A data analysis system for enterprise industrial information

Abstract

The present invention relates to the technical field of industrial informatization, and discloses a data analysis system for enterprise industrial information. A plurality of layers of ultra-thin conductive films are arranged in a layered manner above the pineapple planting area; an electric pulse release unit is used to emit pulsed current from the edge of the ultra-thin conductive film; a current signal capture unit is used to collect current signal data at set coordinates around each pineapple plant; a data preprocessing module generates corresponding current signal data sequences at every interval of a set time T starting from when the pineapple grows through the first layer of ultra-thin conductive film; a data processing module is used to input all the current signal data sequences into a yield prediction model for processing and then output the pineapple yield. The present invention can accurately obtain the current signals at the fixed-point positions around the pineapple plants on the ultra-thin conductive film, collect them once at every interval of the set time, and can accurately predict the growth and yield of the pineapple plants after data processing, effectively realizing the control of yield risks before the picking node.

Description

Technical Field

[0001] The present invention relates to the technical field of industrial informatization, and more specifically, it relates to a data analysis system for enterprise industrial information. Background Art

[0002] When crop R & D enterprises carry out pineapple cultivation, they adopt the form of uniform fertilizer spraying. However, the growth conditions of each pineapple plant are different, and it is impossible to accurately predict and analyze the pineapple yield, and it is impossible to control the yield risk before the harvest node.

[0003] And there is a current method of arranging multiple layers of ultra-thin conductive films above the pineapple growth area. After the pineapple grows through the film, corresponding holes are formed on the film. Then, an electric pulse release unit applies a current signal to one end of the ultra-thin conductive film. The pulse signal changes in amplitude according to a sine law. When encountering a hole, the attenuation degree of the electric pulse current is significantly higher than when there is no hole. The more holes there are in the same straight line, the more the attenuation degree. Then, the lateral position of the electric pulse is judged according to the amplitude, so as to calculate the hole area on the corresponding ultra-thin conductive film. However, at this time, the amplitude of the collected electric pulse signal is the current attenuation on the entire film. It is unreliable to accurately calculate the area of each hole, and it is impossible to form accurate plant growth condition data, with a large error, and it is even more impossible to be used as reference data for pineapple plant growth and yield prediction. Summary of the Invention

[0004] The purpose of the present invention is to provide a data analysis system for enterprise industrial information in order to solve the above problems.

[0005] The present invention provides a data analysis system for enterprise industrial information, including:

[0006] Several layers of ultra-thin conductive films, which are arranged in a layered manner above the pineapple planting area;

[0007] An electric pulse release unit, which is used to emit a pulsed current from the edge of the ultra-thin conductive film;

[0008] A current signal capture unit, which is used to collect current signal data at set coordinates around each pineapple plant;

[0009] A data preprocessing module, starting from when the pineapple grows through the first layer of ultra-thin conductive film, collects the current signal of the current signal capture unit once every set time T, and generates a corresponding current signal data sequence. Among them, the current signal data sequence generated by the t-th collection is represented as A t ={A t，1 ,…A t，N}, where the c-th sequence unit is represented as Where It represents the current signal value of the current signal capture unit at the i-th row and j-th column of the c-th layer of the ultra-thin conductive film collected at the t-th time;

[0010] A data processing module, which is used to input all current signal data sequences into a yield prediction model for processing and then output the pineapple yield.

[0011] As a further optimization scheme of the present invention, the yield prediction model includes a CNN layer, a first hidden layer, a splicing layer, an RNN layer and a classifier.

[0012] As a further optimization scheme of the present invention, when all current signal data sequences are input into the yield prediction model for processing, specifically, the sequence units of the current signal data sequence are input into the CNN layer to generate a current signal feature sequence It represents A t The current signal feature vector obtained by inputting the N-th sequence unit of into the CNN layer.

[0013] As a further optimization scheme of the present invention, when all current signal data sequences are input into the yield prediction model for processing, specifically, the current signal feature sequence is input into the first hidden layer for encoding to generate an encoded output sequence.

[0014] As a further optimization scheme of the present invention, the calculation formula of the first hidden layer is as follows:

[0015]

[0016] Among them, It represents the i-th sequence unit of the encoded output sequence, F i and F j respectively represent the i-th and j-th sequence units of the current signal feature sequence, W1 represents the weight vector of the first hidden layer, and N is the number of sequence units of the current signal feature sequence.

