Sensors

Research

16 pages, 1927 KiB

Open AccessArticle

Enhancing the Tracking of Seedling Growth Using RGB-Depth Fusion and Deep Learning

by Hadhami Garbouge, Pejman Rasti and David Rousseau

Sensors 2021, 21(24), 8425; https://doi.org/10.3390/s21248425 - 17 Dec 2021

Cited by 10 | Viewed by 3506

The use of high-throughput phenotyping with imaging and machine learning to monitor seedling growth is a tough yet intriguing subject in plant research. This has been recently addressed with low-cost RGB imaging sensors and deep learning during day time. RGB-Depth imaging devices are [...] Read more.

The use of high-throughput phenotyping with imaging and machine learning to monitor seedling growth is a tough yet intriguing subject in plant research. This has been recently addressed with low-cost RGB imaging sensors and deep learning during day time. RGB-Depth imaging devices are also accessible at low-cost and this opens opportunities to extend the monitoring of seedling during days and nights. In this article, we investigate the added value to fuse RGB imaging with depth imaging for this task of seedling growth stage monitoring. We propose a deep learning architecture along with RGB-Depth fusion to categorize the three first stages of seedling growth. Results show an average performance improvement of

5 %

correct recognition rate by comparison with the sole use of RGB images during the day. The best performances are obtained with the early fusion of RGB and Depth. Also, Depth is shown to enable the detection of growth stage in the absence of the light. Full article

(This article belongs to the Special Issue Recent Advances in Depth Sensors and Applications)

► Show Figures

Figure 1

19 pages, 5677 KiB

Open AccessArticle

Metrological Characterization and Comparison of D415, D455, L515 RealSense Devices in the Close Range

by Michaela Servi, Elisa Mussi, Andrea Profili, Rocco Furferi, Yary Volpe, Lapo Governi and Francesco Buonamici

Sensors 2021, 21(22), 7770; https://doi.org/10.3390/s21227770 - 22 Nov 2021

Cited by 38 | Viewed by 5469

Abstract

RGB-D cameras are employed in several research fields and application scenarios. Choosing the most appropriate sensor has been made more difficult by the increasing offer of available products. Due to the novelty of RGB-D technologies, there was a lack of tools to measure [...] Read more.

RGB-D cameras are employed in several research fields and application scenarios. Choosing the most appropriate sensor has been made more difficult by the increasing offer of available products. Due to the novelty of RGB-D technologies, there was a lack of tools to measure and compare performances of this type of sensor from a metrological perspective. The recent ISO 10360-13:2021 represents the most advanced international standard regulating metrological characterization of coordinate measuring systems. Part 13, specifically, considers 3D optical sensors. This paper applies the methodology of ISO 10360-13 for the characterization and comparison of three RGB-D cameras produced by Intel^® RealSense™ (D415, D455, L515) in the close range (100–1500 mm). ISO 10360-13 procedures, which focus on metrological performances, are integrated with additional tests to evaluate systematic errors (acquisition of flat objects, 3D reconstruction of objects). The present paper proposes an off-the-shelf comparison which considers the performance of the sensors throughout their acquisition volume. Results have exposed the strengths and weaknesses of each device. The D415 device showed better reconstruction quality on tests strictly related to the short range. The L515 device performed better on systematic depth errors; finally, the D455 device achieved better results on tests related to the standard. Full article

(This article belongs to the Special Issue Recent Advances in Depth Sensors and Applications)

► Show Figures

Figure 1

18 pages, 10173 KiB

Open AccessArticle

AdjustSense: Adaptive 3D Sensing System with Adjustable Spatio-Temporal Resolution and Measurement Range Using High-Speed Omnidirectional Camera and Direct Drive Motor

by Mikihiro Ikura, Sarthak Pathak, Jun Younes Louhi Kasahara, Atsushi Yamashita and Hajime Asama

Sensors 2021, 21(21), 6975; https://doi.org/10.3390/s21216975 - 21 Oct 2021

Cited by 1 | Viewed by 2498

Abstract

Many types of 3D sensing devices are commercially available and were utilized in various technical fields. In most conventional systems with a 3D sensing device, the spatio-temporal resolution and the measurement range are constant during operation. Consequently, it is necessary to select an [...] Read more.

Many types of 3D sensing devices are commercially available and were utilized in various technical fields. In most conventional systems with a 3D sensing device, the spatio-temporal resolution and the measurement range are constant during operation. Consequently, it is necessary to select an appropriate sensing system according to the measurement task. Moreover, such conventional systems have difficulties dealing with several measurement targets simultaneously due to the aforementioned constants. This issue can hardly be solved by integrating several individual sensing systems into one. Here, we propose a single 3D sensing system that adaptively adjusts the spatio-temporal resolution and the measurement range to switch between multiple measurement tasks. We named the proposed adaptive 3D sensing system “AdjustSense.” In AdjustSense, as a means for the adaptive adjustment of the spatio-temporal resolution and measurement range, we aimed to achieve low-latency visual feedback for the adjustment by integrating not only a high-speed camera, which is a high-speed sensor, but also a direct drive motor, which is a high-speed actuator. This low-latency visual feedback can enable a large range of 3D sensing tasks simultaneously. We demonstrated the behavior of AdjustSense when the positions of the measured targets in the surroundings were changed. Furthermore, we quantitatively evaluated the spatio-temporal resolution and measurement range from the 3D points obtained. Through two experiments, we showed that AdjustSense could realize multiple measurement tasks: 360

