![在计算机芯片上发现缺陷](http://m.czl106.com/library/media/products/suakit/suakit---carousel-2.jpg?sc_lang=ko-kr&h=350&w=500&la=ko-KR&hash=8C3966FD3692A74B08FC0F9FF11FDE3C)
제품품질및처리량개선
정교하게최적화된SuaKIT의딥러닝모델은매우정교한검사결과를제공합니다。딥러닝알고리즘의내부분석프로세스는품질과처리량을최적화하기위해서과검(过度)와미검(underkill)비율을감소시킵니다。
비용절감
자동화된시스템을구현하면신뢰도가낮은수작업검사에대한의존성을감소시킵니다。지속적으로검사를작업할수있다면처리량을최적화하고단위생산시간을개선해서고객수요를충족시킬수있습니다。또한SuaKIT의높은감지율은검사하드웨어추가설치비용을절감할수있습니다。
![橘色汽水瓶有缺陷](http://m.czl106.com/library/media/products/suakit/suakit---carousel-3.jpg?sc_lang=ko-kr&h=350&w=500&la=ko-KR&hash=B7EDF7AA58E3F6E7E0DB145B80EC5D59)
![零件上发现的缺陷](http://m.czl106.com/library/media/products/suakit/suakit---carousel-4.jpg?sc_lang=ko-kr&h=350&w=500&la=ko-KR&hash=66BEA7D1066C4FD5E0094FB9BEA63ADC)
신뢰할수있는검증된결과보장
모든각각의라인,작업시간,공장에대해서도SuaKIT의일관성이높은검사로동일한결과를보장합니다。이소프트웨어는저장된이미지와문서화된결과를오프라인에서검토및검증할수있습니다。가치가높은이데이터는애플리케이션을최적화하고변칙적인결과를이해할수있도록지원합니다。
핵심적인기능
감지
![发现蓝色和黄色汽车的SuaKIT检测特征的例子](http://m.czl106.com/library/media/products/suakit/features-_detection.jpg?sc_lang=ko-kr&h=500&w=700&la=ko-KR&hash=0B1D3CD2C6F849D024FC20F5B677137A)
단일이미지내에서다양한범주내의물체감지
분류
![SuaKIT分类特性的例子](http://m.czl106.com/library/media/products/suakit/features-_classification.jpg?sc_lang=ko-kr&h=500&w=700&la=ko-KR&hash=94753DD5917B84D8C82087294FC07991)
사전에결정된여러범주별로이미지그룹화
세그먼테이션
![蓝色的车,天窗被标记为红色](http://m.czl106.com/library/media/products/suakit/features-_segmentation.jpg?sc_lang=ko-kr&h=500&w=700&la=ko-KR&hash=B32803A613A7F549CA3471E64A620B54)
이미지상의위치,면적,형태,결함을정확하게파악
딥러닝의구조
단일이미지분석
![](http://m.czl106.com/library/media/products/suakit/features-_single-image-analysis.jpg?sc_lang=ko-kr)
각이미지를학습하고결함파악
이미지비교
![](http://m.czl106.com/library/media/products/suakit/features-_image-comparison.jpg?sc_lang=ko-kr)
이미지두장으로구성된한세트사이의차이점에집중함으로써결함을학습하고파악
다중이미지분석
![](http://m.czl106.com/library/media/products/suakit/features-_multi-image-analysis.jpg?sc_lang=ko-kr)
결함감지모델학습을위해이미지들사이의관계분석
단일범주학습
![](http://m.czl106.com/library/media/products/suakit/features-_one-class-learning.jpg?sc_lang=ko-kr)
학습한정상이미지와의편차에기초해서결함파악