In recent years, as accidents due to tire wear have increased, a diagnosis system to alarm the tire wear condition is needed. An image based tire wear monitoring and alarming system is presented in this paper. A camera is installed inside the wheel house and the image acquired by the camera is used ...
In recent years, as accidents due to tire wear have increased, a diagnosis system to alarm the tire wear condition is needed. An image based tire wear monitoring and alarming system is presented in this paper. A camera is installed inside the wheel house and the image acquired by the camera is used as the input of the implemented system. The following algorithm is proposed to detect the number of pixels in a tire tread pattern and to determine tire wear condition. At first, the original image is converted into a grayscale image, and then bilateral filtering is performed to remove the noise in the input image. Binarization operation is applied to the image using a predetermined threshold and the number of pixels of the final image is counted. As the driving distance of the automobile increases, the depth of the tread on the tire wears is lowered. As a result, the number of pixels of the worn tire decreases linearly. This is modeled as a linear function in the paper and using the function and the number of pixels obtained from image processing the driving distance of the tire can be estimated. In order to get realistic experimental data, four images are taken for four different driving distances. In this paper, images are taken after 1,000 km, 10,000 km, 20,000 km, and 40,000 km driving and those are used to make a linear modeling equation for tire wear condition estimate. After then the equation is used to estimate the driving distances of the used tires and meaningful result was obtained. Thus, if the driving distance of the tire exceeds safety limit, currently about 40,000 km, the system can warn the tire status to the driver in real time. The proposed system is implemented in a commercially available vehicle and the performance is verified through real vehicle experiments. It is found that the processing time is very short because of its simple algorithm and hardware can be implemented at a very low price if it is optimized at the current detection condition. However, more research is needed to detect tire wear on images that are aggravated by weather, road surface conditions, and camera lens contamination. Also, more experiments for each model of tires of different brands are needed to guarantee the performance of the proposed approach.
In recent years, as accidents due to tire wear have increased, a diagnosis system to alarm the tire wear condition is needed. An image based tire wear monitoring and alarming system is presented in this paper. A camera is installed inside the wheel house and the image acquired by the camera is used as the input of the implemented system. The following algorithm is proposed to detect the number of pixels in a tire tread pattern and to determine tire wear condition. At first, the original image is converted into a grayscale image, and then bilateral filtering is performed to remove the noise in the input image. Binarization operation is applied to the image using a predetermined threshold and the number of pixels of the final image is counted. As the driving distance of the automobile increases, the depth of the tread on the tire wears is lowered. As a result, the number of pixels of the worn tire decreases linearly. This is modeled as a linear function in the paper and using the function and the number of pixels obtained from image processing the driving distance of the tire can be estimated. In order to get realistic experimental data, four images are taken for four different driving distances. In this paper, images are taken after 1,000 km, 10,000 km, 20,000 km, and 40,000 km driving and those are used to make a linear modeling equation for tire wear condition estimate. After then the equation is used to estimate the driving distances of the used tires and meaningful result was obtained. Thus, if the driving distance of the tire exceeds safety limit, currently about 40,000 km, the system can warn the tire status to the driver in real time. The proposed system is implemented in a commercially available vehicle and the performance is verified through real vehicle experiments. It is found that the processing time is very short because of its simple algorithm and hardware can be implemented at a very low price if it is optimized at the current detection condition. However, more research is needed to detect tire wear on images that are aggravated by weather, road surface conditions, and camera lens contamination. Also, more experiments for each model of tires of different brands are needed to guarantee the performance of the proposed approach.
주제어
#타이어 마모 진단 영상처리
※ AI-Helper는 부적절한 답변을 할 수 있습니다.