scispace - formally typeset
Search or ask a question
Author

Liu Weilin

Bio: Liu Weilin is an academic researcher. The author has contributed to research in topics: Image segmentation & Sensor fusion. The author has an hindex of 3, co-authored 4 publications receiving 29 citations.

Papers
More filters
Patent
10 Aug 2016
TL;DR: In this article, a cucumber disease identification method and apparatus based on image information was presented, which is related to the fields of system engineering and information technology and improves the accuracy of identification, and avoids subjectivity and limitation due to artificial identification.
Abstract: The invention discloses a cucumber disease identification method and apparatus based on image information, and relates to the fields of system engineering and information technology. The invention acquires a leaf disease image of a cucumber to be identified; carries out image segmentation for the leaf disease image to obtain segmented leaf patches; extracts features of the leaf patches to obtain disease feature information; carries out disease identification for the disease feature information according to a preset disease feature sample, to obtain the disease of the cucumber to be identified; and therefore, improves the accuracy of identification, and avoids subjectivity and limitation due to artificial identification.

14 citations

Patent
06 Jul 2016
TL;DR: In this paper, a greenhouse vegetable environmental parameter data fusion method, a data fusion device, and data fusion system is presented. But the method does not consider the influence of the arrangement positions of the sensors on the greenhouse integrated environmental parameters.
Abstract: The invention provides a greenhouse vegetable environmental parameter data fusion method, a data fusion device, and a data fusion system. The greenhouse vegetable environmental parameter data fusion method is characterized in that at least one environmental parameter value of greenhouse vegetables on various preset positions in a greenhouse at different moments can be acquired; various environmental parameter values at the same moment can be used to form environmental parameter sequences, and various environmental parameter sequences at different moments can be acquired; the weighting fusion calculation of the environmental parameter sequences can be carried out to acquire the fusion value of various environmental parameter sequences at different moments. By adopting the greenhouse vegetable environmental parameter data fusion method, the data fusion device, and the data fusion system, the technical problems of the prior art of the inaccurate monitoring of the greenhouse integrated environmental parameters caused by the lack of consideration of the influence of the arrangement positions of the sensors on the greenhouse integrated environmental parameters can be solved.

7 citations

Patent
12 Oct 2016
TL;DR: In this article, a disease spot image segmentation method and system for a greenhouse vegetable leaf was proposed, which consists of vegetable leaf disease-spot images collected at a greenhouse field are divided into a training set and a testing set and enhancement processing is carried out on the images; initial color classification features and gradient features of the training set images after enhancement processing are extracted.
Abstract: The invention provides a disease spot image segmentation method and system for a greenhouse vegetable leaf. The method comprises: vegetable leaf disease-spot images collected at a greenhouse field are divided into a training set and a testing set and enhancement processing is carried out on the images; initial color classification features and gradient features of the training set images after enhancement processing are extracted; the training set images after enhancement processing are classified into disease spot samples and leaf samples, and initial color classification feature data and gradient feature data of the disease spot samples and the leaf samples are obtained based on the extracted initial color classification features and gradient features; with a rough set method, color classification features are selected for the initial color classification feature data of the disease spot samples and leaf samples, thereby obtaining a color feature set; according to the color feature set and the gradient feature data of the disease spot samples and leaf samples, a condition random field model is constructed; and on the basis of the condition random field model, the testing set images after enhancement processing are segmented and disease-spot images are extracted. According to the invention, a disease-spot image can be extracted accurately and the speed is fast.

5 citations

Patent
31 Aug 2016
TL;DR: In this paper, a hierarchical analysis method is employed, and a weight value of each factor relative to the target layer is calculated, and each factor is divided into four different specific scoring values according to scoring standards.
Abstract: The invention provides an organic plant type product species evaluation method. The method comprises steps that organic plant type product risk evaluation is set as a target layer; determining main control factors of organic plant type product risk evaluation is set as a criterion layer; determining sub factors belonging to each main control factor is set as an index layer; 20-50 types of organic plant products are set as a scheme layer; a hierarchical analysis method is employed, and a weight value of each factor relative to the target layer is calculated; each factor is divided into four different specific scoring values according to scoring standards; according to the weight value of each factor relative to the target layer and the specific scoring values, a risk value of each species of the organic plant products of the scheme layer is calculated. According to the method, based on the hierarchical analysis method, in combination with the practical production condition of the organic products, risks of 39 types of organic plant products such as wheat, corn and paddy rice in the production process are identified, the risk of each species is evaluated, and the 39 organic plant type products are ordered according to the risk values.

3 citations


Cited by
More filters
Patent
22 Feb 2017
TL;DR: In this article, a crop disease identification method based on incremental learning is presented, where a negative correlation integrated neural network (NIL) classifier is used to identify crop pests and diseases.
Abstract: The invention provides a crop disease identification method based on incremental learning. When new data arrive, continuous learning is carried out based on an original learning result, and the capability of progressive learning is achieved, which means that new knowledge can be obtained from new samples obtained by batch and the performance is gradually improved under a condition that original knowledge is effectively kept. Firstly, a crop disease sample database is collected, and simulation incremental learning of disease images in the sample database is carried out using a negative correlation integrated neural network as main technical means, so that an initial parameter of a negative correlation learning system is determined, an integrated neural network classifier based on negative correlation learning is initialized based on the initial parameter, and the classifier is trained using a sample in an initial stage; in an incremental learning stage, when an expert adds a new sample in the sample database, the integrated neural network classifier based on negative correlation learning only is updated by only training the newly-added sample data, so that the object of incremental learning is achieved; and finally, a diagnosis result of a disease picture and control measures are fed back to a user, so that the pest and disease can be accurately identified and diagnosed, and the object of comprehensive crop control is achieved.

