Difference between revisions of "Team:Aix-Marseille/Hardware"

 
(26 intermediate revisions by 5 users not shown)
Line 1: Line 1:
−
{{Aix-Marseille|title=Detection of the disease|toc=__TOC__}}
+
{{Aix-Marseille|title=Detection of the disease|toc=__noTOC__}}
  
 +
[[File:T--Aix-Marseille--transmission chain-1.png|right|400px|The hydric stress detection method.]]
  
−
The purpose of the detection would be to determine if the tree is infected by X.F. For this, there is a method called PCR (Polymerase Chain Reaction) that can detect the presence of viruses or measure viral loads. However, this is a complex method that requires DNA sampling on the tree, which will undergo various treatments in the laboratory, so it is not a way to detect X.F. on the field.
+
[[Team:Aix-Marseille/Project|'''KILL XYL''']] aims not only to cure the disease caused by [[Team:Aix-Marseille/Xylella_fastidiosa|''Xylella fastidiosa'']], but also to detect it. Nowadays, the most effective way to detect the bacteria is a method called PCR (Polymerase Chain Reaction).  
 +
However, this method is complex and requires DNA samples from trees and lengthy and complex laboratory treatments.
  
−
Therefore, this project focuses on detecting the drying of the tree which is one of the first symptoms of X.F. In the case of the olive tree, this method is relevant because the olive tree is an evergreen tree, its leaves will not dry naturally if the tree is healthy. However, the drying of the tree can be caused by other factors than XF, the solution presented here serves primarily as a warning device, it is appropriate after detecting the drying of a tree to take a sample and to ensure that it is XF which infects the tree before injecting the remedy developed by the team.
+
Therefore, we focused on another method which will allow us to work more easily in the open with hundreds of acres of crops.
 +
For this, we thought about detecting the first symptoms of the disease: hydric stress.  
 +
The easiest way to measure the hydric stress is by assessing leaf dryness.  
  
−
You can see below a diagram representing a chain of data transmission that would allow you to observe the status of one or more trees in real time. This report relates mainly on the sensor and the data processing part.
+
[[File:T--Aix-Marseille--refraction spectrum leaves-1.png|400px|right|thumb|Healthy leaves, because of photosynthesis, refract more infrared (IR) light than dry leaves.]]
  
−
[[File:T--Aix-Marseille--transmission chain.png]]
+
Olive trees are evergreen which means that leaves will not dry naturally.
 +
However,  dry leaves can be a consequence of other factors than the disease caused by [[Team:Aix-Marseille/Xylella_fastidiosa|''X. fastidiosa'']].
 +
Thus our solution serves primarily as a warning device, it's appropriate to the detection of hydric stress in crops.
 +
If any is detected, then DNA samples should be taken to verify the presence of the bacteria.
  
−
To observe the dryness of the tree, we use NDVI (Normalized Difference Vegetation Index) images, this allows to detect the level of photosynthesis of the leaves of a tree.
+
[[File:T--Aix-Marseille--NDVI comparison-1.jpg|400px|right|thumb|K-State Research and Extension
 +
Soybean NVDI photo]]
  
−
You can see on the diagram below that healthy leaves return more infrared light, there is an index, the NDVI, which allows to determine this level of photosynthesis of a plant.
+
As healthy leaves reflect more infrared light than dry leaves we can calculate the Normalized Difference Vegetation Index (NDVI), which allows determining the relative level of photosynthesis of a plant. NDVI is calculated by : $$\text{NDVI} = \frac{\text{NIR}-\text{RED}}{\text{NIR}+\text{RED}}$$
  
−
$$\text{Index NDVI} = \frac{\text{NIR}-\text{RED}}{\text{NIR}+\text{RED}}$$
+
The calculated index is between -1 and 1 and is associated with a color scale (1, red and -1, blue) that allows you to observe easily and quickly if the tree is healthy or not.
  
