Investigation sugar beet density using remote sensing

One of the major branches of Precision Agriculture (P.A) is Remote Sensing (R.S). This study was in Eghlid Township to investigate the plant population of sugar beet and through this research the capability of LISS-III sensor images taken from satellite IRS-1D, P6 was investigated. 1D image was taken in 86/5/17 and simultaneously 51 data from 4 farms were taken systematically random for regression relationship, and investigation of models accuracy.. The farm “A” had more coverage than other farms for being early planted, therefore for first stage investigation was performed without the “A” farm and the next stage all farms were included. The second image was taken from satellite P6 which was close to sugar been harvesting time 86/7/13 and simultaneously 46 data from 2 farms were taken systematically random for regression relationship, and investigation of models accuracy. Images with less than one pixel accuracy were geo-referenced. For investigation of the relationship between measured plant population and satellite images, liner regression was applied. 1D image in the first stage (B, C, D) had the highest R square (R2=0.74) related to IPVI, NDVI indexes and in the second stage (all the farms), IPVI, NDVI, NRR had the highest R square (R2=0.43). For P6 image the highest R square was 0.27 related to NDVI index. For investigation of models accuracy we used the RMS method. It is concluded that when all the farms are not covered, the images of LISS-III satellite enable us to estimate the sugar beet population. But this needs highly correct image taking period. If the period exceeds 10 days form correct time, it will cause the plant population by satellite images be underestimated.