scholarly journals Intelligent System for Estimation of the Spatial Position of Apples Based on YOLOv3 and Real Sense Depth Camera D415

Symmetry ◽  
2022 ◽  
Vol 14 (1) ◽  
pp. 148
Author(s):  
Nikita Andriyanov ◽  
Ilshat Khasanshin ◽  
Daniil Utkin ◽  
Timur Gataullin ◽  
Stefan Ignar ◽  
...  
Keyword(s):  
Coordinate System ◽  
Intelligent System ◽  
Complete Information ◽  
Depth Camera ◽  
Network Architectures ◽  
Linear Transformations ◽  
Aggregate Information ◽  
Bounding Boxes ◽  
Object Relative ◽  
Detection And Recognition

Despite the great possibilities of modern neural network architectures concerning the problems of object detection and recognition, the output of such models is the local (pixel) coordinates of objects bounding boxes in the image and their predicted classes. However, in several practical tasks, it is necessary to obtain more complete information about the object from the image. In particular, for robotic apple picking, it is necessary to clearly understand where and how much to move the grabber. To determine the real position of the apple relative to the source of image registration, it is proposed to use the Intel Real Sense depth camera and aggregate information from its depth and brightness channels. The apples detection is carried out using the YOLOv3 architecture; then, based on the distance to the object and its localization in the image, the relative distances are calculated for all coordinates. In this case, to determine the coordinates of apples, a transition to a symmetric coordinate system takes place by means of simple linear transformations. Estimating the position in a symmetric coordinate system allows estimating not only the magnitude of the shift but also the location of the object relative to the camera. The proposed approach makes it possible to obtain position estimates with high accuracy. The approximate root mean square error is 7–12 mm, depending on the range and axis. As for precision and recall metrics, the first is 100% and the second is 90%.

scholarly journals Calibration of the Relative Orientation between Multiple Depth Cameras Based on a Three-Dimensional Target

Sensors ◽  
2019 ◽  
Vol 19 (13) ◽  
pp. 3008 ◽  
Author(s):  
Zhe Liu ◽  
Zhaozong Meng ◽  
Nan Gao ◽  
Zonghua Zhang
Keyword(s):  
Coordinate System ◽  
Camera Calibration ◽  
Three Dimensional ◽  
Field Of View ◽  
Depth Information ◽  
Depth Camera ◽  
Depth Cameras ◽  
Calibration Methods ◽  
3D Shape Reconstruction ◽  
Single Coordinate

Depth cameras play a vital role in three-dimensional (3D) shape reconstruction, machine vision, augmented/virtual reality and other visual information-related fields. However, a single depth camera cannot obtain complete information about an object by itself due to the limitation of the camera’s field of view. Multiple depth cameras can solve this problem by acquiring depth information from different viewpoints. In order to do so, they need to be calibrated to be able to accurately obtain the complete 3D information. However, traditional chessboard-based planar targets are not well suited for calibrating the relative orientations between multiple depth cameras, because the coordinates of different depth cameras need to be unified into a single coordinate system, and the multiple camera systems with a specific angle have a very small overlapping field of view. In this paper, we propose a 3D target-based multiple depth camera calibration method. Each plane of the 3D target is used to calibrate an independent depth camera. All planes of the 3D target are unified into a single coordinate system, which means the feature points on the calibration plane are also in one unified coordinate system. Using this 3D target, multiple depth cameras can be calibrated simultaneously. In this paper, a method of precise calibration using lidar is proposed. This method is not only applicable to the 3D target designed for the purposes of this paper, but it can also be applied to all 3D calibration objects consisting of planar chessboards. This method can significantly reduce the calibration error compared with traditional camera calibration methods. In addition, in order to reduce the influence of the infrared transmitter of the depth camera and improve its calibration accuracy, the calibration process of the depth camera is optimized. A series of calibration experiments were carried out, and the experimental results demonstrated the reliability and effectiveness of the proposed method.

scholarly journals Implementation of intelligent model for pneumonia detection

2019 ◽  
Vol 13 (4) ◽  
pp. 315-322
Author(s):  
Željko Knok ◽  
Klaudio Pap ◽  
Marko Hrnčić
Keyword(s):  
Neural Networks ◽  
Intelligent System ◽  
Input Parameter ◽  
Unintended Consequences ◽  
Accurate Diagnosis ◽  
Medical Professionals ◽  
Network Architectures ◽  
X Ray ◽  
Learning Principles ◽  
Intelligent Model