−
[[File:T--Aix-Marseille--refraction spectrum leaves.png|400px]]
+
==Detection==
  
−
The calculated index is between -1 and 1 and it is associated with a color scale that allows you to observe easily and quickly if the tree is healthy or not, you can see an example of NDVI image below.
+
First, we built the camera.
 +
For this, we used a Raspberry Pi 2 which supports the necessary acquisition and treatment software and a Black Pi camera.
 +
The infrared blocking filter was removed from the camera and replaced with a Blue filter (ROSCO#2007).
 +
This modification allows us to detect the NIR and RED light necessary to construct the NDVI image.
  
−
[[File:T--Aix-Marseille--NDVI comparison.png]]
+
==Software==
  
−
==Sensor Operation==
+
[[File:T--Aix-Marseille--NDVI-process.png|right|500px|thumb|]]
  
−
We use a Raspberry Pi 2 which will support the different software and programs used, a Black Pi camera (an infrared camera) and a ROSCO # 2007 blue filter, which is used for the detection of the dewatering of the tree. the manufacture of the NDVI image.
+
We used Python and OpenCV to calculate our NDVI images.  
−
 
+
NDVI index computation is done pixel by pixel on the image to create a new grayscale image, which is then shown with a linear colormap (jet) for a better visual contrast.
−
===Pi camera===
+
−
 
+
−
[[File:T--Aix-Marseille--Pi camera and operation normal camera.png]]
+
−
 
+
−
Above, the black Pi camera used is shown on the left, and on the right the operation of a normal camera, an infrared camera works as a normal camera but without the anti-infrared filter.
+
−
 
+
−
However, we add another filter in front of the camera that will be used to create the NDVI image.
+
−
 
+
−
===Blue filter ROSCO#2007===
+
−
 
+
−
[[File:T--Aix-Marseille--blue filter rosco.png|left]]
+
−
 
+
−
This filter is used to filter the light and keep only blue and near infrared (see graph below).
+
−
 
+
−
[[File:T--Aix-Marseille--graph filter rosco.png]]
+

Latest revision as of 02:29, 2 November 2017

Detection of the disease

The hydric stress detection method.

KILL XYL aims not only to cure the disease caused by Xylella fastidiosa, but also to detect it. Nowadays, the most effective way to detect the bacteria is a method called PCR (Polymerase Chain Reaction). However, this method is complex and requires DNA samples from trees and lengthy and complex laboratory treatments.

Therefore, we focused on another method which will allow us to work more easily in the open with hundreds of acres of crops. For this, we thought about detecting the first symptoms of the disease: hydric stress. The easiest way to measure the hydric stress is by assessing leaf dryness.

Healthy leaves, because of photosynthesis, refract more infrared (IR) light than dry leaves.

Olive trees are evergreen which means that leaves will not dry naturally. However, dry leaves can be a consequence of other factors than the disease caused by X. fastidiosa. Thus our solution serves primarily as a warning device, it's appropriate to the detection of hydric stress in crops. If any is detected, then DNA samples should be taken to verify the presence of the bacteria.

K-State Research and Extension Soybean NVDI photo

As healthy leaves reflect more infrared light than dry leaves we can calculate the Normalized Difference Vegetation Index (NDVI), which allows determining the relative level of photosynthesis of a plant. NDVI is calculated by : $$\text{NDVI} = \frac{\text{NIR}-\text{RED}}{\text{NIR}+\text{RED}}$$

The calculated index is between -1 and 1 and is associated with a color scale (1, red and -1, blue) that allows you to observe easily and quickly if the tree is healthy or not.

Detection

First, we built the camera. For this, we used a Raspberry Pi 2 which supports the necessary acquisition and treatment software and a Black Pi camera. The infrared blocking filter was removed from the camera and replaced with a Blue filter (ROSCO#2007). This modification allows us to detect the NIR and RED light necessary to construct the NDVI image.

Software

T--Aix-Marseille--NDVI-process.png

We used Python and OpenCV to calculate our NDVI images. NDVI index computation is done pixel by pixel on the image to create a new grayscale image, which is then shown with a linear colormap (jet) for a better visual contrast.