Automobile surface defects like scratches or dents occur during the process of manufacturing and cross-border transportation. This will affect consumers’ first impression and the service life of the car itself. In most worldwide automobile industries, the inspection process is mainly performed by human vision, which is unstable and insufficient. The combination of artificial intelligence and the automobile industry shows promise nowadays. However, it is a challenge to inspect such defects in a computer system because of imbalanced illumination, specular highlight reflection, various reflection modes and limited defect features. This paper presents the design and implementation of a novel automatic inspection system (AIS) for automobile surface defects which are the located in or close to style lines, edges and handles. The system consists of image acquisition and image processing devices, operating in a closed environment and noncontact way with four LED light sources. Specifically, we use five plane-array Charge Coupled Device (CCD) cameras to collect images of the five sides of the automobile synchronously. Then the AIS extracts candidate defect regions from the vehicle body image by a multi-scale Hessian matrix fusion method. Finally, candidate defect regions are classified into pseudo-defects, dents and scratches by feature extraction (shape, size, statistics and divergence features) and a support vector machine algorithm. Experimental results demonstrate that automatic inspection system can effectively reduce false detection of pseudo-defects produced by image noise and achieve accuracies of 95.6% in dent defects and 97.1% in scratch defects, which is suitable for customs inspection of imported vehicles.

The plastic zone developed during elastoplastic contact may be effectively modeled as an inclusion in an isotropic half space. This paper proposes a simple but efficient computational method to analyze the stresses caused by near surface inclusions of arbitrary shape. The solution starts by solving a corresponding full space inclusion problem and proceeds to annul the stresses acting normal and tangential to the surface, where the numerical computations are processed by taking advantage of the fast Fourier transform techniques with a parallel computing strategy. The extreme case of a cuboidal inclusion with one facet on the surface of the half space is chosen to validate the method. When the surface truncation domain is extended sufficiently and the grids are dense enough, the results based on the new approach are in good agreement with the exact solutions. When solving a typical elastoplastic contact problem, the present analysis is roughly two times faster than the image inclusion approach and six times faster than the direct method. In addition, the present work demonstrates that a significant enhancement in the computational efficiency can be achieved through the introduction of parallel computation.

Gears are widely used in machines to transmit torque from one shaft to another shaft and to change the speed of a power source. Gear failure is one of the major causes for mechanical transmission system breakdown. Therefore, early gear faults must be immediately detected prior to its failure. Once early gear faults are diagnosed, gear remaining useful life (RUL) should be estimated to prevent any unexpected gear failure. In this paper, an intelligent prognostic system is developed for gear performance degradation assessment and RUL estimation. For gear performance degradation assessment, which aims to monitor current gear health condition, first, the frequency spectrum of gear acceleration error signal is mathematically analyzed to design a high-order complex Comblet for extracting gear fault related signatures. Then, two health indicators called heath indicator 1 and health indicator 2 are constructed to detect early gear faults and assess gear performance degradation, respectively, using two individual dynamic Bayesian networks. For gear RUL estimation, which aims to predict future gear health condition, a general sequential Monte Carlo algorithm is applied to iteratively infer gear failure probability density function (FPDF), which is used to predict gear residual lifetime. One case study is investigated to illustrate how the developed prognostic system works. The vibration data collected from a gearbox accelerated life test are used in this paper, where the gearbox started from a brand-new state, and ran until gear tooth failure. The results show that the developed prognostic system is able to detect early gear faults, track gear performance degradation, and predict gear RUL.

No abstract