基于形态学特征的对虾完整性识别方法构建

目的 针对对虾加工过程中缺损对虾混入完整对虾从而降低对虾产品外观品质的问题, 构建基于形态学特征的对虾完整性识别方法。方法 首先, 使用灰度差异法处理对虾图像, 经过连通域、中值滤波等形态学操作后, 得到较为完整的感兴趣区域图像, 再对其采取二值化、轮廓化等操作; 然后, 对轮廓提取骨架线, 并求轮廓内最大内切圆直径以得到长宽比特征, 并求其圆度特征; 最后, 将以上2个特征作为判别对虾完整性的核心指标, 构建融合特征判别算法。结果 本研究所提算法应用于1063幅生鲜虾测试集图像中识别准确率达到99.25%, 相比于传统曲率法, 识别准确率提升了6.48%, 识别时间降低了1598.6 ms。结论 该方法具有较大优势和应用前景, 为开发大规模南美白对虾在线品质的无损检测装备提供关键技术。

Construction of completeness recognition method for shrimp (Litopenaeus vannamei) based on morphological characteristics

Objective To construct a completeness recognition method for shrimp (Litopenaeus vannamei) based on morphological characteristics to solve the problem of appearance quality deterioration by mixing the incomplete shrimp mixing into the sound clustered shrimp during the shrimp production processing. Methods Firstly, the shrimp image was processed by background grayscale difference method, the region of interest (ROI) of shrimp image were obtain after the morphological operation, median filtering, double-value, contouring and other operations; then, the skeleton line was extracted from the contour, and the maximum inscribed circle diameter in the contour was calculated to obtain the length width ratio of shrimp and its roundness characteristics; finally, taking the above 2 features as the core indexes to judge the integrity of shrimp, the fusion feature discrimination algorithm was constructed. Results The algorithm established in this study was applied to 1063 test set images of fresh shrimp and its recognition accuracy reached 99.25%, the recognition accuracy was improved by 6.48%, and the recognition time was reduced by 1598.6 ms compared with the traditional curvature method. Conclusion The proposed method has great advantages and application prospects, which providing the key technology for the development of nondestructive testing equipment for large-scale online quality of Litopenaeus vannamei.