Obstacle Detection and Distance Estimation for Autonomous Electric Vehicle Using Stereo Vision and DNN

Emani, Sarma; Soman, K P; Variyar, V. V. Sajith; Adarsh, S.

doi:10.1007/978-981-13-3393-4_65

Cited by 23 publications

(7 citation statements)

References 4 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Furthermore, various studies [16,17] have been conducted using deep learning networks to detect various objects by training the model based on the features of objects. One of those studies [18] identifies objects using the SSD network among deep learning networks and estimates the distance to the identified object using a stereo camera. However, as many different obstacle shapes cannot be learned in disaster sites, we used the object patterns and image features.…”

Section: Related Work 21 Computer Visionmentioning

confidence: 99%

Intelligent SLAM Algorithm Fusing Low-Cost Sensors at Risk of Building Collapses

Kim¹,

Ahn²

2023

Computers, Materials &Amp; Continua

View full text Add to dashboard Cite

When firefighters search inside a building that is at risk of collapse due to abandonment or disasters such as fire, they use old architectural drawings or a simple monitoring method involving a video device attached to a robot. However, using these methods, the disaster situation inside a building at risk of collapse is difficult to detect and identify. Therefore, we investigate the generation of digital maps for a disaster site to accurately analyze internal situations. In this study, a robot combined with a low-cost camera and twodimensional light detection and ranging (2D-lidar) traverses across a floor to estimate the location of obstacles while drawing an internal map of the building. We propose an algorithm that detects the floor and then determines the possibility of entry, tracks collapses, and detects obstacles by analyzing patterns on the floor. The robot's location is estimated, and a digital map is created based on Hector simultaneous localization and mapping (SLAM). Subsequently, the positions of obstacles are estimated based on the range values detected by 2D-lidar, and the position of the obstacles are identified on the map using the map update method in semantic SLAM. All equipment are implemented using low-specification devices, and the experiments are conducted using a low-cost robot that affords near-real-time performance. The experiments are conducted in various actual internal environments of buildings. In terms of obstacle detection performance, almost all obstacles are detected, and their positions identified on the map with a high accuracy of 89%.

show abstract

Section: Related Work 21 Computer Visionmentioning

confidence: 99%

Intelligent SLAM Algorithm Fusing Low-Cost Sensors at Risk of Building Collapses

Kim¹,

Ahn²

2023

Computers, Materials &Amp; Continua

View full text Add to dashboard Cite

show abstract

“…Además, se han desarrollado propuestas enfocadas en el reconocimiento y estimación de la distancia de objetos, basadas en visión artificial. Por ejemplo, en (Emani et al, 2019), proponen un sistema de visión estéreo, que parte del reconocimiento de los objetos a través de una arquitectura MobileNet, con la cual obtienen recuadros que encierran a los objetos de interés. Posteriormente, estiman la distancia de los objetos detectados, aplicando una función de triangulación que toma como datos el desplazamiento en el eje X de los recuadros generados en las imágenes tomadas por ambas cámaras, la distancia entre las cámaras y la longitud focal de estas.…”

Section: Antecedentesunclassified

Reconocimiento y estimación de posición relativa de objetos en entornos controlados

Luna-Taylor

Rosas

Torres

et al. 2022

ICBI

View full text Add to dashboard Cite

En este proyecto, se presenta un sistema de reconocimiento y clasificación de objetos que se encuentran en su entorno, para un robot de asistencia, así como la estimación de la posición relativa de estos con respecto al robot. Para el reconocimiento y clasificación de los objetos, se aplicaron técnicas de visión artificial basadas en herramientas de segmentación semántica, como son redes neuronales convolucionales. Para la estimación de la posición relativa de los objetos, una vez identificados, se implementó una técnica de visión estereoscópica. Los resultados de los experimentos muestran un 90.6% de precisión en el reconocimiento y clasificación, y un error medio de 5 cm al estimar la posición relativa de los objetos.

show abstract

“…[18] built a vision system based on two cameras and the results obtained using the VS2013 environment have shown that the error is less than 5% for distances within 2 m. Distance values found from two experiments, one using a webcam rig and the other one using ZED camera, were compared in the work of Ref. [19] and showed that they are almost identical. Lai et al.…”

Section: Introductionmentioning

confidence: 99%

“…Sun et al [18] built a vision system based on two cameras and the results obtained using the VS2013 environment have shown that the error is less than 5% for distances within 2 m. Distance values found from two experiments, one using a webcam rig and the other one using ZED camera, were compared in the work of Ref. [19] and showed that they are almost identical. Lai et al [20] proposed a distance measurement system composed of a Digital Signal Processor and two cameras; their experiments showed that a general PC system takes about 0.751 s to calculate the distance.…”

mentioning

confidence: 99%

A novel algorithm for distance measurement using stereo camera

Adil

Mikou

Mouhsen

2022

CAAI Trans on Intel Tech

View full text Add to dashboard Cite

Distance estimation can be achieved by using active sensors or with the help of passive sensors such as cameras. The stereo vision system is generally composed of two cameras to mimic the human binocular vision. In this paper, a Python‐based algorithm is proposed to find the parameters of each camera, rectify the images, create the disparity maps and finally use these maps for distance measurements. Experiments using real‐time images, which were captured from our stereo vision system, of different obstacles positioned at multiple distances (60–200 cm) prove the effectiveness of the proposed program and show that the calculated distance to the obstacle is relatively accurate. The accuracy of distance measurement is up to 99.83%. The processing time needed to calculate the distance between the obstacle and the cameras is less than 0.355 s.

show abstract

Obstacle Detection and Distance Estimation for Autonomous Electric Vehicle Using Stereo Vision and DNN

Cited by 23 publications

References 4 publications

Intelligent SLAM Algorithm Fusing Low-Cost Sensors at Risk of Building Collapses

Intelligent SLAM Algorithm Fusing Low-Cost Sensors at Risk of Building Collapses

Reconocimiento y estimación de posición relativa de objetos en entornos controlados

A novel algorithm for distance measurement using stereo camera

Contact Info

Product

Resources

About