Beyond Bounding Boxes: Using Bounding Shapes for Real-Time 3D Vehicle Detection from Monocular RGB Images

2020 ◽

Vol 39 (3) ◽

pp. 2693-2710 ◽

Cited By ~ 2

Author(s):

Wael Farag

Keyword(s):

Real Time ◽

Vehicle Detection ◽

Machine Learning Algorithms ◽

Tracking Technique ◽

Detection And Tracking ◽

Vehicle Detection And Tracking ◽

Driving Assistance ◽

Rgb Images ◽

Assistance Systems ◽

Self Driving Cars

In this paper, an advanced-and-reliable vehicle detection-and-tracking technique is proposed and implemented. The Real-Time Vehicle Detection-and-Tracking (RT_VDT) technique is well suited for Advanced Driving Assistance Systems (ADAS) applications or Self-Driving Cars (SDC). The RT_VDT is mainly a pipeline of reliable computer vision and machine learning algorithms that augment each other and take in raw RGB images to produce the required boundary boxes of the vehicles that appear in the front driving space of the car. The main contribution of this paper is the careful fusion of the employed algorithms where some of them work in parallel to strengthen each other in order to produce a precise and sophisticated real-time output. In addition, the RT_VDT provides fast enough computation to be embedded in CPUs that are currently employed by ADAS systems. The particulars of the employed algorithms together with their implementation are described in detail. Additionally, these algorithms and their various integration combinations are tested and their performance is evaluated using actual road images, and videos captured by the front-mounted camera of the car as well as on the KITTI benchmark with 87% average precision achieved. The evaluation of the RT_VDT shows that it reliably detects and tracks vehicle boundaries under various conditions.

Download Full-text

A Fast and Reliable Balanced Approach for Detecting and Tracking Road Vechiles

Recent Advances in Computer Science and Communications ◽

10.2174/2666255813999200727163102 ◽

2020 ◽

Vol 13 ◽

Author(s):

Wael Farag

Keyword(s):

Real Time ◽

Vehicle Detection ◽

Machine Learning Algorithms ◽

Detection And Tracking ◽

Balanced Approach ◽

Vehicle Detection And Tracking ◽

Driving Assistance ◽

Rgb Images ◽

Assistance Systems ◽

Self Driving Cars

In this paper, an advanced-and-reliable vehicle detection-and-tracking technique is proposed and implemented. The Real-Time Vehicle Detection-and-Tracking (RT_VDT) technique is well suited for Advanced Driving Assistance Systems (ADAS) applications or Self-Driving Cars (SDC). The RT_VDT is mainly a pipeline of reliable computer vision and machine learning algorithms that augment each other and take in raw RGB images to produce the required boundary boxes of the vehicles that appear in the front driving space of the car. The main contribution of this paper is the careful fusion of the employed algorithms where some of them work in parallel to strengthen each other in order to produce a precise and sophisticated real-time output. In addition, the RT_VDT provides fast enough computation to be embedded in CPUs that are currently employed by ADAS systems. The particulars of the employed algorithms together with their implementation are described in detail. Additionally, these algorithms and their various integration combinations are tested and their performance is evaluated using actual road images, and videos captured by the front-mounted camera of the car as well as on the KITTI benchmark. The evaluation of the RT_VDT shows that it reliably detects and tracks vehicle boundaries under various conditions

Download Full-text

Deteksi Kendaraan secara Real Time menggunakan Metode YOLO Berbasis Android

Jurnal Sustainable Jurnal Hasil Penelitian dan Industri Terapan ◽

10.31629/sustainable.v9i1.1401 ◽

2020 ◽

Vol 9 (1) ◽

pp. 8-14

Author(s):

Junita Sri Wisna Hutauruk ◽

Tekad Matulatan ◽

Nurul Hayaty

Keyword(s):

Real Time ◽

Vehicle Detection ◽

Training Process ◽

Congestion Problems ◽

Bounding Boxes

The Activities on the highway that involve vehicles often have congestion problems due to the tightening of the quantity of vehicles on the highway. In addition, there are also problems of order and violations, the use of improper routes, such as vehicles entering the lane that are not intended for the vehicle. Therefore, researchers designed a vehicle detection application in real time based on Android using the YOLO method (You Only Look Once). The analysis carried out using 200 datasets, 4 classes, 10 batches, and 200 epochs. The training process was carried out up to 4000 steps, and the storage of checkpoints to the form of file protob was done at steps 800, 1000, 1200, 1400, 1600, 1800, 2000, 3000, and 4000. Bounding boxes successfully detected and classified objects correctly. This test is done using a Xiaomi Redmi 4X smartphone with a video resolution measuring 768x432 pixels.

Download Full-text