Advances in visual perception for agricultural robotics

Abstract

Visual perception is key to most agricultural robots in order to make sense of their surroundings, necessary to make decisions and apply actions. This chapter will start with an overview of image sensors, systems, and typical vision tasks in agricultural robotics. Despite great progress in the past decade due to the introduction of deep neural networks, robots face two big challenges in perceiving the agricultural environment: the challenges of variation and incomplete information. Both challenges and available solutions are discussed in this chapter. However, the challenges are not yet solved. The chapter therefore ends with an outline of future research to approach the challenges and to allow successful use of robots in agriculture.