Classification of Maize in Complex Smallholder Farming Systems Using UAV Imagery

Abstract

Yield estimates and yield gap analysis are important for identifying poor agricultural productivity. Remote sensing holds great promise for measuring yield and thus determining yield gaps. Farming systems in sub-Saharan Africa (SSA) are commonly characterized by small field size, intercropping, different crop species with similar phenologies, and sometimes high cloud frequency during the growing season, all of which pose real challenges to remote sensing. Here, an unmanned aerial vehicle (UAV) system based on a quadcopter equipped with two consumer-grade cameras was used for the delineation and classification of maize plants on smallholder farms in Ghana. Object-oriented image classification methods were applied to the imagery, combined with measures of image texture and intensity, hue, and saturation (IHS), in order to achieve delineation. It was found that the inclusion of a near-infrared (NIR) channel and red-green-blue (RGB) spectra, in combination with texture or IHS, increased the classification accuracy for both single and mosaic images to above 94%. Thus, the system proved suitable for delineating and classifying maize using RGB and NIR imagery and calculating the vegetation fraction, an important parameter in producing yield estimates for heterogeneous smallholder farming systems.

Impact of Texture Information on Crop Classification with Machine Learning and UAV Images

“…This is a common issue in Korea, where various types of crops are cultivated in small areas. The use of high-resolution satellite images and aerial photos can contribute to resolving the mixed pixel issues [11,12]. Despite the increased discrimination capability of high-resolution images, it is difficult to collect time-series datasets over the full growth cycles of crops.…”

“…In recent years, there was a growing interest in imaging of unmanned aerial vehicles (UAV) [11][12][13][14][15]. The advantage of UAV images over satellite images is their ability to provide local thematic information with much higher spatial and temporal resolutions [15].…”

“…Despite the great potential of UAV imaging, the technique has several practical issues. Firstly, the ultra-high spatial resolution of UAV images usually causes noise effects due to increased detectable targets when conventional pixel-based approaches are applied for classification [11,18,19]. The common approach to mitigate noise effects is to either use spatial contextual information or apply an object-oriented classification approach.…”

“…These two approaches are known to achieve better classification accuracy than the pixel-based approach based purely on spectral information [19,22]. The second issue is heavy computational load related to data preprocessing and processing [11]. Most UAV images are acquired using a narrow field-of-view, which requires mosaicking of many sub-images to obtain a complete image set.…”

“…If the sub-images are taken at different solar conditions and flight altitudes, radiometric calibration should be employed during mosaicking. The ultra-high spatial resolution of UAV images makes preprocessing complex and requires much processing time for classification [11].…”