Image Detection Method for Broccoli Seedlings in Field Based on Faster R-CNN

被引：0

作者：

Sun Z. ^{[1
]}

Zhang C. ^{[1
]}

Ge L. ^{[1
]}

Zhang M. ^{[1
]}

Li W. ^{[1
]}

Tan Y. ^{[1
]}

机构：

[1] College of Engineering, China Agricultural University, Beijing

来源：

Nongye Jixie Xuebao/Transactions of the Chinese Society for Agricultural Machinery | 2019年 / 50卷 / 07期

关键词：

Broccoli seedlings; Convolutional neural network; Crop recognition; Deep learning; Faster R-CNN;

D O I：

10.6041/j.issn.1000-1298.2019.07.023

中图分类号：

学科分类号：

摘要：

Traditional methods of image processing for crop detection under agricultural natural environment are easily affected by small samples and subjective judgment, so they have many disadvantages such as low recognition rate and low robustness. Deep learning can self-study according to data set, and has a strong ability to express feature. Therefore, a new broccoli seedlings detection approach based on Faster R-CNN model was proposed. Data acquisition was the first step to build deep learning model, and the diversity of data can improve the generalization ability of the model. According to the characteristics of field environment, broccoli seedlings images with different light intensities, different ground moisture contents and different weed densities were collected. The sample size was expanded by images rotation and noise enhancement, and data set was transformed as PASCAL VOC format. And then the Faster R-CNN model was trained by using data set. Contrast experiment was designed on ResNet101, ResNet50 and VGG16 networks. The results showed that ResNet101 network with the deepest network layer and smaller parameter space was the best feature extraction network. The average detection accuracy was 90.89%, and the average time-consuming was 249 ms. Based on that, the network super-parameters were optimized and the average accuracy of model detection reached 91.73%, when Dropout value was 0.6. The results showed that this approach can effectively detect broccoli seedlings in agricultural natural environment, and provided a hopeful solution for crop detection in the field of agriculture. © 2019, Chinese Society of Agricultural Machinery. All right reserved.

引用

页码：216 / 221

页数：5

共 23 条

[1] Liu W., Wang E., Cai J., Et al., High density cultivation techniques of early-middle maturing broccoli at different seedling ages, Journal of Agriculture, 8, 11, pp. 49-53, (2018)
[2] Piron A., Selection of the most efficient wavelength bands for discriminating weeds from crop, Computers and Electronics in Agriculture, 62, 2, pp. 141-148, (2008)
[3] Li X., Zhu W., Ji B., Et al., Weed identification based on features optimization and LS-SVM in the cotton field, Transactions of the Chinese Society for Agricultural Machinery, 41, 11, pp. 168-172, (2010)
[4] Hu L., Luo X., Zeng S., Et al., Plant recognition and localization for intra-row mechanical weeding device based on machine vision, Transactions of the CSAE, 29, 10, pp. 12-18, (2013)
[5] Li N., Grift T.E., Yuan T., Et al., Image processing for crop/weed discrimination in fields with high weed pressure, American Society of Agricultural and Biological Engineers, (2016)
[6] Christine M., Onyango J.A., Segmentation of row crop plants from weeds using colour and morphology, Computers and Electronics in Agriculture, 39, 3, pp. 141-155, (2003)
[7] Si C., Weed-plant discrimination by machine vision and artificial neural network, Biosystems Engineering, 83, 3, pp. 275-280, (2002)
[8] Liu H., Chen Z., Feng Y., Real-time weed recognition method based on color and morphological features, Opto-Electronic Engineering, 33, 7, pp. 96-100, (2006)
[9] Li Y., Zhang L., Yan W., Et al., Weed identification using imaging spectrometer data, Journal of Remote Sensing, 17, 4, pp. 855-871, (2013)
[10] Chen Y., Jin X., Tang L., Et al., Intra-row weed recognition using plant spacing information in stereo images, American Society of Agricultural and Biological Engineers, (2013)

← 1 2 3 →