Pseudo-Image and Sparse Points: Vehicle Detection With 2D LiDAR Revisited by Deep Learning-Based Methods

Abstract

Detecting and locating surrounding vehicles robustly and efficiently are essential capabilities for autonomous vehicles. Existing solutions often rely on vision-based methods or 3D LiDAR-based methods. These methods are either too expensive in both sensor pricing (3D LiDAR) and computation (camera and 3D LiDAR) or less robust in resisting harsh environment changes (camera). In this work, we revisit the LiDAR based approaches for vehicle detection with a less expensive 2D LiDAR by utilizing modern deep learning approaches. We aim at filling in the gap as few previous works conclude an efficient and robust vehicle detection solution in a deep learning way in 2D. To this end, we propose a learning based method with the input of pseudo-images, named Cascade Pyramid Region Proposal Convolution Neural Network (Cascade Pyramid RCNN), and a hybrid learning method with the input of sparse points, named Hybrid Resnet Lite. Experiments are conducted with our newly 2D LiDAR vehicle dataset recorded in complex traffic environments. Results demonstrate that the Cascade Pyramid RCNN outperforms state-of-the-art methods in accuracy while the proposed Hybrid Resnet Lite provides superior performance of the speed and lightweight model by hybridizing learning based and non-learning based modules. As few previous works conclude an efficient and robust vehicle detection solution with 2D LiDAR, our research fills in this gap and illustrates that even with limited sensing source from a 2D LiDAR, detecting obstacles like vehicles efficiently and robustly is still achievable.

Sparse representation optimization of Gaussian mixed feature of image based on convolution neural network

“…The neural network algorithm used in the sparse representation of image is an intelligent algorithm with "human biological characteristics" based on the overall structure of human neurons and the brain's direct two-way regulation and automatic processing of neurons. When we carry out random processing, mutual coupling analysis of bidirectional image information classification and vector processing analysis of multiple coupling combinations for a single neuron structure with multiple neuron structure (synapse) characteristics, we can achieve the optimization of multiple software types and reusable framework processing of software development [19]. If the image type with high modularity requirement is found from the total objects to be processed, the probability of being selected for secondary or multiple analysis is high.…”

Dialectical Analysis of Comparative Pedagogy Based on Multiple Intelligences Evaluation

“…rough the intelligent multielement evaluation system based on micro-nano flexible sensor, the evaluation and construction based on intelligent algorithm and evaluation method should be strengthened, the multi-design and attention of intelligent evaluation scheme should be improved, and the evaluation quality of comparative education should also be paid attention [4]. According to the multi-factor relationship theory in comparative education, researchers propose a new intelligent evaluation solution and analyze the relationship between the traditional dialectic analysis field and the integrated solution of intelligent evaluation [5]. e scholars improved the evaluation method by combining the principal component analysis and other relevant theories, constructing the intelligent evaluation system based on the traditional particle swarm optimization, and explaining the practical significance of the dialectical evaluation system with the theory of modern analytical science [6].…”

Development of a dynamic intelligent recognition system for a real-time tracking robot

“…It is a simple board containing a microcontroller, peripheral interfaces, power supply circuits, which is programmed by an existing software platform [22]. A portable mini-HD Wi-Fi camera is attached on the robot platform; it adopts P2P technology, which allows users to easily configure the camera mounted on an SG-90 servo motor for object tracking purpose [23]. An SRF-05 model ultrasonic sensor is also attached over an SG-90 servo motor for obstacle avoidance, it can sense obstacles from 0.01-to 4 meters space, and its ability connects to the Arduino board easily.…”