关键词: 玉米播种, 出苗率, 无人驾驶, 多光谱影像, 无人机

Abstract: In order to verify the effect of driverless, assisted driving and manual driving on the emergence rate of corn seeding, fresh corn was taken as the experimental crop, and driverless, assisted driving and manual driving were conducted in different areas. Under the condition that the soil benefits and management methods are basically the same, UAV is used to collect multispectral images during seedling emergence. The method of combining supergreen value (ExG) and maximum interclass difference method is used to extract the characteristic image of maize seedlings in the farmland background, and the normalized vegetation index (NDVI) is calculated. Based on the measured and calculated emergence rate and regional NDVI, establish the corresponding relationship. Use spatial analysis technology to calculate the emergence rate under unmanned driving, assisted driving and manual driving modes. The results show that: Based on the UAV multispectral image calculation method, the seeding rate was 67.47%, 66.30% and 52.83% respectively by driverless, assisted and manual driving. The actual emergence rate of maize was 72.19%, 66.23% and 55.49%.The degree of difference was -4.72%, 0.07% and -2.66% respectively. The result of seedling emergence rate is: driverless area > auxiliary driving area > manual driving area. The emergence rate in driverless areas is the highest and relatively uniform. The calculation method based on UAV multispectral image is more accurate, which provides a theoretical basis for realtime estimation of maize emergence rate.

Key words: corn sowing, emergence rate, unmanned driving, multispectral image, UAV