[0017] As a further optimization scheme of the present invention, when all current signal data sequences are input into the yield prediction model for processing, specifically, the row vectors in the encoded output sequence are spliced through the splicing layer to obtain an encoded feature vector Among them, the encoded output sequence generated by the current signal data sequence collected at the t-th time is Splice the sequence units in B t directly to obtain an encoded feature vector

[0018] As a further optimization scheme of the present invention, when all current signal data sequences are input into the yield prediction model for processing, specifically, all encoded feature vectors are input into the RNN layer, where denotes the encoded feature vector generated from the current signal data sequence generated by the last acquisition, and M is the total number of acquisitions.

[0019] As a further optimization scheme of the present invention, all current signal data sequences are input into the yield prediction model for processing. Specifically, the input at the t-th time step of the RNN layer unit The output at the last time step is input into the classifier, and the classifier outputs the pineapple yield.

[0020] The beneficial effects of the present invention are as follows: The present invention can accurately obtain the current signal at the fixed-point position around the pineapple plant on the ultra-thin conductive film, collect it once at a set time interval, and after processing the collected data, it can relatively accurately predict the growth and yield of the pineapple plant, effectively realizing the control of the yield risk before the picking node. Description of the Drawings

[0021] Figure 1 is the system block diagram of the present invention. Detailed Embodiments

[0022] Now, the subject matter described herein will be discussed with reference to exemplary embodiments. It should be understood that discussing these embodiments is only to enable those skilled in the art to better understand and thus implement the subject matter described herein. Without departing from the scope of protection of the content of this specification, changes can be made to the functions and arrangements of the elements discussed. Each example can omit, substitute, or add various processes or components as needed. Additionally, the features described relative to some examples can also be combined in other examples.

[0023] As Figure 1 shown, a data analysis system for enterprise industry information includes:

[0024] Several layers of ultra-thin conductive films, which are arranged in an upper and lower layered manner above the pineapple planting area;

[0025] In an embodiment of the present invention, several layers of ultra-thin conductive films are evenly distributed.

[0026] An electric pulse release unit for emitting pulsed current from the edge of the ultra-thin conductive film.

[0027] A current signal capture unit for collecting current signal data at set coordinates around each pineapple plant;

[0028] Specifically, N layers of ultra-thin conductive films are arranged above the pineapple plant planting area, and the surface area of the ultra-thin conductive film is divided into grids. The collection ends of the current signal capture unit are arranged at the grid point positions. Each pineapple plant grows in the inner area of the grid, and the current signal data at the grid points around each pineapple plant is captured by the corresponding current signal capture unit;

[0029] In one embodiment of the present invention, pulsed currents are emitted in both the horizontal and vertical directions for the coordinate positions of each pineapple plant.

[0030] The data preprocessing module starts from the moment when the pineapple growth penetrates the first ultra-thin conductive film, emits a pulsed current every set time T and collects the current signal. The current signal data sequence generated by the t-th collection is denoted as A t ={A t，1 , … A t，N}, where the c-th sequence unit is denoted as where represents the current signal value of the current signal capture unit at the i-th row and j-th column of the c-th ultra-thin conductive film collected at the t-th time;

[0031] The data processing module inputs the current signal data sequence into the yield prediction model. The yield prediction model includes a CNN layer, a first hidden layer, a splicing layer, an RNN layer, and a classifier. The sequence units of the current signal data sequence are input into the CNN layer to generate a current signal feature sequence represents the current signal feature vector obtained by inputting the N-th sequence unit of A t into the CNN layer;

[0032] The current signal feature sequence is input into the first hidden layer for encoding to generate an encoded output sequence;

[0033] The calculation formula of the first hidden layer is as follows:

[0034]

[0035] where, represents the i-th sequence unit of the encoded output sequence, F i and F j respectively represent the i-th and j-th sequence units of the current signal feature sequence, W1 represents the weight vector of the first hidden layer, and N is the number of sequence units of the current signal feature sequence;

[0036] The row vectors in the encoded output sequence are spliced through the splicing layer to obtain an encoded feature vector;

[0037] In one embodiment of the present invention, the encoded output sequence generated by the current signal data sequence generated by the t-th collection is The sequence units in B t are directly spliced to obtain an encoded feature vector

[0038] Input into the RNN layer, where It represents the encoded feature vector generated from the current signal data sequence generated by the last acquisition. M is the total number of acquisitions, and it is the input at the t-th time step of the RNN layer unit. The output at the last time step is input to the classifier, and the classifier outputs the pineapple yield.

