Biochip　plays　more　and　more　important　role　in　medical　analysis　and　test.　There　is　a　need　to　enhance　the　effectiveness　of　biochip　inspection.　To　detect　automatically　the　print　defects　of　the　PET　board　of　glycemia　detection　biochip,　the　fast　Hough　Transform　algorithm　was　used　for　sub-image　segmentation　based　on　circle　detection,　and　a　modified　method　with　limited-value-based　image　subtraction　was　applied　to　recognize　the　print　patterns.　The　corresponding　software　was　developed　to　realize　the　automatic　inspection　function.　The　results　indicated　that　the　presented　approach　is　applicable　and　can　fulfill　real-time　detection　accurately　and　effectively.