Strawberry localization in a ridge planting with an autonomous rover. (March 2023)

• Record Type: Journal Article
• Authors: Mejia, Gesem; Montes de Oca, Andrés; Flores, Gerardo
• Is Part Of: Engineering applications of artificial intelligence. Volume 119(2023)
• Publication Date: 2023-03
• Subjects: Strawberry detection -- Ripeness estimation -- Deep learning -- Mobile robot; Engineering -- Data processing -- Periodicals; Artificial intelligence -- Periodicals; Expert systems (Computer science) -- Periodicals; Ingénierie -- Informatique -- Périodiques; Intelligence artificielle -- Périodiques; Systèmes experts (Informatique) -- Périodiques; Artificial intelligence; Engineering -- Data processing; Expert systems (Computer science); Periodicals; 620.00285
• DOI: 10.1016/j.engappai.2022.105810
• Languages: English
• ISSNs: 0952-1976

Abstract

Abstract: Having complete knowledge of a strawberry crop field is helpful in making crucial decisions and optimizing strawberry production. For that aim, autonomous robots have played a crucial role in getting such information through various sensors in the last few years. In this context, we present an autonomous rover system that detects strawberries and estimates their ripeness in a natural crop field through visual information using a stereo camera. With that aim, we propose an algorithm that uses image processing techniques with deep learning approaches. Furthermore, we construct a strawberry map to provide the farmer with information regarding the growing state of the fruit, health, and even a production estimation. Moreover, the rover and the vision algorithm are designed to work on ridge planting, a kind of strawberry crop field where uneven terrains, narrow spaces, and rugged backgrounds are presented. Experiments were conducted in a natural crop field, showing satisfactory results; from comparison results with the most recent work in the state-of-the-art, we observe that our system detects 14% more strawberries. Furthermore, a video showing the obtained results is presented. Finally, intending to contribute to the community, we release the code corresponding to our proposed method on our GitHub website.