13 citations

Patent
14 Jul 2017
TL;DR: In this article, an instrument positioning method applicable to a substation patrol inspection robot is described, which firstly preliminary positioning is performed on an image to be detected by using an Adaboost classifier, then secondary positioning on a plurality of candidate regions acquired by positioning by using SVM classifier and a region which is judged to be an instruction in the two times of positioning is the regional location of the instrument in the image.
Abstract: The invention discloses an instrument positioning method applicable to a substation patrol inspection robot, which is characterized in that firstly preliminary positioning is performed on an image to be detected by using an Adaboost classifier, then secondary positioning is performed on a plurality of candidate regions acquired by positioning by using an SVM classifier, and a region which is judged to be an instruction in the two times of positioning is the regional location of the instrument in the image. According to the invention, regions similar to the instrument are positioned by using the Adaboost classifier, so that the detection rate of the instrument is greatly improved; deficiencies of the Adaboost classifier are made up by using the SVM classifier, color features and textural features are fused, and finally the candidate regions are accurately classified. The instrument positioning method not only improves the detection rate, but also increases the expandability of the system, and meets requirements of a substation for instrument positioning.

7 citations

Patent
06 Jun 2017
TL;DR: In this paper, a scab segmentation method in a crop disease blade image is presented. But the method is not suitable for field operation, and it cannot be applied to the mobile terminal such as a smart phone.
Abstract: The present invention provides a scab segmentation method in a crop disease blade image. The method comprises the following steps: obtaining a color image including crop disease blades from an image collection device, performing normalization processing, converting the color image to a CIEL*a*b* color space, setting the initial classification number as 2, employing the adaptive feature learning method to learn the scab in the image and the values of the initial classification color features R, G and B of the background blades, calculating the distance from each pixel point to a classification center, and performing data classification; and calculating the distance between a* mean values, and stopping the segmentation if the distance of the a* value obtains the maximum result. The method overcomes the problems that different crops and different scab segmentation results are unstable in the prior art, improves the segmentation precision and the adaptation of the scab segmentation algorithm, can be applied to the mobile terminal such as a smart phone and the like, is suitable for field operation, and widen the application range.

5 citations

Patent
08 Jan 2019
TL;DR: Wang et al. as discussed by the authors presented a rice lesion detection method and system based on deep learning, belonging to the image processing field, the method comprising: providing a photo sample set and a manual labeling sample set, and cutting the photo sample sets and the manual labeling samples set according to a proportion to form a second photo sample subset and a second manual labeling subset, and inputting thesecond photo subset and the second label sample set into the LinkNet network model, and obtaining the optimal model by training the Linknet network model based on the Pytorch deep learning framework; using
Abstract: The invention discloses a rice lesion detection method and system based on deep learning, belonging to the image processing field, the method comprising: providing a photo sample set and a manual labeling sample set, and cutting the photo sample set and the manual labeling sample set according to a proportion to form a second photo sample set and a second manual labeling sample set; inputting thesecond photo sample set and the second label sample set into the Linknet network model, and obtaining the optimal model by training the Linknet network model based on the Pytorch deep learning framework; using the optimal model to identify the rice lesion images needed to be detected at present, and calculating the proportion of rice lesion area and classifying the disease status. Through the Linknet network model of Pytorch deep learning framework, the generalization ability and field practicability of rice leaf lesion identification can be improved, and the utilization rate of information can be improved, which is conducive to the subsequent quantitative application of pesticides and reduce environmental pollution.

4 citations

Patent
17 Sep 2019
TL;DR: In this paper, a variable fruit tree pesticide application robot based on multiple sensors is presented, which consists of a self-propelled system, a laser sensor, a camera, an industrial personal computer and a single chip microcomputer.
Abstract: The invention discloses a precise variable fruit tree pesticide application robot based on multiple sensors. The robot comprises a self-propelled system, a laser sensor (10), a camera (11), an industrial personal computer and a single chip microcomputer (9); the self-propelled system is used for installing equipment and driving the whole robot to move forward; the laser sensor (10) is used for detecting whether or not a target exists, sending a signal to the industrial personal computer after the target is detected and collecting and transmitting distance information to the industrial personal computer; the camera (11) is used for collecting images of the target and transmitting the images to the industrial personal computer for processing; the industrial personal computer is used for receiving the target signal of the laser sensor (10), judging the disease level according to the collected image information, obtaining the total canopy volume of the target according to the distance information of the target to determine the pesticide spraying amount and transmitting the pesticide spraying amount information to the single chip microcomputer (9); the single chip microcomputer (9) is used for receiving the pesticide spraying amount information, controlling a pesticide supply system to supply pesticide to a variable spraying system and controlling the variable spraying system to apply the pesticide. The robot obtains the accurate pesticide spraying amount by accurately detecting the volume of the target and identifying the degree of diseases and pests.

3 citations