add_data(url)
    qr.make(fit=False)
    img = qr.make_image(fill_color="black", back_color="white")
    img.save(file_name)

generate_image(url)
img1 = cv2.imread(file_name)
scale_percent = 80
width = int(img1.shape[1] * scale_percent / 100)
height = int(img1.shape[0] * scale_percent / 100)
dim = (width, height)

resized_img = cv2.resize(img1, dim, interpolation=cv2.INTER_AREA)
# gray_img1 = cv2.cvtColor(img1, cv2.COLOR_BGR2GRAY)
for i in range(0, int(resized_img.shape[0]), 10):
    for j in range(0, int(resized_img.shape[1]), 10):
        if np.mean(resized_img[i:i+10, j:j+10]) == 255:
            cv2.putText(resized_img, '0', (i, j), cv2.FONT_HERSHEY_SIMPLEX, .2, (255, 0, 0))
        else:
            cv2.putText(resized_img, '1', (i, j), cv2.FONT_HERSHEY_SIMPLEX, .2, (255, 0, 0))

cv2.imshow('resized_img', resized_img)
cv2.waitKey(0)

执行效果:

重点

np.mean(图片数组) == 255的判断部分,可以判断图片是不是白色.

重构了一下,优化后:

import qrcode
import cv2
import numpy as np

# 定义生成 qrcode 的 url 链接和生成的文件名
url = 'https://www.dfrobot.com/product-2480.html'
file_name = 'qrcode.png'

def generate_image(url):
    """"
    生成二维码并保存
    """
    qr = qrcode.QRCode()
    qr.add_data(url)
    qr.make(fit=False)
    img = qr.make_image(fill_color="black", back_color="white")
    img.save(file_name)

def process_image(filename):
    """
    :param filename: 放入生成的二维码
    :return: 显示图
    """
    img1 = cv2.imread(file_name)
    scale_percent = 80
    width = int(img1.shape[1] * scale_percent / 100)
    height = int(img1.shape[0] * scale_percent / 100)
    dim = (width, height)
    resized_img = cv2.resize(img1, dim, interpolation=cv2.INTER_AREA)
    for i in range(0, int(resized_img.shape[0]), 10):
        for j in range(0, int(resized_img.shape[1]), 10):
            if np.mean(resized_img[i:i+5, j:j+5]) == 255:
                cv2.putText(resized_img, '0', (i+5, j+5), cv2.FONT_HERSHEY_SIMPLEX, .2, (255, 0, 0))
            else:
                cv2.putText(resized_img, '1', (i+5, j+5), cv2.FONT_HERSHEY_SIMPLEX, .2, (255, 0, 0))

    cv2.imshow('resized_img', resized_img)
    cv2.waitKey(0)

if __name__ == "__main__":
    generate_image(url)
    process_image(file_name)

优化后结果:

感觉还是不太对.

继续优化试试看.

import qrcode
import cv2
import numpy as np

# 定义生成 qrcode 的 url 链接和生成的文件名
url = 'https://www.dfrobot.com/product-2480.html'
file_name = 'qrcode.png'

def generate_image(url):
    """"
    生成二维码并保存
    """
    qr = qrcode.QRCode()
    qr.add_data(url)
    qr.make(fit=False)
    img = qr.make_image(fill_color="black", back_color="white")
    img.save(file_name)

def process_image(filename):
    """
    :param filename: 放入生成的二维码
    :return: 显示图
    """
    img1 = cv2.imread(file_name)
    scale_percent = 200
    width = int(img1.shape[1] * scale_percent / 100)
    height = int(img1.shape[0] * scale_percent / 100)
    dim = (width, height)
    resized_img = cv2.resize(img1, dim, interpolation=cv2.INTER_AREA)
    resized_img = cv2.cvtColor(resized_img, cv2.COLOR_BGR2GRAY)
    for i in range(0, int(resized_img.shape[0]), 10):
        for j in range(0, int(resized_img.shape[1]), 10):
            if np.mean(resized_img[i:i+3, j:j+3]) == 255:
                cv2.putText(resized_img, '1', (i + 5, j + 5), cv2.FONT_HERSHEY_SIMPLEX, .2, (0, 0, 255))
            else:
                cv2.putText(resized_img, '0', (i + 5, j + 5), cv2.FONT_HERSHEY_SIMPLEX, .2, (255, 0, 0))

    cv2.imshow('img', resized_img)
    cv2.waitKey(0)

if __name__ == "__main__":
    generate_image(url)
    process_image(file_name)