[0039] In an embodiment of the present invention, the value range of the pineapple yield is [0, 100], with the unit of kg / square meter. The mean value of this value range is discretized into 101 point values, which respectively correspond to 101 classification labels of the classifier.

[0040] It can effectively predict the value range of the pineapple yield at the harvest time node, accurately predict and analyze the pineapple yield, improve the accuracy of the enterprise's yield prediction result before the harvest time node, enhance the enterprise's risk control of the industrial yield, and can carry out targeted processing in advance.

[0041] The mobile topdressing module is used to spray fertilizers specifically on areas with low predicted yields, or manually fertilize and water the corresponding plants, etc., and record the final yields of the corresponding plants, as well as data such as fertilizer and water spraying, etc., to provide planting improvement parameters for planting pineapples again.

[0042] The above describes this embodiment, but this embodiment is not limited to the above specific implementation manners. The above specific implementation manners are merely illustrative and not restrictive. Under the inspiration of this embodiment, those of ordinary skill in the art can also make many forms, all of which fall within the protection scope of this embodiment.

Claims

1. A data analysis system for enterprise industry information, characterized in that: include: Several layers of ultra-thin conductive film are laid out in layers above the pineapple planting area; an electric pulse releasing unit for emitting a pulse current from the edge of the ultra-thin conductive film; A current signal capture unit is used to collect current signal data at set coordinates around each pineapple plant; The data preprocessing module starts from the time when the pineapple grows through the first layer of ultra-thin conductive film. It collects the current signal of the current signal capture unit every set time T and generates a corresponding current signal data sequence. The current signal data sequence generated by the tth collection is represented by A t =A t，1 ,…A t，N }, where the cth sequence unit is represented as in represents the current signal value of the current signal capturing unit of the i-th row and j-th column of the c-th layer of ultra-thin conductive film acquired for the t-th time; A data processing module is used to input all current signal data sequences into the yield prediction model for processing and then output the pineapple yield; The yield prediction model includes CNN layer, first hidden layer, concatenation layer, RNN layer and classifier; All current signal data sequences are input into the yield prediction model for processing. Specifically, the sequence units of the current signal data sequence are input into the CNN layer to generate the current signal feature sequence. Indicates A t The current signal feature vector obtained by inputting the Nth sequence unit into the CNN layer.

2. The data analysis system for enterprise industry information according to claim 1, characterized in that: All current signal data sequences are input into the yield prediction model for processing. Specifically, the current signal feature sequence is input into the first hidden layer for encoding to generate an encoded output sequence.

3. The data analysis system for enterprise industry information according to claim 2, characterized in that: The calculation formula for the first hidden layer is as follows: in, represents the i-th sequence unit of the encoded output sequence, F i and F j They represent the i-th and j-th sequence units of the current signal feature sequence respectively, W1 represents the weight vector of the first hidden layer, and N is the number of sequence units of the current signal feature sequence.

4. The data analysis system for enterprise industry information according to claim 3, characterized in that: All current signal data sequences are input into the yield prediction model for processing. Specifically, the row vectors in the coded output sequence are spliced through the splicing layer to obtain the coded feature vector Among them, the coded output sequence generated by the current signal data sequence generated by the tth acquisition is To B t The sequence units in are directly concatenated to obtain the encoding feature vector 5. The data analysis system for enterprise industry information according to claim 4, characterized in that: All current signal data sequences are input into the yield prediction model for processing, specifically, all encoded feature vectors Input RNN layer, where It represents the coded feature vector generated by the current signal data sequence generated by the last acquisition, and M is the total number of acquisitions.

6. The data analysis system for enterprise industry information according to claim 5, characterized in that: All current signal data sequences are input into the yield prediction model for processing. Specifically, the t-th time step input of the RNN layer unit is The output at the last time step is input into the classifier, which outputs the pineapple yield.

Images