| Summary: | 碩士 === 朝陽科技大學 === 工業工程與管理系 === 104 === Fish processing products are one of the common food in our daily lives. Since fish products have highly economic value, the main form of direct consumption of cooked fish, raw fish, processed products (preserved fish, canned food, fish floss, etc.) is fairly extensive. When fish is often consumed, fish bones are some of the most frequently ingested foreign objects encountered in foods. In the production of fish flosses, fishbone detection is performed by human inspection using their sense of touch and vision which can lead to misclassification. When consumers eat fish floss with bones, it may cause harm to the health of consumers. Therefore, this study proposes an automated fishbone inspection method of fish floss products.
In this study, we propose a novel approach to automatically detect fish bones on fish floss surfaces. The proposed method applies the Curvelet Transform (CT) with square-ring high-pass energy filtering to remove the structural textures of background and delete the angle direction of background texture. Then the filtered image is inversely transformed back to spatial domain. In the reconstructed image, the background texture is attenuated and the fishbone areas are enhanced. Finally, a thresholding value is determined by statistical interval estimation and the restored image can be easily segmented to into two categories namely dark fish bones, and white background. Experimental results show that the proposed method achieves a high 84.3% flaw detection rate (1-β), a low 0.15% false alarm rate(α), a high 99.77% correct classification rate (CR).
|
|---|