输出结果:

今天实现了对物体检测并控制舵机运动的应用。
舵机使用了pca9685控制，因此备份一下pac9685舵机控制板的驱动：

PCA9685.py

import time
import math

class PWM:
    _mode_adr      = 0x00
    _base_adr_low  = 0x08
    _base_adr_high = 0x09
    _prescale_adr  = 0xFE

    def __init__(self, bus, address = 0x40):
        '''
        Creates an instance of the PWM chip at given i2c address.
        @param bus: the SMBus instance to access the i2c port(0 or 1).
        @param address: the address of the i2c chip (default: 0x40).
        '''
        self.bus = bus
        self.address = address
        self._writeByte(self._mode_adr, 0x00)

    def setFreq(self, freq):
        '''
        Sets the PWM frequency. The value is stored in the device.
        @param freq: the frequency in Hz (approx.)
        '''
        prescaleValue = 25000000.0  # 25MHz
        prescaleValue /= 4096.0     # 12-bit
        prescaleValue /= float(freq)
        prescaleValue -= 1.0  
        prescale = math.floor(prescaleValue + 0.5)
        oldmode = self._readByte(self._mode_adr)
        if oldmode == None:
            return
        newmode = (oldmode & 0x7F) | 0x10
        self._writeByte(self._mode_adr, newmode)
        self._writeByte(self._prescale_adr, int(math.floor(prescale)))
        self._writeByte(self._mode_adr, oldmode)
        time.sleep(0.005)
        self._writeByte(self._mode_adr, oldmode | 0x80)

    def setDuty(self, channel, duty):
        '''
        Sets a single PWM channel. The value is stored in the device.
        @param channel: one of the channels 0..15
        @param duty: the duty cycle 0..100
        '''
        data = int(duty * 4096 /100)  # 0..4096 (included)
        self._writeByte(self._base_adr_low + 4 * channel, data & 0xFF)
        self._writeByte(self._base_adr_high + 4 * channel, data >> 8)

    def _writeByte(self, reg, value):
        try:
            self.bus.write_byte_data(self.address, reg, value)
        except:
            print("Error while writing to I2C device")

    def _readByte(self, reg):
        try:
            result = self.bus.read_byte_data(self.address, reg)
            return result
        except:
            print("Error while Reading from I2C device")
            return None

调用方法简单：

import time
from smbus2 import SMBus
from pca9685 import PWM

freq = 50
addr = 0x40
channels = [0, 1, 2] 
a = 12.5
b = 2

bus = SMBus(1)
pwm = PWM(bus, addr)
pwm.setFreq(freq)

def setPos(channel, pos):
    duty = a / 180 * pos + b 
    pwm.setDuty(channel, duty)
    time.sleep(0.1)

while True:
    try:
        for pos in range(60, 120, 2):
            setPos(channels[0], pos)
            time.sleep(0.1)

        for pos in range(60, 120, 3):
            setPos(channels[1], pos)
            time.sleep(0.5)

        for pos in range(0, 90, 3):
            setPos(channels[2], pos)
            time.sleep(0.01)

    except KeyboardInterrupt:
        for i in channels:
            setPos(i, 0)
        break

