摘要
针对金鱼游动方向自动监测,提出了一种基于机器视觉金鱼游动方向的监测算法。金鱼图像首先通过阈值分割得到分割图像,然后通过分割图像或者腐蚀后的分割图像提取金鱼的像素点,再分别使用最小二乘法拟合一次、二次曲线从而得到金鱼的游动方向。15幅金鱼图像的实验表明在室内单条金鱼监测中通过分割图像提取像素后一次曲线拟合得到的结果最好,仅6.67%的误差。
A new algorithm based on machine vision for the gold fish swim direction monitoring was proposed.The original image that first segmented by thresholding segmentation algorithm,and the pixels of gold fish were obtained from segmentation image or erosion image that was eroded from segmentation image.The swim direction was found from linear line or cure line that was processed via least square method.The results of 15 gold fish images indoor experiments for single gold fish monitoring showed that the swim direction from linear line and segmentation image had the best effect,and the error of it only had 6.67%.
出处
《安徽农业科学》
CAS
北大核心
2010年第12期6567-6568,共2页
Journal of Anhui Agricultural Sciences
基金
广西工学院硕士科研启动基金资助项目(院科硕0816210)
