Application Models For Monitoring Cotton Growth In South Xinjiang Based On Digital Image Technology

[Objective] The main objective of this study was to develop an image processing techniques for monitoring cotton growth and N status. There were significant exponential relationships between the canopy cover of color indices and the three growth variables(the N content, leaf area index(LAI), and dry matter accumulation(DMA)). The models were validated in three large and representative fields of high-yielding cotton of the first Crops in Xinjiang Production and Construction Corps.[Methods] The experiment design included two cultivars of cotton xinluzao-39(XLZ-39) and xinluzao-56(XLZ-56). In this study, the irrigation mode was drip irrigation under mulch. We selected three large and representative fields. The canopy cover was extracted from canopy images of the cotton with digital camera in 2012 and 2013. The exponential function models were calibrated with N content, leaf area index(LAI), and dry matter accumulation(DMA). And canopy cover(CC) were even more highly correlated with canopy N content, LAI, and DMA in ten color indices.[Results] By the 1:1 lines and simulated values and observed values of the color index CC, the RMSE values were 184.981 g·m-2and the R2 values were 0.892 for DMA, the RMSE values were 0.802 m2·m-2 and the R2 values were 0.917 for LAI, the RMSE values were 1.9041 g·m-2and the R2 values were 0.949 for N content. These data indicate that the color indice CC can be used to accurately estimate the growth and nutrition status of the cotton crop.[Conclusion] The digital image processing techniques in this study of the cotton canopy is reliable, simple, rapid, cost effective, and applicable to different types of green plants. The color index canopy cover extracted by image processing techniques was closely related to cotton growth and nutrition status. The CC extracted from canopy images of cotton was exponentially related to cotton N content, LAI, and DMA. So canopy cover to monitor cotton growth and nutrition status has to be done.Overall, the results of my study suggest that image processing technology has potential to increase yield and improve fertilizer use efficiency in cotton production,thereby reducing environmental pollution. Future work is needed to improve the methods of image processing and image analysis in order to utilize the full potential of this approach.