Lars Lindner

Bio: Lars Lindner is an academic researcher from Autonomous University of Baja California. The author has contributed to research in topics: Machine vision & Laser scanning. The author has an hindex of 13, co-authored 52 publications receiving 488 citations.

Mobile robot vision system using continuous laser scanning for industrial application

Abstract: The purpose of this paper is the presentation and research of a novel robot vision system, which uses laser dynamic triangulation, to determine three-dimensional (3D) coordinates of an observed object. The previously used physical operation principle of discontinuous scanning method is substituted by continuous method. Thereby applications become possible that were previously limited by this discretization.,The previously used prototype No. 2, which uses stepping motors to realize a discontinuous laser scan, was substituted by the new developed prototype No. 3, which contains servomotors, to achieve a continuous laser scan. The new prototype possesses only half the width and turns out to be significantly smaller and therefore lighter than the old one. Furthermore, no transmissions are used, which reduce the systematic error of laser positioning and increase the system reliability.,By using a continuous laser scan method instead of discontinuous laser scan method, dead zones in the laser scanner field can be eliminated. Thereby, also by changing the physical operation principle, the implementation of applications is allowed, which previously was limited by the fixed step size or by the object distance under observation. By using servomotors instead of stepping motors, also a significant reduced positioning time can be accomplished maintaining the relative positioning error less than 1 per cent.,The originality is based on the substitution of the physical operation principle of discontinuous by continuous laser scan. The previously used stepping motors discretized the laser scanner field and thereby produced dead zones, where 3D coordinates cannot be detected. These stepping motors were substituted by servomotors to revoke these disadvantages and provide a continuous laser scan, where dead zones in the field of view get eliminated and the step response of the laser scanner accelerated.

Improve three-dimensional point localization accuracy in stereo vision systems using a novel camera calibration method:

Abstract: Computer vision systems have demonstrated to be useful in applications of autonomous navigation, especially with the use of stereo vision systems for the three-dimensional mapping of the environmen...

Influence of data clouds fusion from 3D real-time vision system on robotic group dead reckoning in unknown terrain

Abstract: This paper proposes the solution of tasks set required for autonomous robotic group behavior optimization during the mission on a distributed area in a cluttered hazardous terrain. The navigation scheme uses the benefits of the original real-time technical vision system ( TVS ) based on a dynamic triangulation principle. The method uses TVS output data with fuzzy logic rules processing for resolution stabilization. Based on previous researches, the dynamic communication network model is modified to implement the propagation of information with a feedback method for more stable data exchange inside the robotic group. According to the comparative analysis of approximation methods, in this paper authors are proposing to use two-steps post-processing path planning aiming to get a smooth and energy-saving trajectory. The article provides a wide range of studies and computational experiment results for different scenarios for evaluation of common cloud point influence on robotic motion planning.

Exact laser beam positioning for measurement of vegetation vitality

Abstract: Purpose The purpose of this paper is to present a novel application for a newly developed Technical Vision System (TVS), which uses a laser scanner and dynamic triangulation, to determine the vitality of agriculture vegetation. This vision system, installed on an unmanned aerial vehicle, shall measure the reflected laser energy and thereby determine the normalized differenced vegetation index. Design/methodology/approach The newly developed TVS shall be installed on the front part of the unmanned aerial vehicle, to perform line-by-line scan in the vision system field-of-view. The TVS uses high-quality DC motors, instead of previously researched low-quality DC motors, to eliminate the existence of two mutually exclusive conditions, for exact positioning of a DC motor shaft. The use of high-quality DC motors reduces the positioning error after control. Findings Present paper emphasizes the exact laser beam positioning in the field-of-view of a TVS. By use of high-quality instead of low-quality DC motors, a significant reduced positioning time was achieved, maintaining the relative angular position error less than 1 per cent. Best results were achieved, by realizing a quasi-continuous control, using a high pulse-width modulated duty cycle resolution and a high execution frequency of the positioning algorithm. Originality/value The originality of present paper is represented by the novel application of the newly developed TVS in the field of agriculture. The vitality of vegetation shall be determined by measuring the reflected laser energy of a scanned agriculture zone. The paper’s main focus is on the exact laser beam positioning within the TVS field-of-view, using high-quality DC motors in closed-loop position control configuration.

Experimental image and range scanner datasets fusion in SHM for displacement detection

Abstract: Summary Optical images and signals can be used to detect displacement in civil engineering structures. This paper presents a technical experimentation of a vision-based technology and artificial intelligence algorithms methodology for structural health monitoring of new and aging structures, by a noncontact and nondestructive system. The experimental study emphasis is on the outdoor urban environment, by the detection of spatial coordinate displacement on the structures, in order to perform a damage assessment. Also, the experimental study contains both theoretical and experimental aspects of the fusion of image and range scanner datasets created using intelligent algorithms. A camera and an optical scanning system were used to generate high resolution and quality images for 2D imaging, and 3D accuracy range data from optoelectronic sensor signals. Scans at a specific area of an engineering structure were performed to measure spatial coordinates displacements, successfully verifying the effectiveness and the robustness of the proposed non-contact and non-destructive monitoring approach.

