Simultaneous Multi-vehicle Detection and Tracking Framework with Pavement Constraints Based on Machine Learning and Particle Filter Algorithm

Due to the large variations of environment with ever-changing background and vehicles with different shapes, colors and appearances, to implement a real-time on-board vehicle recognition system with high adaptability, efficiency and robustness in complicated environments, remains challenging. This paper introduces a simultaneous detection and tracking framework for robust on-board vehicle recognition based on monocular vision technology. The framework utilizes a novel layered machine learning and particle filter to build a multi-vehicle detection and tracking system. In the vehicle detection stage, a layered machine learning method is presented, which combines coarse-search and fine-search to obtain the target using the AdaBoost-based training algorithm. The pavement segmentation method based on characteristic similarity is proposed to estimate the most likely pavement area. Efficiency and accuracy are enhanced by restricting vehicle detection within the downsized area of pavement. In vehicle tracking stage, a multi-objective tracking algorithm based on target state management and particle filter is proposed. The proposed system is evaluated by roadway video captured in a variety of traffics, illumination, and weather conditions. The evaluating results show that, under conditions of proper illumination and clear vehicle appearance, the proposed system achieves 91.2% detection rate and 2.6% false detection rate. Experiments compared to typical algorithms show that, the presented algorithm reduces the false detection rate nearly by half at the cost of decreasing 2.7%–8.6% detection rate. This paper proposes a multi-vehicle detection and tracking system, which is promising for implementation in an on-board vehicle recognition system with high precision, strong robustness and low computational cost.

Hardening Effect on Machined Surface for Precise Hard Cutting Process with Consideration of Tool Wear

During hard cutting process there is severe thermodynamic coupling effect between cutting tool and workpiece, which causes quenching effect on finished surfaces under certain conditions. However, material phase transformation mechanism of heat treatment in cutting process is different from the one in traditional process, which leads to changes of the formation mechanism of damaged layer on machined workpiece surface. This paper researches on the generation mechanism of damaged layer on machined surface in the process of PCBN tool hard cutting hardened steel Cr12MoV. Rules of temperature change on machined surface and subsurface are got by means of finite element simulation. In phase transformation temperature experiments rapid transformation instrument is employed, and the effect of quenching under cutting conditions on generation of damaged layer is revealed. Based on that, the phase transformation points of temperature under cutting conditions are determined. By experiment, the effects of cutting speed and tool wear on white layer thickness in damaged layer are revealed. The temperature distribution law of third deformation zone is got by establishing the numerical prediction model, and thickness of white layer in damaged layer is predicted, taking the tool wear effect into consideration. The experimental results show that the model prediction is accurate, and the establishment of prediction model provides a reference for wise selection of parameters in precise hard cutting process. For the machining process with high demanding on surface integrity, the generation of damaged layer on machined surface can be controlled precisely by using the prediction model.