A method for image recognition of lotus plant targets for picking robots
A picking robot and target image technology, applied in the field of lotus pod target image recognition, can solve the problem of large fruit damage
Active Publication Date: 2018-04-24
HUAWEI TEHCHNOLOGIES CO LTD
View PDF3 Cites 0 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
[0002] The development and development of picking robots began in the United States in 1968. There are mainly mechanical shaking and pneumatic shaking, but both of them are more harmful to the fruit.
Method used
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View moreImage
Smart Image Click on the blue labels to locate them in the text.
Smart ImageViewing Examples
Examples
Experimental program
Comparison scheme
Effect test
specific Embodiment approach
[0040] Describe in detail below in conjunction with accompanying drawing and main steps, and specific implementation is as follows:
[0041] Such as figure 1 Shown is the main flow chart of the present invention, mainly comprises the training of image recognition and the test part, obtains m invariant moment principal component components and K=4 cluster centers of K-Means clustering by the training part, and the main The components and cluster centers are transmitted to the test part for the purpose of testing the accuracy of lotus pod recognition.
[0042] 1. Using the combination of image Gaussian filter and super green index method, using the designed super green Gaussian filter to remove complex background. Capture images such as figure 2 As shown, the figure includes lotus pods, lotus leaves, lotus flowers, and stems, which are used for image processing to obtain image 3 Binary images of the lotus pod, lotus leaf, lotus flower, and stem shown. The image preprocessi...
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More PUM
Login to View More Abstract
The invention discloses a lotus pod target image recognition method for a picking robot, which adopts the method of combining the image Gaussian filter with the ultra-green index method, designs the ultra-green Gaussian filter, and removes the complex background; adopts the improved morphological watershed segmentation algorithm to The overlapped images can be separated and segmented; the Hu invariant moment algorithm is improved to calculate the invariant moments an of lotus pods, lotus leaves, lotus flowers, and stems, and the calculated nth order invariant moments are linearly combined to obtain Moment invariant principal components zm representing the different shape features of lotus pods, lotus leaves, lotus flowers, and stems; for image target recognition, K-Means clustering algorithm is used to classify the moment invariant principal components zm of images of lotus pods, lotus flowers, lotus leaves, and stems. The nearest connected component of the main component separated from the cluster center of the pod is the pod. The invention can effectively distinguish and recognize lotus pods, lotus leaves, lotus flowers and stems, and is the core algorithm technology of the vision system of the lotus pod picking robot.
Description
technical field [0001] The invention belongs to the field of agricultural machinery, and relates to a lotus pod target image recognition method for a picking robot. Background technique [0002] The development and development of picking robots began in the United States in 1968. There are mainly mechanical shaking and pneumatic shaking, but both of them are more harmful to the fruit. With the development of intelligence, automation, and industrialization, the development of foreign picking robots has advanced by leaps and bounds. Japan, the Netherlands, France, the United Kingdom and other countries have successfully tested many picking robots, such as tomato picking robots, grape picking robots, cucumber picking robots, watermelon picking robots, etc. Picking robots, cabbage picking robots, mushroom picking robots, etc. In China, the research on fruit and vegetable picking robots has just started. Lu Huaimin of Northeastern University successfully tested a forest fruit ba...
Claims
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More Application Information
Patent Timeline
Login to View More Patent Type & Authority Patents(China)
IPC IPC(8): G06K9/00G06K9/46
Inventor 赵德安唐书萍陈玉贾伟宽
Owner HUAWEI TEHCHNOLOGIES CO LTD