Marco Bietresato

Evaluation of a LiDAR-based 3D-stereoscopic vision system for crop-monitoring applications

Abstract When dealing with unmanned agricultural vehicles (remotely-controlled vehicles, robots), vision systems are a key-factor for implementing field-solutions having direct interactions with crops. Among all the possible information given by a vision system, the punctual estimation of the canopy volume is surely an interesting parameter: it is related to the crop vegetative status and, hence, it is fundamental for performing and setting-up properly some important field-operations (e.g., pruning/thinning, spraying). A system able to recognize the canopy volume can provide either the input-signals for implementing a robotic real-time site-specific farming system or relevant information for a proper crop management. However, there are many practical difficulties in the field implementation of such a system: complex canopy shapes, different colours, textures and illumination conditions with projected shadows. Terrestrial/aerial vision systems working on visible-light wavelengths and/or 2D-images of crops, although capable of excellent performances, have a computationally-heavy post-processing; therefore, they are unsuitable for implementing low-cost real-time servo-actuated cropping systems (e.g., robotised sprayers). Instead, a vision system composed by two LiDAR sensors aligned vertically, scanning the same targets, could give a sort of stereoscopic vision, here named “ lateral-linear-stereoscopic vision ”. The aim of this study is assessing the opportunity to use such a system on an automatic or autonomous/robotised implement by performing some preliminary tests in a controlled environment. The resulting system is independent of the lighting conditions (it works also in the dark), is highly reliable (no projected shadows) and data processing is very fast. Although further studies are required to overcome the issues that could arise in a future field implementation, this system has all the premises to be successfully embedded in an automatized monitoring system.

Development of a Dynamic Stability Simulator for Articulated and Conventionaltractors Useful for Real-Time Safety Devices

The safety of agricultural tractors drivers is a very actual topic, especially when tractors operate on side slopes, such as in terraced vineyards. This work approaches the stability problem of articulated tractors by modelling, simulating and quantifying the safety of the driver with respect to both roll and pitch overturns. First of all, an articulated tractor has been modelled and simplified, after that a stability index has been defined and calculatedin several simulated slope conditions when the tractor travels along a circular trajectory; then, the obtained results have beencompared with respect to a conventional tractor. This work is a preliminary studyfor a tilting test platform for real vehicles, capable to reproduce real field conditions (slope, obstacles, roughness). Finally, some directives on how exploiting the obtained results for real-time safety devices have been formulated.

Technical aspects concerning the detection of animal waste nutrient content via its electrical characteristics.

The variables influencing corrosion of three metals (galvanised steel, stainless steel, brass) usable for a manure nutrient probe were examined, identifying the best material for field applications. The nutrients in 18 liquid manures were then estimated through the voltage drop between the terminals of a prototype probe. Response Surface Modelling gave the regression functions relating each investigated response only to the statistically-significant factors. After 168h in the manure, it was determined that: stainless steel was the most suitable material for very close electrodes (mass: -1.8% at 15mm), brass can be used with any inter-electrode distance (mass: -13.0% maximum at 35mm). The prototype probe gave reliable estimates (R(2)⩾0.744) of Ntot, Namm, Ptot, Ktot when dry matter and temperature were also accounted for in the regression analysis. Not considering dry matter but just electronically-detectable quantities (temperature, voltage drop), the estimates were only reliable (R(2)⩾0.656) above 20°C.

Design and First Tests of a Vision System on a Tele-operated Vehicle for Monitoring the Canopy Vigour Status in Orchards

A vision system able to give a punctual estimation of the canopy vigour (volume, leaves’ chlorophyll content) of an orchard is a key-system for implementing Precision Agriculture. Indeed, such a system, composed by Lidar and NDVI sensors, can give all the information necessary for performing some important field-operations (e.g., pruning, spraying) and, above all, for setting-up automatically and in real-time the relative machines. The first issues when implementing a vision system concern: which type and how many sensors using, how making this system move within an orchard. As proved in some preliminary lab trials, the use of two Lidar sensors, vertically-aligned to give a sort of lateral-linear-stereoscopic vision, manages to avoid the presence of the large “projected shadows” (or “blind spots”) originating when using a single sensor to scan a target. Then, this article presents a compact “mobile lab”, based on an electric tracked bins-carrier, able to move off-road within the orchards and equipped with an ad-hoc developed adjustable tubular frame, designed to carry two Lidar sensors in the individuated configuration, together with other six (NDVI) sensors. This frame allows placing the sensors at different heights to ensure the complete scan of the canopy (even with high fruit trees).

A numerical approach for evaluating and properly setting self-propelled forage harvesters.

The choice and setup of self-propelled forage harvesters (FHs) in the cost-effective production of quality silage must consider both energetic and zootechnical issues. A global evaluation of whole-plant harvesting operations includes an assessment of the fundamental in-field performance of the machines (theoretical field time and fuel consumption) and the physical and nutritional traits of the corn silage (particle size, dimensional fractions, and level of kernel breakage). However, finding the optimal settings for a machine while taking into account the variability of the field is difficult and experimentally expensive, as there are many factors that can influence these parameters and whose action is not yet completely known. All of these parameters were investigated through an experimental plan including different adjustments of the theoretical cut length (TCL; 10 to 20 mm), processor roller clearance (PRC), and roller speed difference (13%, 25%, and 60%) of the corn conditioner device (CCD) on three different models of self-propelled forage harvesters (425 to 606 kW engine power) with two headers (6.0 and 7.5 m), used for harvesting whole-plant corn silage at different sites (crop yield of 37.12 to 66.96 t ha-1, and dry matter content of 35.93% to 44.52%). The collected data were then processed by performing several statistical analyses (ANOVA), and the coefficients of the regression models (response surface modeling, RSM) were calculated. From this analysis, it was evident which factors influenced the analyzed responses, and several numerical models, approximating the real behavior of the machines within the ranges of the tested factors, were generated. The aim of this work was to demonstrate the effectiveness of this approach in investigating the relationships among these variables to evaluate and properly set up the machines. The results revealed some general correlations among variables and responses, particularly with regard to the quality of the final product, and they provided some basic criteria for refining the settings of the cutterhead of the forage harvesters, particularly the TCL, and of the crop processor, specifically the clearance and speed difference of the rollers.

A tracked mobile robotic lab for monitoring the plants volume and health

Precision agriculture has been increasingly recognized for its potential ability to improve agricultural productivity, reduce production cost, and minimize damage to the environment. In this work, the current stage of our research in developing a mobile platform equipped with different sensors for orchard monitoring and sensing is presented. In particular, the mobile platform is conceived to monitor and assess both the geometric and volumetric conditions as well as the health state of the canopy. To do so, different sensors have been integrated and effective data-processing algorithms implemented for a reliable crop monitoring. Experimental tests have been performed allowing to obtain both a precise volume reconstruction of several plants and an NDVI mapping suitable for vegetation state evaluations.