OpenCV 通过颜色检测并控制：

import cv2
import time
import numpy as np
from pca9685 import PWM
from smbus2 import SMBus

freq = 50
addr = 0x40
ch1  = 0
ch2  = 1

bus = SMBus(1)
pwm = PWM(bus, addr)
pwm.setFreq(freq)

def setPos(channel, position):
    pwm.setDuty(channel, position)
    time.sleep(0.01)

def posMap(x, in_min, in_max, out_min, out_max):
    return int((x - in_min) * (out_max - out_min) / (in_max - in_min) + out_min)

cap = cv2.VideoCapture(0)
cap.set(3, 320)
cap.set(4, 160)
ret, frame = cap.read()

cx = int(frame.shape[1] / 2)
center = int(frame.shape[1] / 2)
pos = 90

while True:
    ret, frame = cap.read()
    frame = cv2.flip(frame, 1)
    hsv_frame = cv2.cvtColor(frame, cv2.COLOR_BGR2HSV)
    ly = np.array([51, 0, 0])
    hy = np.array([180, 255, 255])

    mask = cv2.inRange(hsv_frame, ly, hy)
    contours, hir = cv2.findContours(mask, cv2.RETR_TREE, cv2.CHAIN_APPROX_SIMPLE)
    contours = sorted(contours, key=lambda x: cv2.contourArea(x), reverse=True)
    # cv2.drawContours(frame, contours[1], -1, (0, 0, 255), 2)
    for cnt in contours[1:]:
        (x, y, w, h) = cv2.boundingRect(cnt)
        cv2.rectangle(frame, (x, y), (x+w, y+h), (255, 255,0), 2)
        cx = int((x + x + w)/ 2)
        # cv2.drawContours(frame, cnt, -1, (0, 0, 255), 2)
        cv2.line(frame, (cx, 0), (cx, frame.shape[0]), (0, 0, 255), 1)
        if cx < center:
            pos -= 1.5
        elif cx > center:
            pos += 1.5  
        # pos = posMap(pos, 1, 480, 60, 120)
        setPos(ch1, pos)

        break
    # mask = cv2.cvtColor(mask, cv2.COLOR_GRAY2BGR)
    # hstack = np.hstack([mask, frame])
    cv2.imshow('frame', frame)
    if cv2.waitKey(1) & 0xFF == 27:
        break

cap.release()
cv2.destroyAllWindows()
setPos(ch1, 0)

总结

就是控制dutyCycle，频率肯定要50hz，一般的多级可能需要a， b函数作为微调。根据自己的舵机微调a和b的值。

可以尝试 a 设置8.5， b设置2

前言

最近讲监控, 又将很久以前的东西重新拾起来, 尝试在我的小 mac 上的虚拟机里面进行各种安装测试.

安装 SNMP

我以前的印象中,安装 snmp 需要用的软件包的名字是: net-snmp,但是在 ubuntu 上我发现,就直接安装snmp 和 snmpd 好像就行了.

   sudo apt-get -y install snmp snmpd 

安装 cacti

我之前的经验是需要安装一套 LAMP 架构.先.
但是发现 ubuntu 上面已经让我可以变得非常懒了, 直接一条命令就都给你安装上来了.

sudo apt-get -y install cacti

顺手更新了一把系统.

疑惑

官方网站上说 cacti 账户和密码默认是 admin,但是好像不行.真是折腾.
后来用了账户 admin,密码用的数据库的密码才正常登陆.

事情的起因

由于要上课,因此安装 centos8.3.2011一个短命的系统,但是网卡不好用.

实际的电脑信息

我电脑是:

然后我的 parallel 的虚拟机网卡不好用, 尝试联网各种失败,开机无法联网,CentOS 也不好用.

解决方法:

于是搜了半天, 尝试了半天,终于找到一个看似能用的方法:
->关闭所有VMS和Parallels Desktop-

打开终端并键入：sudo -b /Applications/Parallels\Desktop.app/Contents/MacOS/prl_client_app

总结:

Parallels将打开，也许您需要打开您的VM再次...网络将正常工作...
这是一个临时解决方案，而官方的并行支持不起作用..

