For this week assignment the objectives were:
- group assignment:
- probe an input device’s analog levels and digital signals.
- individual assignment:
- measure something: add a sensor to a micro controller board that you have designed and read it.
1. Group Assignment
The full group assigment can be found:
The full group assignment can be found
here
For the group assignment I was in charge of measuring digital signals produced by pressing a button. For this purpose I used the board I had designed in the electronic production week. I connected the digital analyzer to the ground of the board and to Pin number eight that is connected to the button circuit. Whenever the button was pressed I could see a 1 while it was not pressed a 0. The one signal would be as long as the button was pressed.
2. Individual Assignment
For the individual assignment I choose the DHT11 sensor. This sensor is capable of measuring both Temperature and Relative Humidity. From this I am interested particularly in temperature as it has a very important impact on insect development. Development in poikilotherms that is, animals that do not regulate their body temperature physiologically, are impacted by the environmental temperature which in turn will impact in their developmental rates. Developmental rates is the time they take to pass from one life stage to the other (emergence of a butterflty from a chyrsais). Higher temperatures will lead to higher developmental rates up to a point, were temperature will impact developmental rates negatively and eventually kill the organism. Given that we know an infestation date due to the presence/discovery of a pest life stage we could predict the life stages presence of the insect at different points in time and tune control measures being biological or chemical. To give a general idea of the matter below you can an adult whitefly which is a very common pest of several species of plants. Below you can see the developmental rates (1/time) at different temperatures. For example at 27 degrees Celcius an egg laid by an adult would take about 6 days to hatch.
The DHT11
For my assignment I worked with the DHT11 sensor, this is a sensor used to measure temperature and humidity.A humidity sensing component measures humidity using two electrodes with a moisture-holding substrate, typically salt or a conductive plastic polymer, between them. As the substrate absorbs water vapor, it releases ions, increasing conductivity between the electrodes. The resistance between the electrodes changes in proportion to relative humidity—higher humidity decreases resistance, while lower humidity increases it.
The DHT11 sensor measures temperature using an NTC thermistor, a resistor whose resistance decreases as temperature rises. Unlike standard resistors, thermistors exhibit significant resistance changes, often 100Ω or more per degreew (summarized from AZ electronics).
Below you can find a technical data sheet for the DHT11 sensor.
Specifications
source: mouser electronics
From the chart there are a few points are interesting to comment. First, the resolution that in temperature and RH wont go below 1C or 1% in RH. Also that the Relative humidity sensor will have some sensing restrictions depending at which temperature is working at. Similarly, we can observe that the typical accuracy in RH of 4% and 1 to 2C in temperature.
The sensor needs 3 pins: a votage,
ground and data pin. Note that the data pins needs to be a SDA pin. In
terms on circuit design, it is recommended to to have a pull up
resistor, this is something I did not take into account
in my test board design and needs to be addressed for future boards.
Single-bus data format is used for the communication and synchronization between the MCU and the DHT11. There is a 40 bit transmission consisting of:
- 8 bit of integral data RH
- 8 bit of decimal data RH
- 8 bit of integral data T
- 8 bit of decimal data T
- 8 bit check sum.
Communication Process:
- Start Signal send from MCU [this is necessary]
- DHT11 changes from low power to running mode and waits for MCU to complete signal
- DHT11 sends 40 bit data.
For testing the sensor I used the followiong script in micropython.
from machine import Pin
from time import sleep
import dht
sensor = dht.DHT11(Pin(5))
while True:
sleep(2)
sensor.measure()
temp = sensor.temperature()
hum = sensor.humidity()
print("Temperature ",temp," Relative Humidiy ",hum)
Finally, some additional points on the differences between DHT11 and DHT22 sensors (ariat tech) You can see from the chart that the error rate higher in the DHT11 specially for RH. Also there a finer resolution for measuring temperature in the DHT22. However the price is higher in the latter.
I also looked at some tutorial to set up the XIAO as a server for displaying a value of temperature and a value humidity connecting to Wifi (link ).
I wanted to have a more complete view of how these parameters evolved and modify the script to record values every 10 seconds and plot them using a neat javascript interactive library called “Plotly”.
