Laser Stripe Center Extraction Method of Rail Profile in Train-Running Environment

Robust Analysis and Laser Stripe Center Extraction for Rail Images

“…In the work by Liu et al [2], regions containing the rail waist light stripe are tracked using a Kalman filter, and the extreme value method, center of mass, and Hessian matrix are then combined to extract the precise sub-pixel coordinates of the light stripe centers. In the method proposed by Wang et al [20], the deep learning model ENet was applied to segment the rail laser stripe using the grayscale and gradient direction, which makes it more robust to noise. Nevertheless, the construction of the ENet model is complicated, and due to the influence of changing ambient light, the grayscale and gradient information of the same segment may not be consistent.…”

A real time detection method for abnormal strapping of steel coil based on CCD active imaging

“…This method adopts fixed templates in specific environment and has certain anti-interference ability; Cai et al [11] used Otsu segmentation algorithm to extract laser stripe, and then combined with principal component analysis to obtain stripe normals. In terms of extraction efficiency, this method avoids multiple convolution calculations, but there are certain segmentation errors in Otsu method in uneven illumination environment; Li et al [12] combined the traditional center of gravity algorithm with the threshold contour tracking algorithm to avoid scanning the non laser stripe area in the image, so as to improve the extraction speed of the stripe; Wang et al [13] used the deep learning model to segment the laser stripe, and then used the center of gravity method to calculate the center of the light stripe, which increased the extraction accuracy, but the algorithm had a certain time complexity; Liu et al [14] used the gradient threshold to effectively segment the stripe, took the cross-correlation maximum as the initial light strip center, and then accurately obtained the stripe center through curve fitting; Wang et al [15] removed the influence of uneven fixed background based on the difference method, and then segmented the stripe. After eliminating the influence of speckle noise by the regional growth statistical method, the center of gravity method was used to extract the stripe center; Feng et al [16] used the convex hull theory of Harris corner detection to detect the target region of the image, which can effectively separate the object and background, and finally segment it based on fuzzy c-means; Chmelar P et al [17] used RGB color space (Red,Green,Blue) to highlight the area of laser stripe.…”