Apparatus and method for performing quality inspection of a product surface

A method for performing quality inspection of a product surface is disclosed. The method comprises the steps of: generating, by a lighting device, a plurality of different lighting environments; obtaining, by a camera device, a plurality of surface images by photographing a sample to be inspected in the plurality of lighting environments; and analyzing, by a vision inspection device, the plurality of surface images based on a deep neural network model to determine surface defects of the sample to be inspected. On mass production lines that inspect the surface quality of samples to be inspected, the plurality of surface images of the sample to be inspected obtained in the plurality of lighting environments are inferred using a deep neural network, and based on the inference results, surface defects (micro scratches, waviness, etc.) of the sample to be inspected can be inspected. 제품 표면의 품질 검사를 수행하기 위한 방법이 개시된다. 이 방법은, 조명 장치가, 서로 다른 복수의 조명 환경들을 생성하는 단계, 카메라 장치가, 상기 복수의 조명 환경들에서 하나의 검사 대상 샘플을 촬영하여 복수의 표면 이미지들을 획득하는 단계, 및 비전 검사 장치가, 심층 신경망 모델을 기반으로 상기 복수의 표면 이미지들을 분석하여, 상기 검사 대상 샘플의 표면 불량을 판정하는 단계를 포함한다.