Abstract: In response to the challenges of laborintensive and timeconsuming manual aquatic plant cleaning in crab farming, a method for imageguided navigation line fitting of water grass in crab ponds is proposed. The proposed approach involves several steps. First, RGB images of aquatic plants are read and converted into HSV color space. The H component is selected for binary segmentation, and the segmented image is then filled. Secondly, based on the differences in the water grass image area, parameters are set to delete the small area parts, and the processed image is processed by morphology to retain the target area. Then, the midpoint of nonzero pixels in the binary image is searched by column to obtain the morphological characteristic curve of the aquatic plant image. Finally, based on the characteristic curve, the navigation line is fitted using the least square method. The experimental results show that the relative error of the fitted navigation line is within 0.498%. The average absolute value of the relative error is 0.247%. The average processing time is 0.005s. The method proposed in this paper shows good accuracy and realtime performance in fitting navigation lines for aquatic plant images, which can lay a theoretical and technical foundation for the development of visually guided aquatic plant cleaning workboats.

Key words: aquatic plant image, HSV color space, image segmentation, navigation line fitting

摘要： 针对河蟹养殖中水草清理是以人工完成为主，存在费工费时、劳动效率低、成本高等问题，提出一种图像处理的蟹塘水草图像导航线拟合方法。首先通过采集水草RGB图像，并转换为HSV颜色空间，并选用H分量二值分割，对分割图像进行填充处理；其次根据水草图像面积差异，设定参数删除非目标区域，进行形态学运算，保留目标区域；再按列查找二值图像非零像素点的中点位置，得到水草图像形态特征曲线；最后根据查找的特征曲线，利用最小二乘法拟合出导航线。试验结果表明本文方法拟合导航线相对误差在0.498%以内，平均相对误差绝对值为0.247%，平均耗时为0.005s。本文方法能够较好地满足水草图像导航线拟合的准确性和实时性，可为智能化水草清理作业船的研制奠定理论与技术基础。

关键词: 水草图像, HSV颜色空间, 图像分割, 导航线拟合