The new technologies are changing the agriculture more than ever. The goal of the precision agriculture is to integrate these new methods to optimise the decision-making process in farming.
The IMS Laboratory of the University of Bordeaux has a decade old experience in this field, obtained in several projects in the field of agricultural image processing and remote sensing.
To promote this experience, IMS Bordeaux created imAgro, a group specialized in computer vision in near-field remote sensing applications.
The IMS Laboratory of the University of Bordeaux has a decade old experience in this field, obtained in several projects in the field of agricultural image processing and remote sensing.
To promote this experience, IMS Bordeaux created imAgro, a group specialized in computer vision in near-field remote sensing applications.
Main projects :
Projet EARN
Early yield estimation using embedded camera
Based on over ten years of research at the IMS Laboratory, the EARN project aims to develop a new tool for early harvest estimation for vine parcels, with a fiability of 90%. This device, mounted on a tractor or a quad, is made of a camera, which takes a photo of each individual vine plant, and a small computer capable of analysing the photos and detect the brunches, count the visible berries and measure their volume. This process can be done during other agricultural works.
The early harvest estimation is an important task, as it allows the farmers to estimate during the whole season the volume of the harvest. Till now, it was performed by hand, a tiresome work which could lead to large estimation errors. The results of the EARN device can help to anticipate some tasks (like the green harhest), to optimize the agricultural works, organise the harvesting and anticipate the commercialization.
The EARN project was awarded with a bronze medal at Vinitech-Sifel 2016
The early harvest estimation is an important task, as it allows the farmers to estimate during the whole season the volume of the harvest. Till now, it was performed by hand, a tiresome work which could lead to large estimation errors. The results of the EARN device can help to anticipate some tasks (like the green harhest), to optimize the agricultural works, organise the harvesting and anticipate the commercialization.
The EARN project was awarded with a bronze medal at Vinitech-Sifel 2016
Pixfel
Early harvest quality estimation for apples and plums
The fruit quality estimation at harvest time is an essential tool in the commercial management of the batches. It allows to build an optimal marketing strategy for the stored fruits, to anticipate the commercial scheduling and increase the value of the harvest.
The appearance of low-cost vision systems was an opportunity to propose to the professionals a management system for quality prevision which is adapted to the orchards. This prevision can be done at harvest time, at the entrance to the storage facilities by photographing the surface layer of bulk shipping bins.
The project aims an automatic estimation of the harvest quality: the diameter distribution and the coloration of the apple skin; and in a future version will detect the eventual surface faults. This project had a CasDar founding during the 2005-2007 period.
The Pixfel device consists of a digital camera with a 5 megapixel or higher resolution, and a lens with a 42mm equivalent focal length mounted on a gantry; and a computer controlling the camera. The Pixfel software analyses the photos taken by this device when the bins pass under the camera. The results of the image processing - the diameter and color distribution of the fruits - is added to a database for further analysis.
Pixfel was awarded with a silver medal at Sitevi 2011
The appearance of low-cost vision systems was an opportunity to propose to the professionals a management system for quality prevision which is adapted to the orchards. This prevision can be done at harvest time, at the entrance to the storage facilities by photographing the surface layer of bulk shipping bins.
The project aims an automatic estimation of the harvest quality: the diameter distribution and the coloration of the apple skin; and in a future version will detect the eventual surface faults. This project had a CasDar founding during the 2005-2007 period.
The Pixfel device consists of a digital camera with a 5 megapixel or higher resolution, and a lens with a 42mm equivalent focal length mounted on a gantry; and a computer controlling the camera. The Pixfel software analyses the photos taken by this device when the bins pass under the camera. The results of the image processing - the diameter and color distribution of the fruits - is added to a database for further analysis.
Pixfel was awarded with a silver medal at Sitevi 2011
VVINNER
Embedded vision for an autonomous vine robot
VVINNER (Vineyard Vigilant & INNovative Ecological Rover) is project founded by CIP Eco-innovation initiative of the European Union. It aims to transform Vitirover into a multifunctional robot capable of gathering useful information for monitoring the ecosystem of the vineyard, for early detection of disease risk and pest infestations, for the measurement of meteorological data with high resolution and for early yield estimation.
In this project, Imagro developped the vision and sensor systems for Vitirover.
In this project, Imagro developped the vision and sensor systems for Vitirover.
MecaVision
Apple tree monitoring from blossoming to harvest
This project aims to measure the load of apple trees during the growth season, between blossoming and harvest. Mecavision follows the evolution and growth of the fruits, evaluates the efficiency of mechanical and/or chemical thinning and predict the harvest yield. The algorithms developed during this project can detect and count the blossoms, flowers and fruits on pictures taken with a compact digital camera.
Aventuria
Détection précoce de la tavelure utilisant des images hyperspectrales
This project, developed in collaboration with Irstea Montpellier and CTIFL Lanxade aims to detect the apple scab infection before the symptoms become visible.
This project is founded by CasDar.
This project is founded by CasDar.