Accurate Identification of Seed Maize Fields Based on Histogram of Stripe Slopes

Abstract

In China, the process of emasculation of seed maize usually occurs in August. To prevent self-pollination by removing the tassels from every few rows of maize and to ensure the production of high-quality hybrid maize seeds, resulting in the formation of distinctive stripe-like textural features that show up in high-resolution satellite images. These features can be used as distinctive features to differentiate seed maize fields from common maize fields. In this study, Beijing-3A1 satellite image data with a resolution of 0.5 m were used to identify seed maize fields in Zhangye City, Gansu Province. First, the extraction of maize plots in remote sensing images is performed using a modified U2-net with a new field boundary loss function. Second, a new feature named “histogram of stripe slopes (HoSS)” was developed for seed maize field classification. We compare the classification accuracy obtained using HoSS features with different classifiers using other conventional features. The results show that HoSS features exhibit superiority for both single-feature classification and feature set classification. The feature set including entropy and HoSS with the K-nearest neighbor classification model as the method chosen in this study achieved 93.7% accuracy in identifying seed maize fields.

Keywords

• $_2$ -net%22%2C%22default_operator%22%3A%22AND%22%7D%7D%7D"> Improved U<named-content xmlns:xlink="http://www.w3.org/1999/xlink" xmlns:ali="http://www.niso.org/schemas/ali/1.0/" xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" content-type="math" xlink:type="simple"> <inline-formula> <tex-math notation="LaTeX">$_2$</tex-math> </inline-formula> </named-content>-net
• parcel segmentation
• seed maize
• seed maize field identification