Home /Research /Research on the Location of Citrus Picking Point Based on Structured Light Camera
OTHER

Research on the Location of Citrus Picking Point Based on Structured Light Camera

Xiaomei Hu, Bowen Ni

Year
2019
Citations
9

Abstract

Citrus picking robots have not been commercialized on a global scale. Based on the particularity of citrus fruit and fruit stem growth, a method of citrus picking point location based on structured light camera is proposed. Firstly, the citrus contours are extracted by color space conversion, threshold segmentation, watershed algorithm, opening and closing operation and ellipse fitting. After recognizing the citrus, the sticking citrus is corroded and combined so that the connected citrus region is fitted to a large ellipse. Then, the picking target region and the picking pixel are obtained by extracting the affine rectangle, the circumscribed circle and other regions. Finally, the depth information is extracted by the structured light camera to obtain the three-dimensional picking coordinate point. The experimental results show that the method can effectively recognize mature citrus under natural conditions, and the accuracy of picking points is more than 80%.

Keywords

RectangleEllipseComputer visionArtificial intelligenceComputer scienceAffine transformationPoint (geometry)SegmentationPixelClosing (real estate)

Related papers

Browse all OTHER papers