Multisensory System for Fruit Harvesting Robots. Experimental Testing in Natural Scenarios and with Different Kinds of Crops

Abstract

The motivation of this research was to explore the feasibility of detecting and locating fruits from different kinds of crops in natural scenarios. To this end, a unique, modular and easily adaptable multisensory system and a set of associated pre-processing algorithms are proposed. The offered multisensory rig combines a high resolution colour camera and a multispectral system for the detection of fruits, as well as for the discrimination of the different elements of the plants, and a Time-Of-Flight (TOF) camera that provides fast acquisition of distances enabling the localisation of the targets in the coordinate space. A controlled lighting system completes the set-up, increasing its flexibility for being used in different working conditions. The pre-processing algorithms designed for the proposed multisensory system include a pixel-based classification algorithm that labels areas of interest that belong to fruits and a registration algorithm that combines the results of the aforementioned classification algorithm with the data provided by the TOF camera for the 3D reconstruction of the desired regions. Several experimental tests have been carried out in outdoors conditions in order to validate the capabilities of the proposed system.

Modern viticulture in southern Europe: Vulnerabilities and strategies for adaptation to water scarcity

“…Therefore, tourism can be an extra trigger for farme and winery managers to implement best management practices in the vineyard and winery, and ultimately contribute to minimize the environmental impact of the wine industry. Birch et al (2011), Costantini and Barbetti (2008), Eibach and Töpfer (2015), Fernández et al (2014), Fusi et al (2014), Lopes et al (2004), Ruggieri et al (2009), andTorres-Ruiz et al (2014).…”

Multisensory System for the Detection and Localization of Peripheral Subcutaneous Veins

“…RANSAC samples the solution space of $\left(R,T\right)$ and estimates its fitness by counting the number of inliers, $f_0$:$f_0\left(F_1,F_2,R,T\right)={\displaystyle \sum }_i^{N}L\left(X_1^i,X_2^i,R,T\right)$ where:$L\left(X_1^i,X_2^i,R,T\right)=\{\begin{array}{cc}1,& e=\Vert R{X}_1^i+T-X_2^i\Vert <ϵ\\ 0,& otherwise\end{array}$ and $ϵ$ is the threshold beneath which a features match $\left(X_1^i,X_2^i\right)$ is determined to be an inlier. RANSAC chooses the transform with the largest number of inlier matches [38,39]. The resulting $\left(R,T\right)$ is then utilized online for matching the range data provided by the TOF camera with the processed SWIR data that contains the resulting peripheral veins detection.…”

VIS-NIR, SWIR and LWIR Imagery for Estimation of Ground Bearing Capacity

“…Meanwhile, in the near-infrared wavelengths (700 nm–1300 nm) photosynthesising plants reflect large proportions of the incident sunlight [ 40 , 41 ]. Therefore, processing of these images acquired with the aforementioned filters will allow us to focus only on those pixels of the scene that belong to the ground and that are required for further analysis [ 42 ]. Table 2 summarises technical specifications of the selected filters.…”