7D14472C-76FA-4A8B-8F89-48E01C0F066C

5D6EE7E4-2708-4D3E-A977-262828B51A93

68953889-B34D-41D2-A6A5-600B287DCECD

643EB079-4341-4D3C-B4DC-F60840A70FB5

C5155CC5-9E1A-49D3-9BC0-2F33686D5D18

56DC67E6-80C5-4F0D-B8A5-206EA8E9E67E

2021年，想试试幻灯片的功能给我的小破站装修一下.

如果出问题了

如果您遇到的问题似乎是由于内核错误引起的，请检查文件维护者以查看是否有特定人员与您遇到麻烦的内核部分相关联。如果没有列出有任何人，那么第二个最好的办法是将它们邮寄给(torvalds@linux-foundation.org)，以及可能的其他任何有关的邮件列表或新闻组。

在所有错误报告中，请告诉您您在操作的是什么内核，如何复制问题以及您的设置是什么（使用常识）。如果问题是新问题，请告诉他，如果问题是旧问题，请在首次发现问题时告诉他，他才是Linux内核最大的BOSS。

例如：
如果该错误导致显示如下消息：

unable to handle kernel paging request at address C0000010
Oops: 0002
EIP:   0010:XXXXXXXX
eax: xxxxxxxx   ebx: xxxxxxxx   ecx: xxxxxxxx   edx: xxxxxxxx
esi: xxxxxxxx   edi: xxxxxxxx   ebp: xxxxxxxx
ds: xxxx  es: xxxx  fs: xxxx  gs: xxxx
Pid: xx, process nr: xx
xx xx xx xx xx xx xx xx xx xx

或屏幕上或系统日志中的类似内核调试信息，请准确复制。转储对您来说似乎令人难以理解，但是它确实包含有助于调试问题的信息。转储上方的文本也很重要：它说明了为什么内核转储代码的原因（在上面的示例中，这是由于内核指针错误引起的）。有关寻找转储的更多信息，请参见“错误查找”

如果使用CONFIG_KALLSYMS编译内核，则可以按原样发送转储，否则，您将不得不使用该ksymoops程序来理解转储（但通常首选使用CONFIG_KALLSYMS进行编译）。可以从https://www.kernel.org/pub/linux/utils/kernel/ksymoops/下载该实用程序 。另外，您可以手动进行转储查找：

在上面的调试转储中，如果您可以查找EIP值的含义，那么它将大有帮助。这样的十六进制值对我或其他人没有太大帮助：它将取决于您的特定内核设置。您应该做的是从EIP行中获取十六进制值（忽略0010:），然后在内核名称列表中查找该值，以查看哪个内核函数包含有问题的地址。

要找出内核函数名称，您需要找到与表现出该症状的内核相关的系统二进制文件。这是文件“ linux/vmlinux”。要提取名称列表并将其与内核崩溃时的EIP匹配，请执行以下操作：

nm vmlinux | sort | less

这将为您提供按升序排序的内核地址列表，从中可以轻松找到包含违规地址的函数。请注意，内核调试消息给出的地址不一定与功能地址完全匹配（实际上，这不太可能），因此您不能只是“ grep”列表：但是，列表将为您提供每个内核函数的起始点，因此，通过查找起始地址比您要查找的起始地址低的函数，然后查找具有较高地址的函数，可以找到所需的起始地址。实际上，在问题报告中包括一些“上下文”可能是个好主意，在有趣的地方加上几行即可。

如果由于某种原因您不能执行上述操作（您有预编译的内核映像或类似文件），请尽可能多地告诉他有关您的设置的信息。有关详细信息，请阅读“报告问题”。

另外，您可以在运行的内核上使用gdb。（只读；即，您不能更改值或设置断点。）为此，首先使用-g编译内核；适当地编辑arch/x86/Makefile，然后执行。您还需要启用CONFIG_PROC_FS（通过）。

make cleanmake config

使用新内核重新引导后，请执行。现在，您可以使用所有常用的gdb命令。查找系统崩溃点的命令是。（用EIP值替换XXXes。）

gdb vmlinux /proc/kcorel *0xXXXXXXXX

当前，对非运行内核进行gdb操作失败，因为gdb（错误地）无视了为其编译内核的起始偏移量。