Deep Learning for Soil and Crop Segmentation from Remotely Sensed Data

Jack Dyson,Primo Zingaretti,Emanuele Frontoni,Adriano Mancini

doi:10.3390/rs11161859

Abstract

One of the most challenging problems in precision agriculture is to correctly identify and separate crops from the soil. Current precision farming algorithms based on artificially intelligent networks use multi-spectral or hyper-spectral data to derive radiometric indices that guide the operational management of agricultural complexes. Deep learning applications using these big data require sensitive filtering of raw data to effectively drive their hidden layer neural network architectures. Threshold techniques based on the normalized difference vegetation index (NDVI) or other similar metrics are generally used to simplify the development and training of deep learning neural networks. They have the advantage of being natural transformations of hyper-spectral or multi-spectral images that filter the data stream into a neural network, while reducing training requirements and increasing system classification performance. In this paper, to calculate a detailed crop/soil segmentation based on high resolution Digital Surface Model (DSM) data, we propose the redefinition of the radiometric index using a directional mathematical filter. To further refine the analysis, we feed this new radiometric index image of about 3500 × 4500 pixels into a relatively small Convolution Neural Network (CNN) designed for general image pattern recognition at 28 × 28 pixels to evaluate and resolve the vegetation correctly. We show that the result of applying a DSM filter to the NDVI radiometric index before feeding it into a Convolutional Neural Network can potentially improve crop separation hit rate by 65%.

Highlights

In precision agriculture applications, the typical object of interest is a prescription map that will direct tractor operations like spraying nitrogen or weed treatment over the agricultural field
We develop an example using a simple supervised Convolution Neural Network (CNN) to test how effective the strategy presented in Section 2.2 can be when using a Digital Surface Model (DSM)/normalized difference vegetation index (NDVI) index
Summarizing, after carrying out the proposed benchmark testing, we find that the DSM/NDVI index produces an improvement of about four times compared to its baseline NDVI marker

Summary

Introduction

The typical object of interest is a prescription map that will direct tractor operations like spraying nitrogen or weed treatment over the agricultural field. The planning of agricultural tasks requires a deep knowledge of crop state [2]. Vineyards and fruit plants are especially good examples of complex exercises in both crop detection and study for agricultural image segmentation [4] problems such as weed detection [5], nitrogen application at hot-spots and selective harvesting. Segmentation of images from unmanned aerial vehicles (UAVs) is usually an important and essential step for tasks such as estimating the physiological status of plants [6], identifying plant or weed [5,7,8,9], fruit grading and picking [10,11], crop energy inputs and productivity analysis [12] and crop disease detection [13,14]

Objectives

Findings

Discussion

Conclusion