Abstract
Diversified studies on smart agriculture adopting various approaches have been conducted. Many of these targets large-scale agriculture. However, only a few studies have focused on small-scale agriculture. In large-scale agriculture, single-function agricultural robots are highly demanded, whereas in small-scale agriculture, agricultural robots with multiple functions are required. This study aims at developing an agricultural robot with multiple functions. Moreover, the study focuses on the functions of fruit harvesting and truss pruning and proposes a method for detecting the cutting points to achieve these functions. This method comprises three processes. First, the input image was classified into five classes (fruit, flower, calyx, truss, and other including leaf and background etc.) using semantic segmentation via deep learning. Post-processing was performed based on strawberry characteristics to remove false recognitions from the semantic segmentation results. Next, the properties of the fruit, calyx, and truss were determined. Finally, based on these properties, the cutting points for fruit harvesting and truss pruning were identified. This method detects cutting points based on plant features (size, shape, and position) of strawberries. Additionally, the unworkable area of the agricultural robot was estimated, and the possibility of cutting was determined by considering the unworkable area. The proposed method was evaluated using 1 000 images acquired from a strawberry greenhouse. The F-measure for harvesting and pruning were 0.93 and 0.86 in the cutting point detection and 0.96 and 0.88 in the mature and immature fruits detection, respectively. This study discusses the practical issues for implementing the proposed method in the agricultural robot.
Abstract
Diversified studies on smart agriculture adopting various approaches have been conducted. Many of these targets large-scale agriculture. However, only a few studies have focused on small-scale agriculture. In large-scale agriculture, single-function agricultural robots are highly demanded, whereas in small-scale agriculture, agricultural robots with multiple functions are required. This study aims at developing an agricultural robot with multiple functions. Moreover, the study focuses on the functions of fruit harvesting and truss pruning and proposes a method for detecting the cutting points to achieve these functions. This method comprises three processes. First, the input image was classified into five classes (fruit, flower, calyx, truss, and other including leaf and background etc.) using semantic segmentation via deep learning. Post-processing was performed based on strawberry characteristics to remove false recognitions from the semantic segmentation results. Next, the properties of the fruit, calyx, and truss were determined. Finally, based on these properties, the cutting points for fruit harvesting and truss pruning were identified. This method detects cutting points based on plant features (size, shape, and position) of strawberries. Additionally, the unworkable area of the agricultural robot was estimated, and the possibility of cutting was determined by considering the unworkable area. The proposed method was evaluated using 1 000 images acquired from a strawberry greenhouse. The F-measure for harvesting and pruning were 0.93 and 0.86 in the cutting point detection and 0.96 and 0.88 in the mature and immature fruits detection, respectively. This study discusses the practical issues for implementing the proposed method in the agricultural robot. Leer más
Related posts:
A Novel Approach Extracting Extreme Points Combining Multidirection Phase Superposition and Weighted Moment Diagram for Matching Multi-Modal Images
A passion fruit counting method based on the lightweight YOLOv5s and improved DeepSORT
An autonomous navigation method for orchard rows based on a combination of an improved a-star algorithm and SVR
Proposal of classes for umbu commercialization
What would it look like? Visualizing a future US Corn Belt landscape with more table food production
Ecosystem carbon stock and socio-ecological determinants in selected agroforestry practices at the eastern escarpment of Chercher Massive, Ethiopia