Bionic vision detection system for surface defects of castings

Abstract

The invention relates to a detection system for surface defects of castings, in particular to a bionic vision detection system for surface defects of castings. On the basis of the binocular vestibulo-ocular compound motion control mechanism of the human eyes, a binocular camera adopting liquid lenses similar to the self-adaptive focusing of the human eyes acquires the image information of a to-be-detected casting sample, and autonomous acquisition, transmission, processing, storage and understanding are carried out for the image information; the computer image technology is sufficiently combined to process and store the image information, the machine learning technology is applied, an advanced neural network detection algorithm is adopted to compare the acquired product image information with sample information feature parameters in an acceptable product database, and thereby acceptable products and defective products are effectively distinguished; and the system does not require workers to operate on the site in a workshop, solves the problems of manual detection, i.e. low efficiency and poor precision, and the problems of conventional visual surface detection, such as the existence of blind areas and fuzzy images, ensures that detection operation is more intelligent, increases the efficiency and precision of detection, and reduces labor cost.

Description

technical field [0001] The invention relates to a casting surface defect detection system, in particular to a casting surface defect bionic visual detection system. Background technique [0002] In the process of casting production, defects such as pores, cracks, and scratches will inevitably occur on the surface of casting products. For a long time, casting companies have carried out necessary defect detection on the surface of castings to ensure product quality. At present, domestic foundries mostly use manual sampling and traditional machine vision methods to inspect the surface of castings. Manual sampling inspection requires quality inspectors to inspect the appearance of the sampled products with naked eyes, and manually sort the qualified and unqualified products one by one, which requires a lot of labor and adds huge labor and management costs to the enterprise. In addition, manual The sampling detection method mainly relies on manual experience, and there are prob...

AI Technical Summary

Problems solved by technology

[0004] The present invention mainly aims at the deficiencies of the prior art, and provides a bionic visual detection system for casting surface defects. The system adopts the control mechanism of binocular, triangular, and eye-movement compound movements of human eyes, and adopts an advanced surface detection algorithm to independently detect the surface of casting products. The system does not require staff to work on-site in the workshop, and solves the problems of low efficiency and poor accuracy of manual inspection, blind spots and blurred images in traditional surface visual inspection, making inspection operations more intelligent, improving inspection efficiency and accuracy, and reducing labor costs

Images