Bt Research 2025, Vol.16, No.6, 242-250 http://microbescipublisher.com/index.php/bt 248 7 Technical Challenges and Areas for Improvement 7.1 Influence of spatial and temporal resolution on Bt monitoring accuracy Sometimes, no matter how clear the image is, if the timing is not right, the key changes cannot be captured. The monitoring of Bt pesticide spraying is precisely such an application scenario that is quite picky about both "time" and "space". High spatial resolution can help us see which plots in the farmland have been sprayed and which have not. However, if the imaging frequency is insufficient, it is easy to miss the actual time point for spraying pesticides. Satellite data often faces such a dilemma - either to capture details or to capture frequently, and it is difficult to achieve both (Liu, 2025). Moreover, unexpected conditions such as clouds and weather may also interfere with imaging. In contrast, although drones have advantages in resolution and time flexibility, they cannot fly too far and their costs are not acceptable to all farmers. These restrictions, when combined, make it somewhat difficult to accurately locate the location and dosage of Bt application, ultimately affecting the reliability of the monitoring results (Sasakaros et al., 2025). 7.2 Challenges in multi-source data fusion and error control Data from different sources inherently have distinct "personalities", and it is naturally not an easy task to integrate them. Satellite shooting conditions, drone altitude, ground observation time points... Even a slight difference may lead to the data not matching when spliced together. Besides, each sensor has its own imaging characteristics and error range. If these are incorporated into a system without a unified standard, problems are likely to arise. Often, even with the use of fusion algorithms, it is difficult to completely eliminate noise and bias. Errors will accumulate layer by layer and eventually affect the judgment results of the Bt application mode. To control these errors, it is necessary to rely on ground measurement data for calibration and also require strong computing power support. The problem is that not every region has these resources (Kang et al., 2019). 7.3 User-side technology thresholds and adoption barriers in agriculture Advanced technologies are not necessarily all suitable for grassroots promotion, and this is particularly evident in the application of remote sensing and GIS monitoring of Bt crops. Many farmers have no idea where to start installing and maintaining the system, let alone how to process the data or interpret the results. Technical training and infrastructure construction are not well developed in many small-scale farmers or remote areas. Furthermore, even if someone is willing to give it a try, the cost of high-resolution images and drone flights is often the first "threshold". Therefore, although remote sensing and GIS are useful, their actual implementation at the field level still needs to take into account the acceptance of users. Unless there is a simpler and more user-friendly platform or the cost can be reduced, the popularization of such monitoring tools may still have a long way to go (Sasakaros et al., 2025). 8 Conclusion and Future Perspectives In the past, the monitoring of Bt biopesticides relied more on ground observations or empirical judgments. However, the introduction of remote sensing and GIS has enabled us to grasp the health of crops and the dynamics of pests and diseases from a large range, specific locations and time for the first time. This clear and rapid assessment method for space is obviously more efficient than traditional methods. However, technology is not omnipotent and has encountered some limitations in application. However, for now, integrating multi-source data (such as satellite images, drone flight records, ground sensor data, etc.) has indeed made the planting management and environmental assessment of Bt crops more scientific, and also made the implementation of precision agriculture more feasible. For Bt monitoring to go further, the intelligence and adaptive capabilities of the model must keep up. An ideal monitoring system may have to handle different types of data such as hyperspectral, lidar and time series images simultaneously, and also run smoothly in various complex environments. Of course, this places higher demands on modeling. Many studies have begun to attempt to combine machine learning with GIS for more dynamic identification of crop-pest interactions. Whether the platform can be made intuitive and user-friendly, and whether it can help farmers save trouble instead of causing trouble, also determines the speed of technology promotion.
RkJQdWJsaXNoZXIy MjQ4ODYzNA==