A method for agricultural machine guidance on row crops based on the vanishing point

• Main Authors: de Oliveira Neris, Luciano, Obac Roda, Valentin, Trindade, Onofre
• Format: Conference Proceeding
• Language: English
• Subjects: Agriculture; Control systems; Crops; Data mining; Fatigue; Feedback control; Image processing; Lighting control; Safety; Stability
• DOI: 10.1109/ICVES.2007.4456365

Automatic guidance systems have became important in agriculture to reduce the operator's work and fatigue, resulting in higher efficiency and safety in field operations. Most crops are cultivated in rows, hence enabling the development of a row-recognition system to accurately follow a row of plants without overrunning the crop. This paper presents a novel approach for extracting agricultural machine position from field images based on the look-ahead method and vanishing points. The proposed method facilitates the development of the guidance controller, simplifying the tuning and increasing the guidance system response and stability. The goal was to develop a quick position extraction method, based on the fact that by decreasing the image processing time increases the position data rate, providing fast response feedback control. Results demonstrated that the use of look- ahead method and vanishing point simplifies the guidance controller development and satisfactorily works with shadows, gaps, weeds and different light conditions.