Bt Research 2025, Vol.16, No.6, 242-250 http://microbescipublisher.com/index.php/bt 244 3.2 Image processing and crop classification techniques Whether remote sensing images can be used or not still depends on the processing method. For instance, when images are brought back, they first need to undergo radiation correction, and then they might also need to be "merged" with other images. Only in this way can the data quality be improved. It was only then that the classification algorithm came into play. Many methods determine the species or growth stage of different crops based on their spectral performance differences, while others track changes through time series. There are also many classification methods, such as supervised classification and unsupervised classification, each with its own usage. When encountering complex plots, the combined use of different data sources, such as combining hyperspectral data with liDAR or thermal imaging data, can sometimes significantly improve the accuracy of recognition (Figure 1) (Omia et al., 2023; Allu and Mesapam, 2025). However, not every crop is suitable for the same processing procedure. In practical application, specific problems still need to be analyzed specifically. Figure 1 The electromagnetic spectrum, different wavelengths and regions, bands’ energy levels, and some examples of their use in agricultural remote sensing applications (Adopted from Omia et al., 2023) 3.3 Vegetation health and pest detection indicators (e.g., NDVI, EVI) based on remote sensing When evaluating the growth of crops, you may have heard of vegetation indices such as NDVI and EVI to some extent. These indices are not calculated out of thin air. They rely on the reflection of crops to different wavelengths of light, especially the differences between near-infrared and visible light. Simply put, the greener the leaves are, the higher the index is usually. Conversely, once the index shows abnormal fluctuations, it may be a sign of pest or disease, even earlier than the yellowing of leaves as seen by the human eye. This ability to issue early warnings is of great value to agricultural management. Furthermore, if these indices can be combined with time series data, the changing trend of the entire field can be tracked, such as when pests and diseases start to spread and to what stage they have developed, facilitating timely measures to be taken, such as precise application of Bt biopesticides (Segarra et al., 2020). Of course, all of this is predicated on the fact that the remote sensing data is detailed enough and someone knows how to use it.
RkJQdWJsaXNoZXIy MjQ4ODYzNA==