摘要
在养殖虾夷扇贝(Patinopecten yessoensis)的生产过程中需要多次将其按大小分级。传统方法采用筛子和分级机筛分,会使扇贝受到振动、碰撞。振动影响扇贝生长发育,碰撞使扇贝边缘受到损伤,贝壳碎裂外套膜裸露在壳外,造成病贝、死贝,且机械筛分分级精度低,人工筛分劳动强度大、效率低。本文研究一种新的方法,利用机器视觉检测扇贝大小。通过摄像头获取扇贝图像、计算机对输入的图像进行预处理、图像分割、膨胀腐蚀,提取扇贝的面积等特征值,建立扇贝的几何模型、数学模型,确定面积与壳长的关系,进一步识别扇贝的大小。试验表明,该方法检测速度快,正确率高,能够满足虾夷扇贝分级要求。摄像头与扇贝不接触,可以避免机械振动、碰撞对扇贝的损伤。
Grading is necessary in scallops （ Patinopecten yessoensis ） culturing and is commonly implemented by using screener or grading machine. Screening and mechanical grading are time-consuming and unreliable in scallops grading. Furthermore, scallops are often damaged or lead to death due to the vibration and impact of grading machine and screener, which can result in crushing of the shells and organs exposure of the scallops. A new detecting method of scallops grading based on machine vision was studied. The digital image of the scallop was firstly taken by a camera, and sent to a computer for image segmentation, dilation and erosion. The represented area of the scallop was estimated, and the geometric and mathematic models were accordingly established which could be used to determine the size of the scallop by recognizing the real area and the length of the scallop. Results showed that machine vision was feasible in scallop grading with fairly high efficiency and accuracy. Scallop didn＇ t contact with the camera in the grading process, therefore, vibration and impact damages were eliminated.
出处
《水产学报》
CAS
CSCD
北大核心
2006年第3期397-403,共7页
Journal of Fisheries of China
基金
国家自然科学基金(50475152)
高等学校博士学科点专项科研基金(98014106)
关键词
机器视觉
虾夷扇贝
分级
machine vision
Patinopecten yessoensis
grading
