| 摘要: |
| 针对实况海域大黄鱼摄食行为识别受环境影响大的问题,以大黄鱼为对象,提出一种波高信息和图像纹理特征融合的大黄鱼摄食行为量化方法。研究基于GLCM纹理特征的提取方法,利用灰度共生矩阵计算鱼群摄食前后纹理特征的四个二次统计量(能量、相关性、对比度、逆差距)。同时考虑纹理特征受海面波高影响的问题,以小波高变化为例研究其对纹理特征的识别影响。以人工分类大黄鱼480帧图像的摄食状态为基础,判断单一纹理特征模型的识别精度,结果模型准确率达到了80％。结果表明波高在0.31cm以上时对应的部分图像存在识别误差,误差可能性随波高增加而扩大。通过引入波高修正参数,优化单一纹理特征模型,并以另480帧大黄鱼摄食前后图像测试。海试结果说明该方法在实况海域识别准确率达到93％,能有效实现实况海域大黄鱼摄食行为量化。 |
| 关键词: 大黄鱼 摄食行为 波高 图像纹理 |
| DOI： |
| 分类号:S951.2??????? |
| 基金项目:宁波市科技创新2025专项(2020Z076)；宁波市自然科学(2023J172) |
|
| Quantitative study on feeding behavior of large yellow croaker in real sea area |
|
zhaojiahao1, JIANG Chu-hua2, CHEN Jun-hua2, LI Hao1
|
|
1.Ningbo University;2.College of Science and Technology, Ningbo University
|
| Abstract: |
| Aiming at the problem that the feeding behavior recognition of Larimichthys crocea in the live sea area is greatly affected by the environment, taking Larimichthys crocea as an object, a quantitative method of feeding behavior of Larimichthys crocea based on the fusion of wave height information and image texture features was proposed. Research on the extraction method of texture features based on GLCM, using the grayscale co-occurrence matrix to calculate four quadratic statistics (energy, correlation, contrast, inverse difference) of texture features before and after fish feeding. At the same time, considering the issue of texture feature methods being affected by sea surface wave height, taking the variation of wavelet height as an example to study its recognition impact on texture feature methods. Artificial experience is used to determine 480 frames of images and classify the feeding intensity of fish schools, and to determine the recognition accuracy of a single texture feature model, with an accuracy rate of 80%. The results indicate that there is recognition error in the corresponding part of the image when the wave height is above 0.31cm, and the possibility of error increases with the increase of wave height. The single texture feature model was optimized by introducing wave height correction parameters, and another 480 images of Larimichthys crocea before and after feeding were tested. The sea trial results show that the accuracy of the method in the live sea area is 93%, which can effectively quantify the feeding behavior of Larimichthys crocea in the live sea area. |
| Key words: large yellow croaker Feeding behavior Wave height Image texture |