Search

• RGB & CHM fusion for precise canopy data. • Variable fusion thresholds impact segmentation. • Stand-scale forest mapping with Mask R-CNN & instance merging. The tree canopy represents a fundamental element of tree-related information. However, achieving precise canopy information from remote sensing images remains a significant challenge due to varying canopy sizes, mutual overlap, and diverse woodland environments. This study aims to leverage high-resolution Chinese fir images captured by an unmanned aerial vehicle (UAV) from a state forest farm in Jiangle County, Fujian Province, China. The images are integrated with the Mask R-CNN model to autonomously extract attributes at both individual and stand levels, facilitating precise forest mapping at the level of individual trees. The fusion entails a corresponding band saturation-weighted approach between RGB images and thermally-enhanced Canopy Height Model (CHM) images. The fusion threshold is set equal to the weight assigned to the CHM, and the weight of the RGB is computed as 1 minus the fusion threshold, ranging from 0 to 1 in intervals of 0.1. The dataset is then trained using three instance segmentation models: feature extraction networks based on ResNet50, ResNet101, and ResNeXt101, respectively. An instance merging approach based on the canopy cross-occupancy ratio is introduced, to enhance the accuracy of individual tree and stand-level attributes extraction, as well as forest mapping for individual tree canopies by using extensive stand images. This study focuses on evaluating the performance of instance segmentation models using two critical metrics: Bounding Box Average Precision (Box-AP) and Segmentation Average Precision (Segm-AP). The results highlighted that the Mask R-CNN model integrated with ResNeXt101, coupled with sample fusion using a threshold of 0.1, demonstrated exceptional performance. The accuracy of segmentation is acceptable, with Box-AP at 51.697 % and Segm-AP at 54.946 %. The extractions of individual level attributes, crown area, north–south crown width, and east–west crown width yielded R2 of 0.933, 0.871, and 0.877, respectively. As for stand-level attributes, canopy density, and individual population extraction resulted in R2 values of 0.901 and 0.912, respectively. Relative to the original images (with a fusion threshold of 0), segmentation accuracy was improved for all combinations, with the optimal configuration showing a 0.019 increase in R2 for canopy density and a 0.014 increase in R2 for individual population extraction. Furthermore, Box-AP and Segm-AP exhibit enhancements of 10.795 % and 10.746 %, respectively. The instances-merging method enhances the extraction accuracy of individual population by 5%. This study underscores the precision and efficacy of the Mask R-CNN instance-based segmentation model and fusion strategy, providing a robust support for the integration of deep learning in canopy extraction research. Its implications are of great significance for large-scale forestry surveys and the advancement of precision forestry, highlighting the substantial potential. [ABSTRACT FROM AUTHOR]

• Proposing a solution for tree height estimation based on trunk curve. • Estimating the trunk volume combining segmentation measurement and cylinder fitting. • Achieving automatic measurement of the individual tree parameters using CRP data. The forest plots survey's primary measurement factors are the diameter at breast height (DBH), tree height, and the individual tree volume. The reconstruction of a 3D forest scene based on dense images can contribute to the automatic extraction of individual tree parameters. Close-range photogrammetry (CRP) can achieve the reconstruction of a 3D scene from the sequence images at a highly efficient and low cost. However, the reconstructed point clouds often miss the trees' upper part, resulting in the tree height and volume not being measured directly. This paper proposes an integrating extraction framework and methods of the DBH, tree height, and individual tree volume for achieving automatic measurement of the individual tree parameters with high accuracy based on CRP. Four plantation plots in Wangyedian Forest Farm, Inner Mongolia, and Gaofeng Forest Farm, Guangxi Zhuang, are selected to obtain sequence close-up photographs and construct 3D scenes. The DBH, tree height, and trunk volume of the individual tree are extracted using circle fitting, stem curve fitting, growth equation, and segmentation measurement. The extracted individual tree parameters from Terrestrial laser scanning (TLS) are used as validation data to evaluate the accuracy. It shows a good performance in two Larix gmelinii plots, one Pinus tabulaeformis plot, and one Eucalyptus grandis plot with the accuracy of extracted DBH above 90%, which is 90.18%, 96.54%, 90.17%, and 95.00%, respectively. The accuracy of extracted tree height of the four plots is all above 85%, which is 93.34%, 91.71%, 90.32%, and 85.96%, respectively, and the accuracy of trunk volume is 91.55%, 91.94%, 89.86%, and 89.6%, respectively. The results show that our proposed integrated parameter extraction frame and methods can achieve high accuracy economically, efficiently, and conveniently, and be used for automatic measurements of individual tree factors in forest plot surveys. [ABSTRACT FROM AUTHOR]

