USING PASSIVE AND ACTIVE MULTISPECTRAL SENSORS ON THE CORRELATION WITH THE PHENOLOGICAL INDICES OF COTTON

Abstract

ABSTRACT: The investment in precision agriculture technologies has been growing as well the investment in vegetation index sensors with different types of data collection. This study aimed to evaluate the NDVI potential use obtained from passive and active multispectral sensors as phenological parameters indicator of cotton growth. The experiment was conducted in cultivated cotton fields during the harvest seasons of 2013/14 and 2014/15. The phenological variables analyzed in the field, were cotton plant height and number of branches per sampled cotton plant. The active sensors Crop Circle and Green Seeker were mounted on a manual support. The passive sensor Tetracam was installed on a support positioning the sensor at 4 m height from the ground. NDVI readings are positively correlated with plant height and number of branches per cotton plant. Active sensors present exponential correlations between NDVI and phenological readings. Passive sensor, when installed in a support above the cotton canopy level, presents linear correlation between NDVI and phenological readings. NDVI values obtained by the passive sensor are higher than the active sensors. Filtering the raw NDVI values obtained by passive sensor decreases the adjustment of the regression analyses with cotton height and number of branches per plant.

Keywords

• vegetation index
• variability
• multispectral sensors