Research On Low Contrast Imaging Object Size Measurement System Based On Machine Vision

Prisms are widely used in optical instruments and are precision machined product with extremely strict size requirements.Relying on manual product detection is inefficient and prone to problems such as missed detection and false detection,it is difficult to ensure the accuracy and stability of the detection results.A detection system based on machine vision can solve this problem.Most of the current visual inspection systems study high-contrast images,and there are few studies on low-contrast images.Because the prism has a high light transmittance,the contrast of the image received by the CCD camera is low,and the image area of the prism cannot be well distinguished.Besides,the light-reflecting characteristics of the prism surface also limit the application of traditional machine vision inspection methods in the size measurement of such products.Therefore,this paper studies the size measurement of low-contrast imaging objects.Taking the prism as an example,a non-contact measurement method of object size based on machine vision for low-contrast imaging is proposed.The specific research contents are divided into the following points.(1)Design and construction of a visual inspection system for low-contrast imaging objects.Based on machine vision and aiming at the prism imaging characteristics of low-contrast imaging,complete the hardware selection and measurement system construction,and cooperate with the computer to complete the work of image acquisition,system calibration,and distortion correction.(2)Locate the region of interest in low-contrast images.A method to automatically obtain the region of interest of the part to be tested based on BLOB analysis is proposed,and the experiment shows that it can effectively find the prism area of each image to be tested.(3)Edge positioning method of low-contrast image.The positioning is divided into two parts: pre-processing and post-processing.The pre-processing combines the image average denoising method with the Contrast Limited Adaptive Histogram Equalization algorithm.Realize the low-contrast image enhancement under the premise of controlling the noise amplification.Use filtering and morphological operations to smooth the image,and finally,high-precision edge positioning is achieved by the Zernike moment sub-pixel edge detection algorithm.The optimized low-contrast image edge location method can enhance the contrast of low-contrast prism images under the premise of the reasonable control of noise amplification,and facilitate the detection of sub-pixel edge points.(4)Comparative experiment and result analysis.Taking the right-angle prism as a sample,the size measurement system was verified experimentally on the experimental platform,and the thickness and angle of the two types of right-angle prisms were measured.The accuracy reached ±3～5μm,and the angle measurement accuracy was within ±36″;A variety of special-shaped prisms have completed the thickness measurement,and the measurement accuracy is within ±5μm.The measurement experiment verifies the accuracy and stability of the optimization algorithm proposed in this article.Compared with the expensive high-precision image size measuring instrument on the market,it has certain advantages.The vision measurement system built in this paper and the optimized image processing method proposed in this paper can effectively solve the problem of size measurement of low-contrast imaging objects,and the detection accuracy can meet the requirements of industrial production,providing an effective method for the size measurement of low-contrast imaging objects.