• Method overview of deep learning application in machine vision. • Review of deep learning applications in fruit detection and yield estimation. • Recommendations made for original contributions to the literature in this field. • Recommendation made for the use of common public image sets. • Recommendation made for the use of common evaluation metrics. A review of developments in the rapidly developing field of deep learning is presented. Recommendations are made for original contributions to the literature, as opposed to formulaic applications of established methods to new application areas (e.g., to new crops), including the use of standard metrics (e.g., F1 score, the harmonic mean between Precision and Recall) for model comparison involving binary classification. A recommendation for the provision and use of publically available fruit-in-orchard image sets is made, to allow method comparisons and for implementation of transfer learning for deep learning models trained on the large public generic datasets. Emphasis is placed on practical aspects for application of deep learning models for the task of fruit detection and localisation, in support of tree crop load estimation. Approaches to the extrapolation of tree image counts to orchard yield estimation are also reviewed, dealing with the issue of occluded fruit in imaging. The review is intended to assist new users of deep learning image processing techniques, and to influence the direction of the coming body of application work on fruit detection. [ABSTRACT FROM AUTHOR]

• Image colors vary dramatically at different BBCH flowering phenological stages. • A random forest model classifies BBCH stages by image color ranges accurately. • Growing degree days and biological growth time improve the model accuracy. • In situ flowering phenology monitoring is applicable for peach breeding management. Accurate knowledge of peach flowering phenology is essential for scheduling precise irrigation and managing artificial pollination for breeding. However, in situ monitoring of peach flowering phenology is laborious and challenging. This study divided peach flowering phenology into eight stages covering inflorescence emergence and flowering using BBCH scales. Web-connected recording cameras and weather stations automatically-collected flower images and meteorological data from three peach cultivars. Daily variables, including area proportion of specific color ranges in the images, temperature, illuminance and heat/chilling requirement data were applied for training and testing four machine learning models: random forest (RF), support vector machine (SVM), naïve Bayes (NB), and k-Nearest Neighbors (KNN). Grid tuning and 10-fold cross-validations were conducted on these models to determine the optimal method for identifying peach flowering phenological stages. The RF model obtained highest optimal F1 score of 98.82 % (harmonic average of model accuracy and recall) among the models on the testing set. The area proportion of specific color ranges in the real-time images are essential for model performance. Heat requirement data (biological growth time and growing degree days) can improve the overall accuracy. The classification accuracy at different BBCH phenological stages of the three cultivars is acceptable for real-time monitoring of peach flowering phenology. These are applicable for peach breeding, heat stress management, and irrigation scheduling. The RF model also shows acceptable accuracy with flower images and public meteorological data input, which reveals the extendibility for phenology monitoring of other trees or crops. [ABSTRACT FROM AUTHOR]

• The article presents a trunk detector for tree crops based solely on RGB-D data. • The regular tree distribution is incorporated to a structure-based SLAM system. • The detector parameters have geometric meaning unlike black box-type procedures. • The proposal can run online and is experimentally tested in a challenging tree crop. Agricultural environments with tree plantations usually present a regular structure that can be used by SLAM systems to improve self-location, and therefore, facilitate the autonomous navigation. In this context, tree trunks are natural landmarks that can be used to incorporate the environment structure into the problem modeling. This article presents a trunk detector solely based on RGB-D data obtained from a frontal-view stereo camera, and a SLAM system that incorporates the regular tree distribution of these crops. The trunk detector can be adapted to similar agricultural environments because its parameters have specific geometric meanings, which differentiates it from black box-type procedures. The structure-based SLAM system has theoretical and practical advantages over the well-known SLAM procedures in the mentioned context. This proposal can be executed on-line and is experimentally tested with databases obtained from a challenging agricultural environment. Results show a good performance and robustness when the database is spatially or temporally subsampled, even under adverse lighting conditions. [ABSTRACT FROM AUTHOR]

