Machine learning algorithms are increasingly used to enhance agricultural productivity cost-effectively. A critical task in precision agriculture is locating a plant’s root collar. This is required for the site-specific fertilization of the plants. Though state-of-the-art machine learning models achieve stellar performance in object detection, they are often sensitive to noisy inputs and variation in environment settings. In this paper, we propose an innovative technique of smooth perturbations to improve the robustness of root collar detection tasks using the YOLOv5 neural network model. We train a YOLOv5 model on blueberry image data for root collar detection. A small amount noise is added as a smooth perturbation to the bounding box of dimensions 50× 50, and this perturbed image is fed for training. Furthermore, we introduce an additional test set that represents the out-of-distribution (O.O.D.) case by applying Gaussian blur on test images to simulate particle situation. We use three different image datasets to train our model, the (i) Estonian blueberry, (ii) Serbian blueberry image, and (iii) public dataset sourced from Roboflow datasets, of sample size 118, 2779, and 2993, respectively. We achieve an overall precision of 0.886 on perturbed blueberry images compared to 0.871 on original (unperturbed) images for the O.O.D. test set. Similarly, our smooth perturbation training has achieved an mAP of 0.828, which significantly increases against the result of normal training, which only reaches 0.794. The result proves that our proposed smooth perturbation is an effective method to increase the robustness and generalizability of the object detection task.

For precise fertilization of blueberry plants, it is technologically the easiest and most suitable option to use a volumetric filling, for which it can be presumed that it is possible to precisely dose the fertilizer for each plant by grams. For setting up a volumetric filler, it is necessary to know parameters such as the size of the fertilizer particles and their bulk density. The aim of this research is to determine the granulometric parameters and their effect, which is done by measuring up three different fertilizers (SQM Qrop K, Memon Siforga, Substral): width, height, and length of 100 randomly selected fertilizer particles as well as the volumes and weights of 100 particles in 10 repetitions. According to the measurements, the average diameters of fertilizer particles were found as well as the average mass, volumes, and bulk density. A Mahr Digital Caliper 16EWRi 0–150 mm was used to measure the diameters of the fertilizer granules. A Yxlon FF35 computer tomograph was used to accurately scan particles. The analytical scale, Kern ABJ 220-4NM, was used to determine mass. The volumes were measured, using measuring glasses, with one having a maximum volume of 10 mL in 0.2 mL increments and another having a maximum volume of 100 mL in 1 mL increments. Descriptive statistics analysis was performed in Microsoft Excel. It turned out that the average diameters (3.68 vs. 3.64 vs. 4.29 mm) and bulk densities (0.928 vs. 0.631 vs. 0.824 g cm⁻³) of the three fertilizers differed far from each other, meaning that the given volume could be filled with different amounts of fertilizer. Equations between mass and weight were formed according to the measurements. As a result, it was found that a volumetric filler can be used for fertilizing blueberry plants precisely, but it demands adjusting the filler each time in the situation, which is defined by the variety of blueberry plants: their age, size, and health.

An integrated lignin-focused biorefining approach.

This article presents the development of a new pneumatic device for the precise application of mineral fertilizers, designed for use in precision agriculture systems involving farming robots. The proposed device is mounted on an autonomous agricultural platform and utilizes a machine vision system to determine plant coordinates. Its operating principle is based on accumulating a single dose of fertilizer in a chamber and delivering it precisely to the plant’s root zone using a directed airflow. The study includes a theoretical investigation of fertilizer movement inside the applicator tube under the influence of airflow and rotational motion of the tube. A mathematical model has been developed to describe both the relative and translational motion of the fertilizer. The equations, which account for frictional forces, inertia, and air pressure, enable the determination of optimal structural and kinematic parameters of the device depending on operating conditions and the properties of the applied material. The use of numerical methods to solve the developed mathematical model allows for synchronization of the device’s operating time parameters with the movement of the agricultural robot along the crop rows. The obtained results and the developed device improve the accuracy and speed of fertilizer application, minimize fertilizer consumption, and reduce soil impact, making the proposed device a promising solution for precision agriculture.