Comparative Study on the Recognition Effect of RGB and Multispectral Images on Banana Leaves with Nitrogen Deficiency

It is of great significance for banana planting industry to monitor and evaluate the growth status of banana plants efficiently and non destructively.In sunny and cloudy days,the RGB and multispectral images of banana leaves were collected.The recognition effects of the two images were compared by using a variety of target detection algorithms,and the imaging differences of the two images under different light conditions were analyzed.The experimental results show that under different models,the recognition rate of multispectral image detection is 7%～14% higher than that of RGB image.The recognition effect of multispectral image is the best in cloudy and smooth environment,but the adverse light environment has a serious impact on its imaging.Using multispectral images can quickly and non destructively identify the nitrogen deficiency status of banana plants in the environment without obvious light effect,and provide technical support for the monitoring of banana planting status.