![Institutional Repository [SANDBOX]](images/dias.logo.png "Institutional Repository [SANDBOX]"){kind=logo}
Institutional Repository [SANDBOX]
Technical University of Crete
Technical University of Crete
EN
|
EL
| URI: | http://purl.tuc.gr/dl/dias/D75328D6-C174-47A6-92ED-7B3FCD943C7C | ||
| Year | 2020 | ||
| Type of Item | Diploma Work | ||
| License |
|
||
| Bibliographic Citation | Minas Pantelidakis, "CIWA – A web-based tool for automated irrigation management using intelligent plant canopy temperature identification techniques ", Diploma Work, School of Electrical and Computer Engineering, Technical University of Crete, Chania, Greece, 2020 https://doi.org/10.26233/heallink.tuc.87605 | ||
| Appears in Collections |
Canopy temperature has been recognized as a crop water stress indicator, since it reflects the interaction of crops with soil and the atmosphere. Even though there are several methods to estimate canopy temperature, most of them are time consuming, expensive, inaccurate, or require considerable human input. This work mobilizes Convolutional Neural Networks (CNNs) to identify sunlit leaves and in conjuction with thermal imagery, find the underlying leaf temperatures and calculate the Crop Water Stress Index (CWSI). The results of two different CNN architectures (FRRN, DeepLabV3) have been compared with two minimum input, state of the art methods, namely temperature Histogram Gradient Thresholding and Gaussian Mixture Models. We evaluate our approaches against this baseline using our own dataset of 1432 image-label pairs of pistachio trees. Results indicate that CNNs outperform existing methods. Our dataset is released for the scientific community to use. Finally, a web application was developed, so that researchers/growers can calculate the CWSI in real time, using pictures captured with a thermal camera.
Copyright © DIAS 2013