^{\circ}

3D sensing, 3D sensing at a high spatial resolution around multiple targets, and local 3D sensing at a high spatio-temporal resolution around a single object. Full article

(This article belongs to the Special Issue Recent Advances in Depth Sensors and Applications)

► Show Figures

Figure 1

18 pages, 7968 KiB

Open AccessArticle

Real-Time FPGA Accelerated Stereo Matching for Temporal Statistical Pattern Projector Systems

by Zan Brus, Marko Kos, Matic Erker and Iztok Kramberger

Sensors 2021, 21(19), 6435; https://doi.org/10.3390/s21196435 - 26 Sep 2021

Cited by 1 | Viewed by 2706

Abstract

The presented paper describes a hardware-accelerated field programmable gate array (FPGA)–based solution capable of real-time stereo matching for temporal statistical pattern projector systems. Modern 3D measurement systems have seen an increased use of temporal statistical pattern projectors as their active illumination source. The [...] Read more.

The presented paper describes a hardware-accelerated field programmable gate array (FPGA)–based solution capable of real-time stereo matching for temporal statistical pattern projector systems. Modern 3D measurement systems have seen an increased use of temporal statistical pattern projectors as their active illumination source. The use of temporal statistical patterns in stereo vision systems includes the advantage of not requiring information about pattern characteristics, enabling a simplified projector design. Stereo-matching algorithms used in such systems rely on the locally unique temporal changes in brightness to establish a pixel correspondence between the stereo image pair. Finding the temporal correspondence between individual pixels in temporal image pairs is computationally expensive, requiring GPU-based solutions to achieve real-time calculation. By leveraging a high-level synthesis approach, matching cost simplification, and FPGA-specific design optimizations, an energy-efficient, high throughput stereo-matching solution was developed. The design is capable of calculating disparity images on a 1024 × 1024(@291 FPS) input image pair stream at 8.1 W on an embedded FPGA platform (ZC706). Several different design configurations were tested, evaluating device utilization, throughput, power consumption, and performance-per-watt. The average performance-per-watt of the FPGA solution was two times higher than in a GPU-based solution. Full article

(This article belongs to the Special Issue Recent Advances in Depth Sensors and Applications)

► Show Figures

Figure 1

20 pages, 3724 KiB

Open AccessArticle

Measurement Noise Model for Depth Camera-Based People Tracking

by Otto Korkalo and Tapio Takala

Sensors 2021, 21(13), 4488; https://doi.org/10.3390/s21134488 - 30 Jun 2021

Cited by 5 | Viewed by 3416

Abstract

Depth cameras are widely used in people tracking applications. They typically suffer from significant range measurement noise, which causes uncertainty in the detections made of the people. The data fusion, state estimation and data association tasks require that the measurement uncertainty is modelled, [...] Read more.

Depth cameras are widely used in people tracking applications. They typically suffer from significant range measurement noise, which causes uncertainty in the detections made of the people. The data fusion, state estimation and data association tasks require that the measurement uncertainty is modelled, especially in multi-sensor systems. Measurement noise models for different kinds of depth sensors have been proposed, however, the existing approaches require manual calibration procedures which can be impractical to conduct in real-life scenarios. In this paper, we present a new measurement noise model for depth camera-based people tracking. In our tracking solution, we utilise the so-called plan-view approach, where the 3D measurements are transformed to the floor plane, and the tracking problem is solved in 2D. We directly model the measurement noise in the plan-view domain, and the errors that originate from the imaging process and the geometric transformations of the 3D data are combined. We also present a method for directly defining the noise models from the observations. Together with our depth sensor network self-calibration routine, the approach allows fast and practical deployment of depth-based people tracking systems. Full article

(This article belongs to the Special Issue Recent Advances in Depth Sensors and Applications)

► Show Figures

Figure 1

25 pages, 6406 KiB

Open AccessArticle

3D Reconstruction of Non-Rigid Plants and Sensor Data Fusion for Agriculture Phenotyping

by Gustavo Scalabrini Sampaio, Leandro A. Silva and Maurício Marengoni

Sensors 2021, 21(12), 4115; https://doi.org/10.3390/s21124115 - 15 Jun 2021

Cited by 18 | Viewed by 4669

Abstract

Technology has been promoting a great transformation in farming. The introduction of robotics; the use of sensors in the field; and the advances in computer vision; allow new systems to be developed to assist processes, such as phenotyping, of crop’s life cycle monitoring. [...] Read more.