• A smart sensing system for tree crop sprayers was developed and evaluated. • It utilizes sensor fusion (e.g., LiDAR, machine vision, GPS) and artificial intelligence. • The low-cost prototype was evaluated in citrus orchards. • It was able to detect trees and spray based on tree size and leaf density. • It reduced spraying volume by 28% compared to conventional spraying. Delivering the appropriate amount of chemicals for each plant based on their needs is essential for reducing costs, waste, and labor. Smart agricultural machinery can provide variable-rate technologies using artificial intelligence (AI) and machine vision to optimize spraying applications. In this research, a low-cost smart sensing system for controlling airblast tree crop sprayers was designed and evaluated using citrus as a case study. The prototype comprised a LiDAR, machine vision, GPS, flow meters, sensor fusion, and AI to scan trees for tree height, tree classification, and fruit counting. Specifically, this smart sensing system can detect and classify objects to tree or non-tree (e.g., human, field constructions), measure tree height and canopy density, and detect and count fruit. Based on this information, it controls spraying nozzles to optimize spraying applications. A novel software was written in C++ and ran on an Nvidia Jetson Xavier NX embedded computer to process and control the data utilizing data fusion and AI techniques. The smart sensing system's results for tree height estimate indicated a relatively low average error of 6%. A convolutional neural network (CNN) was used to perform tree classification with an average accuracy of 84% in classifying the collected imagery into mature, young, dead, and non-tree objects. The fruit count module (also a CNN) had an F1 score of 89%, compared to ground-truth labeled images of mature and immature citrus fruits. Finally, the adoption of this new sensing method reduced spraying volume by 28%, compared to traditional spraying applications. [ABSTRACT FROM AUTHOR]

Abstract: The structural characteristics of the canopy are a key consideration for improving the efficiency of the spray application process for tree crops. However, obtaining accurate data in an easy, practical, and efficient way is an important problem to be solved. This paper describes the technical characteristics of a sprayer prototype developed for vineyards, following the principles and previous laboratory tests described in the complementary paper Variable rate sprayer. Part 1 – Orchard prototype: design, implementation and validation. This prototype can modify the sprayed volume application rate according to the target geometry by using an algorithm based on the canopy volume inspired by the tree row volume (TRV) model. Variations in canopy width along the row crop are electronically measured using several ultrasonic sensors placed on the sprayer and used to modify the emitted flow rate from the nozzles in real time; the objective during this process is to maintain the sprayed volume per unit canopy volume (Lm−3). Field trials carried out at different crop stages for Merlot and Cabernet Sauvignon vines (Vitis vinifera) indicated a good relationship between the applied volume and canopy characteristics. The potential pesticide savings were estimated to be 21.9% relative to the costs of a conventional application. This conclusion is in accordance with the results of similar research on automated spraying systems. [Copyright &y& Elsevier]