Unmanned Aerial Vehicles in Agriculture: A Review of Perspective of Platform, Control, and Applications

Abstract: For agricultural applications, regularized smart-farming solutions are being considered, including the use of unmanned aerial vehicles (UAVs). The UAVs combine information and communication technologies, robots, artificial intelligence, big data, and the Internet of Things. The agricultural UAVs are highly capable, and their use has expanded across all areas of agriculture, including pesticide and fertilizer spraying, seed sowing, and growth assessment and mapping. Accordingly, the market for agricultural UAVs is expected to continue growing with the related technologies. In this study, we consider the latest trends and applications of leading technologies related to agricultural UAVs, control technologies, equipment, and development. We discuss the use of UAVs in real agricultural environments. Furthermore, the future development of the agricultural UAVs and their challenges are presented.

Bearing Fault Diagnosis Method Based on Deep Convolutional Neural Network and Random Forest Ensemble Learning.

Abstract: Recently, research on data-driven bearing fault diagnosis methods has attracted increasing attention due to the availability of massive condition monitoring data. However, most existing methods still have difficulties in learning representative features from the raw data. In addition, they assume that the feature distribution of training data in source domain is the same as that of testing data in target domain, which is invalid in many real-world bearing fault diagnosis problems. Since deep learning has the automatic feature extraction ability and ensemble learning can improve the accuracy and generalization performance of classifiers, this paper proposes a novel bearing fault diagnosis method based on deep convolutional neural network (CNN) and random forest (RF) ensemble learning. Firstly, time domain vibration signals are converted into two dimensional (2D) gray-scale images containing abundant fault information by continuous wavelet transform (CWT). Secondly, a CNN model based on LeNet-5 is built to automatically extract multi-level features that are sensitive to the detection of faults from the images. Finally, the multi-level features containing both local and global information are utilized to diagnose bearing faults by the ensemble of multiple RF classifiers. In particular, low-level features containing local characteristics and accurate details in the hidden layers are combined to improve the diagnostic performance. The effectiveness of the proposed method is validated by two sets of bearing data collected from reliance electric motor and rolling mill, respectively. The experimental results indicate that the proposed method achieves high accuracy in bearing fault diagnosis under complex operational conditions and is superior to traditional methods and standard deep learning methods.

Monocular Depth Estimation Based On Deep Learning: An Overview

Abstract: Depth information is important for autonomous systems to perceive environments and estimate their own state. Traditional depth estimation methods, like structure from motion and stereo vision matching, are built on feature correspondences of multiple viewpoints. Meanwhile, the predicted depth maps are sparse. Inferring depth information from a single image (monocular depth estimation) is an ill-posed problem. With the rapid development of deep neural networks, monocular depth estimation based on deep learning has been widely studied recently and achieved promising performance in accuracy. Meanwhile, dense depth maps are estimated from single images by deep neural networks in an end-to-end manner. In order to improve the accuracy of depth estimation, different kinds of network frameworks, loss functions and training strategies are proposed subsequently. Therefore, we survey the current monocular depth estimation methods based on deep learning in this review. Initially, we conclude several widely used datasets and evaluation indicators in deep learning-based depth estimation. Furthermore, we review some representative existing methods according to different training manners: supervised, unsupervised and semi-supervised. Finally, we discuss the challenges and provide some ideas for future researches in monocular depth estimation.

Mobile robot vision system using continuous laser scanning for industrial application

Abstract: The purpose of this paper is the presentation and research of a novel robot vision system, which uses laser dynamic triangulation, to determine three-dimensional (3D) coordinates of an observed object. The previously used physical operation principle of discontinuous scanning method is substituted by continuous method. Thereby applications become possible that were previously limited by this discretization.,The previously used prototype No. 2, which uses stepping motors to realize a discontinuous laser scan, was substituted by the new developed prototype No. 3, which contains servomotors, to achieve a continuous laser scan. The new prototype possesses only half the width and turns out to be significantly smaller and therefore lighter than the old one. Furthermore, no transmissions are used, which reduce the systematic error of laser positioning and increase the system reliability.,By using a continuous laser scan method instead of discontinuous laser scan method, dead zones in the laser scanner field can be eliminated. Thereby, also by changing the physical operation principle, the implementation of applications is allowed, which previously was limited by the fixed step size or by the object distance under observation. By using servomotors instead of stepping motors, also a significant reduced positioning time can be accomplished maintaining the relative positioning error less than 1 per cent.,The originality is based on the substitution of the physical operation principle of discontinuous by continuous laser scan. The previously used stepping motors discretized the laser scanner field and thereby produced dead zones, where 3D coordinates cannot be detected. These stepping motors were substituted by servomotors to revoke these disadvantages and provide a continuous laser scan, where dead zones in the field of view get eliminated and the step response of the laser scanner accelerated.