To address the issues of overexposure distortion and challenging identification of pellet adhesion in pellet images,an enhanced contrast-limited adaptive histogram equalization method was proposed for image enhancement,along with the introduction of a multi-scale Hough detection algorithm for efficient pellet detection.Firstly,the en-hanced contrast-limited adaptive histogram equalization method incorporates a detail security mechanism by utilizing a contrast enhancement function to adjust the output of contrast-limited adaptive histogram equalization(CLAHE).The extent of contrast improvement is limited by this approach,while simultaneously preserving more detailed infor-mation in the images.Then,the image noise induced after enhancement is filtered by the adaptive median filtering method.Finally,the multi-scale Hough detection algorithm is applied to tackle both adhesion pellet identification and detection challenges,so as to realize the accurate,real-time and efficient detection and statistics of pellet size.Practical application results from a steel factory reveal that this method achieves a precision rate of as high as 94.44％,with 96.72％accuracy in determining the number of qualified pellets,fulfilling the requirements of pellet preparation process.This method plays a vital role in enhancing production effectiveness,decreasing labor expen-ses,and enhancing overall quality control rates for pellet preparation.
NETL
NSTL
万方数据
