You can find a complete list of the equipment needed to build your sensor in the Getting Started page.

Firstly, make sure if you are connected to the Internet. Secondly, make sure that the network you are connected to allows hosting a local webserver. If everything is fine with the Internet, you may be missing some required packages. Use the following commands and then go through the installation guide again.

It is possible to find out the IP address using the command ifconfig in the terminal. If you are connected to WiFi the address will be displayed in the wlan0 section. If you are connected via Ethernet cable the address will be displayed in the eth0 section. In either case, you need to look for a line similar to inet addr: 10.99.78.101.

If there is an error you can go back to the home page by typing in the webserver address in your browser again. If the error persists, go through the installation steps again as there may be a package missing.

It is available a Raspberry Pi Camera Extension kit.

If the Raspberry Pi restarts the raspistillWeb application needs to be initialized again from the command line as follows: There is the a way to have the raspistillWeb application start automatically with the following steps.

1. Create a startup script
   $ sudo nano /etc/init.d/raspistillweb
2. Type in the following lines save and exit.
#!/bin/bash cd /home/pi/Development/env/raspistillWeb ../bin/pserve development.ini
3. Make the script executable:
   $ sudo chmod 755 /etc/init.d/raspistillweb
4. Register the script to be run at startup:
   $ sudo update-rc.d /etc/init.d/raspistillweb defaults

Generally speaking, it is possible. If your institute has some firewall blocking any access from the outside, you need to contact your IT service and ask how to configure a VPN in your computer. By configuring a VPN, you can access your Raspberry Pi using the local IP address. Another way we recommend is to install TeamViewer on your Raspberry Pi and your computer.

You can remove raspstillWeb by using the following command: If you have also the start-up script (see above), you may also run the following commands:

Generally, this problem appears when you do not have a suitable C compiler installed and/or configured in your system. Please, follow this instruction for your operating system to solve the issue. If you believe the problem is something else, do not hesitate to contact us.

If you wish to share data with colleagues or move the them to another workstation, you simply need to copy the plant's image files and the database file with .mat extension that was generated by the modules. When you load your data on the new workstation, the software will ask you to provide the path where images are located. Please, provide the new location of the images on the new workstation.

Unfortunately, it is not possible to abort a computation. Since our software is based on MATLAB and it is single-threaded, we could not add any mechanism to cancel long computations. However, if you are running Phenotiki from the Source Code, you may press CTRL+Z, as you would do for any computation in MATLAB, to abort.

If you have errors saying \textit{mclmcr.dll not found}, it could be either of the following: (i) MCR not installed, (ii) MCR installation corrupted, (iii) Older version of the library installed in a different directory. Specifically, we suggest:

1. Download the Installer of Phenotiki and/or make sure you are installing MCR
2. Different solutions can be found on-line. Search for MCR on your system and delete all the files related to it. Then, go to the point (i)
3. Search on your system for the mclmcr.dll. It can happen it is located within the System32 directory of Windows. Locate and remove it. Then, go to the step (ii)

Sometimes on Linux systems, you may experience the error "An error occurred while extracting files". If you have the zip file containing Phenotiki in your Desktop folder, try to move it file in any other folder different than the Desktop and try again.

We experienced issues using MATLAB/Phenotiki on Parallels (for Mac), depending on MATLAB and we were unable to fix them. We do not recommend using Parallels with Phenotiki on a virtual Linux Operating System.

Finding good parameters for the plant segmentation may not be immediate for challenging datasets. If you have fair results and you want to improve them, then we suggest to select a few images to trial different configurations. In the unluckily case that segmentation results are not accurate enough, we suggest to run a grid search. With this method, the software will automatically test a range of parameters and find optimal ones according to some supervision. First, segment manually a few of plant images using an external image editing software (e.g., GIMP, Photoshop). Create a black-and-white (also known as binary) image with just 0's and 1's, where 0 (black) marks background area and 1 (white) marks plant area. When you are done with this, open the Phenotiki Analysis Software and go to Pot Tray Analysis tool and load your dataset. Then, from the list of images, select one plant image for which you have manually created the foreground/background segmentation and press the right-button of your mouse, then select Set ground-truth. When the file selection window appears, select the corresponding mask for this image and repeat this operation for all the plant images for which you have manually obtained the segmentation mask. Then press on the Setting button in the bottom part of the Pot Tray Analysis Tool and press on Advanced Options. The full list of parameters will appear and you may choose the ones you want to include in the grid search, by ticking the check-boxes in the list. When you choose a parameter, you may define a range of values that the software will test against the ground-truth you have provided before. Ranges can be set in the following way.

• Numeric values: MATLAB syntax is accepted. For example:
  • explicit list of values to test: [0, 10, 20, 30, 40, 50];
  • all the values between 0 and 50 with a step of 10: 0:10:50;
• Text values: cell array MATLAB syntax is accepted. Some examples are:
  • {Feature1, Feature2}: the software will use Feature1 and Feature2 separately;
  • {Feature1,Feature2}, {Feature1,Feature3}: the software will test the best combination between Feature1 and Features2, or Feature1 and Feature3.
• Boolean values: for Boolean values you do not need to specify range of values, as they can be either 0 or 1.

Whilst integers parameters may accept any value, we suggest you to vary them by small amounts. For further details on the pararemters selection, we reccomand to read the corresponding sections in the user manual.

We suggest to add an annotation with the same label to the different part of the leaf. Once you annotated one part, select from the annotation list the label corresponding to the newly created label and annotated the other part of the leaf. This will result with two annotations with the same label. Once you are finished, select the <new label> to annotate a new leaf.

If you want to assign the same label to the same leaf across time, we recommend to start the annotation from the first time instant (or the last). Then, move to the next time instant and, from the list of labels, click with the right-button of your mouse and select Copy from the previous plant (or Copy from the next plant). In this way, annotations are copied with consistent numeric identifiers. Then, adapts the copied labels to adjust to growth and movement of the plant and add new labels if necessary.

Use more refined annotation (such as lines or freehand scribbles) to better delineate the image region of each leaf.