The advancement of technology in the field of artificial intelligence and neural networks allows us to improve speed and efficiency in the diagnosis of various types of problems. In the last few years, the rise in the field of convolutional neural networks has been particularly noticeable, showing promising results in problems related to image processing and computer vision. Given that humans have limited ability to detect patterns in individual images, accurate diagnosis can be a problem for even medical professionals. In order to minimize the number of errors and unintended consequences, computer programs based on neural networks and deep learning principles are increasingly used as assistant tools in medicine. The aim of this study was to develop a model of an intelligent system that receives x-ray image of the lungs as an input parameter and, based on the processed image, returns the possibility of pneumonia as an output. The implementation of this functionality was implemented through transfer learning methodology based on already defined convolution neural network architectures.

scholarly journals Deep Learning-Based Pose Estimation of Apples for Inspection in Logistic Centers Using Single-Perspective Imaging

Processes ◽  
2019 ◽  
Vol 7 (7) ◽  
pp. 424 ◽  
Author(s):  
Lino Antoni Giefer ◽  
Juan Daniel Arango Castellanos ◽  
Mohammad Mohammadzadeh Babr ◽  
Michael Freitag
Keyword(s):  
Neural Networks ◽  
Convolutional Neural Networks ◽  
Pose Estimation ◽  
Three Dimensional ◽  
Estimation Algorithm ◽  
Single Shot ◽  
Network Architectures ◽  
Image Area ◽  
Natural Variations ◽  
Bounding Boxes

Fruit packaging is a time-consuming task due to its low automation level. The gentle handling required by some kinds of fruits and their natural variations complicates the implementation of automated quality controls and tray positioning for final packaging. In this article, we propose a method for the automatic localization and pose estimation of apples captured by a Red-Green-Blue (RGB) camera using convolutional neural networks. Our pose estimation algorithm uses a cascaded structure composed of two independent convolutional neural networks: one for the localization of apples within the images and a second for the estimation of the three-dimensional rotation of the localized and cropped image area containing an apple. We used a single shot multi-box detector to find the bounding boxes of the apples in the images. Lie algebra is used for the regression of the rotation, which represents an innovation in this kind of application. We compare the performances of four different network architectures and show that this kind of representation is more suitable than using state-of-the-art quaternions. By using this method, we achieved a promising accuracy for the rotation regression of 98.36%, considering an error range lower than 15 degrees, forming a base for the automation of fruit packing systems.

scholarly journals Review on Deep Learning and Neural Network Implementation for Emotions Recognition

2021 ◽  
Vol 1 (1) ◽  
Author(s):  
Renas Rajab Asaad
Keyword(s):  
Neural Network ◽  
Intelligent System ◽  
Electronic Devices ◽  
Network Architectures ◽  
Digital Information ◽  
Final Model ◽  
Social And Emotional ◽  
Video Application ◽  
Emotional Relationships ◽  
Classification Tasks

     Recently, we humans integrate into the world of our smart phones and our portable electronic devices to the point that that world can numb us in one way or another and separate us from the real world, so that we and the generations that come after us are fully adapted to dealing with huge amounts of digital information, so they are ready to be absorbed faster And ready to deal with it more efficiently than previous generations, but during this process of rapid adaptation and adapting to the digital age gradually begins to lose one thing, this thing is what the machine has not given it yet and it is human emotions. A key step in the humanization of robotics is the ability to classify the emotion of the human operator. In this paper we present the design of an artificially intelligent system capable of emotion recognition trough facial expressions. Three Promising neural network architectures are customized, trained. and subjected to various classification tasks, after which the best performing network is further optimized. The applicability of the final model is portrayed in a live video application that can instantaneously return the emotion of the user. Technology experts have found that we can create empathy through technology as well, which will consequently lead to what is known as “emotional intelligence.” Instead of seeking about digital communication that is losing us to real communication, experts have found that we can employ technology in favor of that type of communication to restore Soul for social and emotional relationships that technology has lost its advantages for many years.

An Efficient General Purpose Contact Search Algorithm Using the Relative Coordinate System for Multibody System Dynamics

2007 ◽  
Vol 120 ◽  
pp. 129-134 ◽  
Author(s):  
Hui Je Cho ◽  
Dae Sung Bae ◽  
Jin Hwan Choi ◽  
Ja Choon Koo
Keyword(s):  
Coordinate System ◽  
Reference Frame ◽  
Contact Force ◽  
Search Algorithm ◽  
General Purpose ◽  
Contact Pair ◽  
Force Model ◽  
Relative Coordinate ◽  
Contact Search ◽  
Bounding Boxes

