Multisensor Fusion of Ground-based and Airborne Remote Sensing Data for Crop Condition Assessment

In this study, the performances of the optical sensors and instruments carried on both ground-based and airborne platforms were evaluated for monitoring crop growing status, detecting the vegetation response to aerial applied herbicides, and identifying crop nitrogen status. Geostatistical analysis on remotely sensed data was conducted to investigate spatial structure of crop canopy normalized difference vegetation index and multispectral imagery.

A computerized crop monitoring system was developed that combined sensors and instruments that measured crop structure and spectral data with a global positioning system. The integrated crop monitoring system was able to collect real-time, multi-source, multi-form, and crop related data simultaneously as the tractor-mounted system moved through the field.

This study firstly used remotely sensed data to evaluate glyphosate efficacy on weeds applied with conventional and emerging aerial spray nozzles. A weedy field was In this study, the performances of the optical sensors and instruments carried on both ground-based and airborne platforms were evaluated for monitoring crop growing status, detecting the vegetation response to aerial applied herbicides, and identifying crop nitrogen status. Geostatistical analysis on remotely sensed data was conducted to investigate spatial structure of crop canopy normalized difference vegetation index and multispectral imagery. A computerized crop monitoring system was developed that combined sensors and instruments that measured crop structure and spectral data with a global positioning system. The integrated crop monitoring system was able to collect real-time, multi-source, multi-form, and crop related data simultaneously as the tractor-mounted system moved through the field.

This study firstly used remotely sensed data to evaluate glyphosate efficacy on weeds applied with conventional and emerging aerial spray nozzles. A weedy field was set up in three blocks and four aerial spray technology treatments were tested. Spectral reflectance measurements were taken using ground-based sensors from all the plots at 1, 8, and 17 days after treatment. The results indicated that the differences among the treatments could be detected with spectral data. This study could provide applicators with guidance equipment configurations that can result in herbicide savings and optimized applications in other crops.

The main focus of this research was to apply sensor fusion technology to ground-based and airborne imagery data. Experimental plots cropped with cotton and soybean plants were set up with different nitrogen application rates. The multispectral imagery was acquired by an airborne imaging system over crop field; at the same period, leaf chlorophyll content and spectral reflectance measurements were gathered with chlorophyll meter and spectroradiometer at canopy level on the ground, respectively. Statistical analyses were applied on the data from individual sensor for discrimination with respect to the nitrogen treatment levels. Multisensor data fusion was performed at data level. The results showed that the data fusion of airborne imagery with ground-based data were capable of improving the performance of remote sensing data on detection of crop nitrogen status. The method may be extended to other types of data, and data fusion can be performed at feature or decision level.