The HuskyLens is an easy-to-use AI machine vision sensor. It is equipped with multiple functions such as:
Face recognition
Object tracking
Object recognition
Line trace
Color recognition
Tag recognition (QR code).
Via the UART / I2C port you can among others: boards connect:
Arduino
micro:bit
Raspberry Pi
Steps to take: Press Face detection, when a cross in a square is displayed, press the button on your HuskyLens
Set your husky protocol to I2C in the settings.
Minimal Code needed
/***************************************************
HUSKYLENS An Easy-to-use AI Machine Vision Sensor
<https://www.dfrobot.com/product-1922.html>
****************************************************/
#include "HUSKYLENS.h"
HUSKYLENS huskylens;
//HUSKYLENS green line >> SDA; blue line >> SCL
int ID0 = 0; //not learned results. Grey result on HUSKYLENS screen
int ID1 = 1; //first learned results. colored result on HUSKYLENS screen
int ID2 = 2; //second learned results. colored result on HUSKYLENS screen
// and so on.....
int arjprevious = 0;
void printResult(HUSKYLENSResult result);
void setup() {
Serial.begin(115200);
Wire.begin();
while (!huskylens.begin(Wire))
{
Serial.println(F("Begin failed!"));
Serial.println(F("1.Please recheck the \"Protocol Type\" in HUSKYLENS (General Settings>>Protocol Type>>I2C)"));
Serial.println(F("2.Please recheck the connection."));
delay(100);
}
huskylens.writeAlgorithm(ALGORITHM_FACE_RECOGNITION);
}
void loop() {
if (huskylens.requestLearned()) //request blocks and arrows tangged ID != 0 from HUSKYLENS
if (huskylens.requestBlocksLearned()) //request blocks tangged ID != ID0 from HUSKYLENS
{
for (int i = 0; i < huskylens.countArrows(ID0); i++)
{
HUSKYLENSResult result = huskylens.getArrow(ID0, i);
}
int arj = huskylens.count(ID1);
if ( arj != arjprevious )
{
if ( arj == 1 )
{
Serial.println("Learned face detected");
}
else
{
Serial.println("Learned face not detected");
}
arjprevious = arj;
}
}
else
{
Serial.println("Fail to request objects from Huskylens!");
}
}
Learned face detected ID1
Learned face not detected
Learned face detected ID1
Learned face not detected
Learned face detected ID1
Learned face not detected
Now i’ve changed the notification flag for Email using Thunderbird. Just connect the wemos to a USB on your computer, no mqtt/wifi needed. (On the road solution)
Steps:
Install FiltaQuilla Add-on in thunderbird select run program in config.
Next create a filter
Create two bash files (i’ve got mine in ~/bin/ ) Change ttyUSB0 if needed
So many things to try .. and probably buy. I’m lucky to have this already, official it’s not even out yet.
Specification
Processor : Broadcom BCM2712 2.4GHz quad-core 64-bit Arm Cortex-A76 CPU, B with cryptography extensions, 512KB per-core L2 caches, and a 2MB shared L3 cache
Features:
VideoCore VII GPU, supporting OpenGL ES 3.1, Vulkan 1.2
Dual 4Kp60 HDMI® display output with HDR support
4Kp60 HEVC decoder
LPDDR4X-4267 SDRAM (4GB and 8GB SKUs available at launch)
Dual-band 802.11ac Wi-Fi
Bluetooth 5.0 / Bluetooth Low Energy (BLE)
microSD card slot, with support for high-speed SDR104 mode
2 × USB 3.0 ports, supporting simultaneous 5Gbps operation
2 × USB 2.0 ports
Gigabit Ethernet, with PoE+ support (requires separate PoE+ HAT)
2 × 4-lane MIPI camera/display transceivers
PCIe 2.0 x1 interface for fast peripherals (requires separate M.2 HAT or other adapter)
5V/5A DC power via USB-C, with Power Delivery support
Raspberry Pi standard 40-pin header
Real-time clock (RTC), powered from external battery
Power button
Some things come to mind to test:
Kubernetes
Dual Camera OpenCV – depthmap and more
The offset between the normal camera and the Flir, is due to being to close to the object.After 5 minutes Wifi/Bluetooth bottom left heats up (didn’t use in this test)
You really want to use a case with an active blower to cool the rpi.
Example uses a shelly sensor, which has a offset mode in its new firmware. So below is not needed any more. But the example can be used for any calibration/adjustments.
I’ve put two examples in the NodeRed function
First the solution
var temperature=msg.payload.tC;
var humidity=msg.payload.tF;
// calc new offset example
// simple offset + 2.3 off real measurement
msg.payload.tC=temperature + 2.3;
// more complex example
// take two measurements (with a big difference)
msg.payload.tF=1.11 * (humidity - 1);
return msg;
First adjustment is plus or minus a value. Second is more precise when the temperature needs more adjusting
Dots on the red line are measured values Blue is how it should be
So measured was 2.8 but should be 2 And measured 14.5 but needs to be 15
slope = (14.5-2.8)/(15-2) = 0.9
To get the multiplication factor = 1/0.9 = 1.1111
=(2.8-heightadjust)*1,1111 should give us 2
2/1.1111 = 2.8 – heightadjust
1.8 = 2.8 – heightadjust = 1
So the formula is realhumid = 1,1111 * ( measuredhumid – adjust )
Under hardware in domoticz install the mqtt broker. Configure your mqtt server (mine is mosquitto) Add devices to floorplans to send these to Mosquitto
Now i can see the topics in MQTTExplorer
Install the home assistant websocket palette in NodeRed
Configure the nodes like above
These are some example flows (I’ll put the function code below)
Above are the NAMED entities in Home Assistant
CODE
//Code for NR temperature filtering on domoticz IDX
//A door censor uses nvalue instead of svalue1
//Humidity can be svalue2
//Check the Mqtt payload in MQTTExplorer which to use!
var varidx = msg.payload.idx;
var varnvalue = msg.payload.svalue1;
if(varidx == 3108)
{
msg.payload = {};
msg.payload = varnvalue;
return msg;
}
Example of the trigger node. When a temperature sensor battery dies, and no new data arrives- in 1 hour, I get a pushover warning. (use pushbullet/email/TV notification whatever you want)
Our bedroom has a shelly dimmer, this one is connected with a wall switch and is being controlled by Domoticz, Home Assistant and NodeRed.
I had to fix some stuff, so this was a perfect time to jot down some notes.
I’ve wired it up like this: (I’ve got a Line wire in my ceiling socket, so i placed the module there instead of the wall socket)
Configure the Shelly as mentioned in the manual. After that do the following: Advance > Developer settings : (Enable CoIot if you want a easy auto detect for Home Assistant) Enable MQTT (This will DISABLE cloud ! ) Server: the ip number of your Mqtt Broker
Next I did was:
Now the wall switch will change/toggle what the current state is. (If the light is off, and you switch it on using Mqtt, you probably needed to hit the switch two times to turn it off again. Not so with this setting.)
My Nodered Flow ( Not using the Shelly palette nodes in this example)
The Mqtt IN node sets the state of the switch when you use the wall switch! Cozy turns light on at a specific level. The slider send an off command when 0% selected, else an on command and the sliders value.
A ledstrip on a Harp, now I can see the strings at night!
Woohoo .. ( Little harp in the background (autoharp))
Well .. its a pity that the distance of the leds is NOT the same as the strings distance.
I could light up the string to be played, or even cooler … When using FFT code (Fast Fourier Transform), I could light up the string being played! I’m probably going to try to implement this at a later time.
"If something is worth doing, it's worth overdoing."