Dynamic analysis of many mechanical systems is often involved with contacts among bodies. This paper presents an efficient and general-purpose contact search algorithm for multibody dynamics in the context of the compliance contact force model. While many conventional collision detection algorithms are based on the absolute coordinate system, this paper proposes to use the relative coordinate system in detecting a contact. A boundary box of a defense surface geometry is divided into many blocks. A contact reference frame is defined on the defense body of a contact pair. Since all geometric variables necessary to detect a contact are measured relative to the contact reference frame attached to the defense body, the variables belonging to the defense body are constant, which significantly reduces computation time associated with the contact search. Therefore, the contact reference frame plays a key role in developing an efficient contact search algorithm. Contour of the defense body is approximated by many piecewise triangular patches, while contour of the hitting body is represented by hitting nodes along its boundary. Bounding boxes inside which contain each body of a contact pair are defined at a preprocessing stage to eliminate an exhaustive contact inspection when two bodies are in a distance. If two bounding boxes are turned out to be in a contact during the pre-search, each node on the hitting boundary is inspected to find out to which block the node belongs in the post-search. Since each block dividing the boundary of the defense body has a list of patches, each node on the hitting boundary is tested for a contact only with the patches in the block that the node belongs. Actual contact calculation is then carried out to find the contact penetration used in calculating the compliant contact force. Numerical example is performed to demonstrate the validity of the proposed method.

Reference Coordinate System Requirements for Geophysics

1975 ◽  
Vol 26 ◽  
pp. 87-92
Author(s):  
P. L. Bender
Keyword(s):  
Coordinate System ◽  
Polar Motion ◽  
Main Part ◽  
Measurement Techniques ◽  
Reference Points ◽  
Angular Motion ◽  
Coordinate Systems ◽  
Axis Of Rotation ◽  
The Earth ◽  
Fixed System

AbstractFive important geodynamical quantities which are closely linked are: 1) motions of points on the Earth’s surface; 2)polar motion; 3) changes in UT1-UTC; 4) nutation; and 5) motion of the geocenter. For each of these we expect to achieve measurements in the near future which have an accuracy of 1 to 3 cm or 0.3 to 1 milliarcsec.From a metrological point of view, one can say simply: “Measure each quantity against whichever coordinate system you can make the most accurate measurements with respect to”. I believe that this statement should serve as a guiding principle for the recommendations of the colloquium. However, it also is important that the coordinate systems help to provide a clear separation between the different phenomena of interest, and correspond closely to the conceptual definitions in terms of which geophysicists think about the phenomena.In any discussion of angular motion in space, both a “body-fixed” system and a “space-fixed” system are used. Some relevant types of coordinate systems, reference directions, or reference points which have been considered are: 1) celestial systems based on optical star catalogs, distant galaxies, radio source catalogs, or the Moon and inner planets; 2) the Earth’s axis of rotation, which defines a line through the Earth as well as a celestial reference direction; 3) the geocenter; and 4) “quasi-Earth-fixed” coordinate systems.When a geophysicists discusses UT1 and polar motion, he usually is thinking of the angular motion of the main part of the mantle with respect to an inertial frame and to the direction of the spin axis. Since the velocities of relative motion in most of the mantle are expectd to be extremely small, even if “substantial” deep convection is occurring, the conceptual “quasi-Earth-fixed” reference frame seems well defined. Methods for realizing a close approximation to this frame fortunately exist. Hopefully, this colloquium will recommend procedures for establishing and maintaining such a system for use in geodynamics. Motion of points on the Earth’s surface and of the geocenter can be measured against such a system with the full accuracy of the new techniques.The situation with respect to celestial reference frames is different. The various measurement techniques give changes in the orientation of the Earth, relative to different systems, so that we would like to know the relative motions of the systems in order to compare the results. However, there does not appear to be a need for defining any new system. Subjective figures of merit for the various system dependon both the accuracy with which measurements can be made against them and the degree to which they can be related to inertial systems.The main coordinate system requirement related to the 5 geodynamic quantities discussed in this talk is thus for the establishment and maintenance of a “quasi-Earth-fixed” coordinate system which closely approximates the motion of the main part of the mantle. Changes in the orientation of this system with respect to the various celestial systems can be determined by both the new and the conventional techniques, provided that some knowledge of changes in the local vertical is available. Changes in the axis of rotation and in the geocenter with respect to this system also can be obtained, as well as measurements of nutation.

2.2. Working Group 2: Conceptual Definitions of Reference Coordinate Systems for Earth Dynamics

1975 ◽  
Vol 26 ◽  
pp. 21-26
Keyword(s):  
Coordinate System ◽  
Working Group ◽  
General Character ◽  
Coordinate Systems ◽  
Reference Coordinate System ◽  
Group 2 ◽  
Definition Of ◽  
Earth Dynamics

An ideal definition of a reference coordinate system should meet the following general requirements:1. It should be as conceptually simple as possible, so its philosophy is well understood by the users.2. It should imply as few physical assumptions as possible. Wherever they are necessary, such assumptions should be of a very general character and, in particular, they should not be dependent upon astronomical and geophysical detailed theories.3. It should suggest a materialization that is dynamically stable and is accessible to observations with the required accuracy.

Sign in / Sign up

Export Citation Format

Share Document