Something to note here is the sharp drop in RH, possibly this might be due to the precision of the sensor.
Xiao ESP32S3 sense Camera
For my project I need to use a camera in for object detection. This camera should be mounted in the trap and face toward the sticky tape. The camera that comes with the sense is detachable OV2640 sensor with a 1600x1200 pixel resolution. It is possible to upgrade this camera for the OV5640 camera with 5MP and supposedly autofocus althouhg still have to find someone who has successuflly tested it. The field of view is 65 degrees.
https://www.waveshare.com/wiki/OV2640_Camera_Board
An important task for the final project is to check if the field depth of the camera is adequate for the project given that the tape will be quite close to the camera. Also whether the field of view will be enough to cover a good area of the tape, one additional thing to check is how dark or light the pictures are. For this I constructed a quick prototype in mdf. Below, you can see some of the Fusion sketches. One important requirement in order to standardize the quality of the photos is that the tape rests flat against the board. Finally, depending on the angle the camera is positioned it might be necessary to ortho-rectify the picture (post processing).
Here you can see the testing set up. Composed of a house like structure,
a roll of sticky tape,
and camera mounted in the roof with case I found here: link
Thereafter I needed to operate the camera and send the pictures over the internet. I decided to implement this part in micropython as it is a more intuitive programming language than C++. Its beneficial to stay in python as it has deep learning libraries and as in C++ openCV library.
I found these web sites useful to try to understand how to connect to WIFI and send information between devices. Most of the code below is adapted or from this sites (video, link 1, link 2 ). I also used ChatGPT to correct or provide suggestions on how to solve certain issues/errors.
Camera Control
import camera
try:
print('Taking a photo')
camera.init()
buffer = camera.capture()
print(len(buffer))
file_path = "captured_image.jpg"
file= open(file_path,"w")
file.write(buffer)
file.close()
except Exception as e:
print('An issue ocurred')
finally:
print('Deintializing Camera and Wifi')
camera.deinit()Server Side
import socket
port = 80
address = socket.getaddrinfo('0.0.0.0', port)[0][-1]
server = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
server.bind(address)
server.listen()
print(f"Server listening on port {port}...")
client_socket, client_address = server.accept()
f=open("server_image2.jpg", "wb")
image_chunk = client_socket.recv(2048)
while image_chunk:
f.write(image_chunk)
image_chunk = client_socket.recv(2048)
f.close
client_socket.close()Client Side
import network
import socket
import time
# Wi-Fi Connection
sta = network.WLAN(network.STA_IF)
if not sta.isconnected():
print('Connecting to network...')
sta.active(True)
sta.connect('*****', '*****') # Replace with actual credentials
while not sta.isconnected():
pass
print('Network config:', sta.ifconfig())
server_ip = "192.168.0.26" # This is the IP address of the computer.
port = 80
### This for connecting
client = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
client.connect((server_ip, port)) # Use correct port
file = open("captured_image.jpg", "rb")
image_data = file.read(2048)
while image_data:
client.send(image_data)
image_data = file.read(2048)
file.close()
client.close()
After I sorted out the programming problems I took some pictures. I placed the camera at different distances from the edge. I found that the position the furthest away from the tape was best as it allow to cover more area. It was evident that the light was not enough, however, it remains to see if by using an acrylic roof this could be improved.
Key Learnings
- I do need a pull-up resistor for the DHT sensor.
- Given the resolution and error rate it might be beneficial to consider a DHT22.
- You can use the Xiao micro controller as a network station/router.
- Did learn a fair bit on networking it will be important in the following weeks.
- Light is super important to get good pictures with the OV
- I might need a wider angle camera maybe this will help
- For the OV2640 definitely it will be necessary to have a transparent roof possibly acrylic?
References
Aregbesola, O. Z., Legg, J. P., Lund, O. S., Sigsgaard, L., Sporleder, M., Carhuapoma, P., & Rapisarda, C. (2020). Life history and temperature-dependence of cassava-colonising populations of Bemisia tabaci. Journal of Pest Science, 93, 1225–1241. https://doi.org/10.1007/s10340-020-01249-z