Pix2Pix Network to Estimate Agricultural Near Infrared Images from RGB Data
Remote sensing has been applied to agriculture, making it possible to acquire a large amount of data far away from crops, providing information for decision making by producers that can impact production costs and crops quality. One way of getting the production information is through vegetation ind...
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Taylor & Francis Group
2022-03-01
|
| Series: | Canadian Journal of Remote Sensing |
| Online Access: | http://dx.doi.org/10.1080/07038992.2021.2016056 |
| _version_ | 1797661137874976768 |
|---|---|
| author | Daniel Caio de Lima Diego Saqui Steve Ataky Tsham Mpinda José Hiroki Saito |
| author_facet | Daniel Caio de Lima Diego Saqui Steve Ataky Tsham Mpinda José Hiroki Saito |
| author_sort | Daniel Caio de Lima |
| collection | DOAJ |
| description | Remote sensing has been applied to agriculture, making it possible to acquire a large amount of data far away from crops, providing information for decision making by producers that can impact production costs and crops quality. One way of getting the production information is through vegetation indices, arithmetic operations that use spectral bands, especially the Near Infrared (NIR). However, sensors that capture this spectral information are very expensive for small producers to afford it. In a previous article, a pixel-to-pixel image synthesis model to estimate NIR images from RGB data using hyperspectral endmembers (pure hyperspectral signatures) was described. In this work, an image-to-image synthesis model, known as Pix2Pix, is used for estimating NIR images from low-cost RGB camera images. Pix2Pix is a kind of Generative Adversarial Networks (GANs), composed by two neural networks, a generator (G) and a discriminator (D), that compete. G learns to create images from a random noise inputs and D learns to verify if these images are real or fake. The results showed that the presented method generated NIR images quite similar to real ones, reaching a value of 0.912 on M3SIM similarity metric, outperforming results obtained with the previous endmembers method (0.775 on M3SIM). |
| first_indexed | 2024-03-11T18:40:05Z |
| format | Article |
| id | doaj.art-b3ceadc6544242f8a5ec6afa9ab6229b |
| institution | Directory Open Access Journal |
| issn | 1712-7971 |
| language | English |
| last_indexed | 2024-03-11T18:40:05Z |
| publishDate | 2022-03-01 |
| publisher | Taylor & Francis Group |
| record_format | Article |
| series | Canadian Journal of Remote Sensing |
| spelling | doaj.art-b3ceadc6544242f8a5ec6afa9ab6229b2023-10-12T13:36:24ZengTaylor & Francis GroupCanadian Journal of Remote Sensing1712-79712022-03-0148229931510.1080/07038992.2021.20160562016056Pix2Pix Network to Estimate Agricultural Near Infrared Images from RGB DataDaniel Caio de Lima0Diego Saqui1Steve Ataky Tsham Mpinda2José Hiroki Saito3Computer Department, UFSCar – Federal University of São CarlosIFSULDEMINAS, Federal Institute of Southern Minas GeraisUQAM, Université du Québec à MontréalComputer Department, UFSCar – Federal University of São CarlosRemote sensing has been applied to agriculture, making it possible to acquire a large amount of data far away from crops, providing information for decision making by producers that can impact production costs and crops quality. One way of getting the production information is through vegetation indices, arithmetic operations that use spectral bands, especially the Near Infrared (NIR). However, sensors that capture this spectral information are very expensive for small producers to afford it. In a previous article, a pixel-to-pixel image synthesis model to estimate NIR images from RGB data using hyperspectral endmembers (pure hyperspectral signatures) was described. In this work, an image-to-image synthesis model, known as Pix2Pix, is used for estimating NIR images from low-cost RGB camera images. Pix2Pix is a kind of Generative Adversarial Networks (GANs), composed by two neural networks, a generator (G) and a discriminator (D), that compete. G learns to create images from a random noise inputs and D learns to verify if these images are real or fake. The results showed that the presented method generated NIR images quite similar to real ones, reaching a value of 0.912 on M3SIM similarity metric, outperforming results obtained with the previous endmembers method (0.775 on M3SIM).http://dx.doi.org/10.1080/07038992.2021.2016056 |
| spellingShingle | Daniel Caio de Lima Diego Saqui Steve Ataky Tsham Mpinda José Hiroki Saito Pix2Pix Network to Estimate Agricultural Near Infrared Images from RGB Data Canadian Journal of Remote Sensing |
| title | Pix2Pix Network to Estimate Agricultural Near Infrared Images from RGB Data |
| title_full | Pix2Pix Network to Estimate Agricultural Near Infrared Images from RGB Data |
| title_fullStr | Pix2Pix Network to Estimate Agricultural Near Infrared Images from RGB Data |
| title_full_unstemmed | Pix2Pix Network to Estimate Agricultural Near Infrared Images from RGB Data |
| title_short | Pix2Pix Network to Estimate Agricultural Near Infrared Images from RGB Data |
| title_sort | pix2pix network to estimate agricultural near infrared images from rgb data |
| url | http://dx.doi.org/10.1080/07038992.2021.2016056 |
| work_keys_str_mv | AT danielcaiodelima pix2pixnetworktoestimateagriculturalnearinfraredimagesfromrgbdata AT diegosaqui pix2pixnetworktoestimateagriculturalnearinfraredimagesfromrgbdata AT steveatakytshammpinda pix2pixnetworktoestimateagriculturalnearinfraredimagesfromrgbdata AT josehirokisaito pix2pixnetworktoestimateagriculturalnearinfraredimagesfromrgbdata |