Technology has been promoting a great transformation in farming. The introduction of robotics; the use of sensors in the field; and the advances in computer vision; allow new systems to be developed to assist processes, such as phenotyping, of crop’s life cycle monitoring. This work presents, which we believe to be the first time, a system capable of generating 3D models of non-rigid corn plants, which can be used as a tool in the phenotyping process. The system is composed by two modules: an terrestrial acquisition module and a processing module. The terrestrial acquisition module is composed by a robot, equipped with an RGB-D camera and three sets of temperature, humidity, and luminosity sensors, that collects data in the field. The processing module conducts the non-rigid 3D plants reconstruction and merges the sensor data into these models. The work presented here also shows a novel technique for background removal in depth images, as well as efficient techniques for processing these images and the sensor data. Experiments have shown that from the models generated and the data collected, plant structural measurements can be performed accurately and the plant’s environment can be mapped, allowing the plant’s health to be evaluated and providing greater crop efficiency. Full article

(This article belongs to the Special Issue Recent Advances in Depth Sensors and Applications)

► Show Figures

Figure 1

20 pages, 7377 KiB

Open AccessArticle

Action Unit Detection by Learning the Deformation Coefficients of a 3D Morphable Model

by Luigi Ariano, Claudio Ferrari, Stefano Berretti and Alberto Del Bimbo

Sensors 2021, 21(2), 589; https://doi.org/10.3390/s21020589 - 15 Jan 2021

Cited by 9 | Viewed by 3211

Abstract

Facial Action Units (AUs) correspond to the deformation/contraction of individual facial muscles or their combinations. As such, each AU affects just a small portion of the face, with deformations that are asymmetric in many cases. Generating and analyzing AUs in 3D is particularly [...] Read more.

Facial Action Units (AUs) correspond to the deformation/contraction of individual facial muscles or their combinations. As such, each AU affects just a small portion of the face, with deformations that are asymmetric in many cases. Generating and analyzing AUs in 3D is particularly relevant for the potential applications it can enable. In this paper, we propose a solution for 3D AU detection and synthesis by developing on a newly defined 3D Morphable Model (3DMM) of the face. Differently from most of the 3DMMs existing in the literature, which mainly model global variations of the face and show limitations in adapting to local and asymmetric deformations, the proposed solution is specifically devised to cope with such difficult morphings. During a training phase, the deformation coefficients are learned that enable the 3DMM to deform to 3D target scans showing neutral and facial expression of the same individual, thus decoupling expression from identity deformations. Then, such deformation coefficients are used, on the one hand, to train an AU classifier, on the other, they can be applied to a 3D neutral scan to generate AU deformations in a subject-independent manner. The proposed approach for AU detection is validated on the Bosphorus dataset, reporting competitive results with respect to the state-of-the-art, even in a challenging cross-dataset setting. We further show the learned coefficients are general enough to synthesize realistic 3D face instances with AUs activation. Full article

(This article belongs to the Special Issue Recent Advances in Depth Sensors and Applications)

► Show Figures

Figure 1

13 pages, 1582 KiB

Open AccessEditor’s ChoiceArticle

Single-Shot 3D Shape Reconstruction Using Structured Light and Deep Convolutional Neural Networks

by Hieu Nguyen, Yuzeng Wang and Zhaoyang Wang

Sensors 2020, 20(13), 3718; https://doi.org/10.3390/s20133718 - 3 Jul 2020

Cited by 109 | Viewed by 10639

Abstract

Single-shot 3D imaging and shape reconstruction has seen a surge of interest due to the ever-increasing evolution in sensing technologies. In this paper, a robust single-shot 3D shape reconstruction technique integrating the structured light technique with the deep convolutional neural networks (CNNs) is [...] Read more.

Single-shot 3D imaging and shape reconstruction has seen a surge of interest due to the ever-increasing evolution in sensing technologies. In this paper, a robust single-shot 3D shape reconstruction technique integrating the structured light technique with the deep convolutional neural networks (CNNs) is proposed. The input of the technique is a single fringe-pattern image, and the output is the corresponding depth map for 3D shape reconstruction. The essential training and validation datasets with high-quality 3D ground-truth labels are prepared by using a multi-frequency fringe projection profilometry technique. Unlike the conventional 3D shape reconstruction methods which involve complex algorithms and intensive computation to determine phase distributions or pixel disparities as well as depth map, the proposed approach uses an end-to-end network architecture to directly carry out the transformation of a 2D image to its corresponding 3D depth map without extra processing. In the approach, three CNN-based models are adopted for comparison. Furthermore, an accurate structured-light-based 3D imaging dataset used in this paper is made publicly available. Experiments have been conducted to demonstrate the validity and robustness of the proposed technique. It is capable of satisfying various 3D shape reconstruction demands in scientific research and engineering applications. Full article

(This article belongs to the Special Issue Recent Advances in Depth Sensors and Applications)

► Show Figures

Figure 1

Journal Menu

Journal Browser

Recent Advances in Depth Sensors and Applications

Share This Special Issue

Special Issue Editors

Special Issue Information

Keywords

Benefits of Publishing in a Special Issue

Published Papers (8 papers)

Research

Further Information

Guidelines

MDPI Initiatives

Follow MDPI