Abstract: Judicious irrigation necessitates development of technologies that apply the precise amount of water demanded by crops. The goal is to schedule irrigation according to the particular requirements of each orchard and to the variability encountered during the season caused by weather conditions, canopy development and other factors. For supporting a precise and low labour management of irrigation here we propose and depict an algorithm that coordinates seven automatable tasks: (1) estimation of irrigation needs, (2) adaptation to a particular irrigation setup, (3) execution of the schedule, (4) soil and/or plant monitoring, (5) interpretation of sensor data, (6) reaction to occasional events and (7) tuning the model of irrigation needs. This is illustrated with three examples from a peach orchard where a software tool implementing the algorithm was configured to manage irrigation under different strategies. They included two alternative methods for estimating the crop water needs: a typical water balance based on reference evapotranspiration (ETo) and a variation based on the measurement of solar radiation intercepted by the canopy. They also included two types of sensors for feedback: soil water probes and dendrometers, the latter for applying regulated deficit irrigation. The results show that the proposed algorithm provided a common framework that could be embraced under different irrigation strategies to support autonomous control through the season, requiring no more data than those acquired and processed by the same system and enabling some degree of tolerance to failures. Overall, some advantages were the fast response to changes in factors such as weather conditions whose effect is rather predictable. At the same time the algorithm self-adjusted through feedback to variation in orchard-specific factors related with crop development and phenology which would be hard to model. The separation between agro-meteorological and hydraulic concerns in different steps of the algorithm facilitated the adaptation to the particularities of each irrigation setup, simplifying the use of commercial equipment from different manufacturers. [Copyright &y& Elsevier]

Abstract: This paper presents the foundations and applications in agriculture of the main systems used for the geometrical characterization of tree plantations, including systems based on ultrasound, digital photographic techniques, light sensors, high-resolution radar images, high-resolution X-ray computed tomography, stereo vision and LIDAR sensors. Amongst these, LIDAR laser scanners and stereo vision systems are probably the most promising and complementary techniques for achieving 3D pictures and maps of plants and canopies. The information about the geometric properties of plants provided by these techniques has innumerable applications in agriculture. Some important agricultural tasks that can benefit from these plant-geometry characterization techniques are the application of pesticides, irrigation, fertilization and crop training. In the field of pesticide application, knowledge of the geometrical characteristics of plantations will permit a better adjustment of the dose of the product applied, improving the environmental and economic impact. However, it is still necessary to resolve several technological and commercial questions. The former include improving detection systems, especially with regard to developing software for the post-processing steps and improving the speed of calculation and decision making. Amongst the latter, it is essential to produce low cost sensors and control systems in order to facilitate large-scale deployment. Obtaining a precise geometrical characterization of a crop at any point during its production cycle by means of a new generation of affordable and easy-to-use detection systems, such as LIDAR and stereo vision systems, will help to establish precise estimations of crop water needs as well as valuable information that can be used to quantify its nutritional requirements. If accurate, this can provide valuable information on which to base more sustainable irrigation and fertilizer dosages. These would be able to meet crop needs and could also be used as part of specific management systems, based on prescription maps, for the application of variable quantities of water and fertilizers. The availability of measurement tools that allow a precise geometric characterization of plantations will also facilitate and enhance research aimed at developing better crop training systems that ensure an optimal distribution of light within the treetops and higher fruit quality. It is therefore of vital importance to continue devoting major efforts to the development of increasingly accurate, robust and affordable systems capable of measuring the geometric characteristics of plantations, which support the development of the different areas of a sustainable and precision agriculture. [Copyright &y& Elsevier]

• Three models for the linking of plots receptors of rice straw with straw suppliers plots. • Direct application of straw to a single plot from several plots without prior storage. • Direct application in several receiving parcels from various suppliers without storage. • Model that groups provider plots with different collection points. This work presents three models for the linking of fruit tree plots which are receptors of rice straw that is applied as mulch, with other straw suppliers plots, whose straw is a byproduct of cultivating cereal. The mulch is applied to fruit tree plots in order to save irrigation water, to reduce the incidence of weeds and to reduce erosion. In other words, they intend to assign a set of supplier plots to each straw receiving plot. Each model solves the problem in one scenario. The first one considers the direct application of straw to a single plot which must be supplied to apply the mulch from several plots. Therefore, these must be selected from a set that act as suppliers. It is not considered prior storage or collection. The second scenario also involves a direct application but in several receiving target plots from several source supplier plots without storage or intermediate storage. The third scenario develops a model that groups provider plots with different collection points; It offers a system that selects the location of the storage points and associates each storage point with a group of receiving plots. [ABSTRACT FROM AUTHOR]