CPC G01N 33/025 (2013.01) [A01G 25/167 (2013.01); G06V 10/751 (2022.01); G06V 20/68 (2022.01)] | 3 Claims |
1. A non-destructive fruit defect detection system based on neural networks, comprising: a user terminal, a storage module, a standard formulation module, a preliminary identification module, a data acquisition module, a non-destructive detection module, a quality judgment module and a server;
wherein when selling fruits, an identification code is attached on each fruit, the user terminal is configured to input the identification code of each fruit, take a real-time photo of each fruit with the identification code and send the real-time photo of each fruit with the identification code to the server, the server is configured to send the identification code of each fruit to the storage module and send the real-time photo of each fruit to the preliminary identification module, and the storage module is configured to send, based on the identification code of each fruit, standard characteristic data of each fruit to the standard formulation module, standard growth data of each fruit to the non-destructive detection module and a standard fruit photo of each fruit to the preliminary identification module;
wherein the standard formulation module is configured to formulate monitoring standards for different batches and varieties of the fruits to obtain standard detection parameters for the different batches and varieties of the fruits, and send the standard detection parameters to the non-destructive detection module by the server, the preliminary identification module is configured to preliminarily identify external conditions of the different batches and varieties of the fruits to obtain a non-destructive detection signal or a fruit abnormal signal, and send the non-destructive detection signal or the fruit abnormal signal to the server, the server is configured to send the fruit abnormal signal to the user terminal in response to the fruit abnormal signal, and send the non-destructive detection signal to the data acquisition module in response to the non-destructive detection signal;
wherein before picking the fruits, the data acquisition module is configured to acquire real-time growth data and real-time characteristic data of the different batches and varieties of the fruits based on the non-destructive detection signal, and send the real-time growth data and the real-time characteristic data to the non-destructive detection module by the server, the non-destructive detection module is configured to non-destructively detect the different batches and varieties of the fruits to obtain a fruit abnormal signal or growth deviation values of the different batches and varieties of the fruits;
wherein the quality judgment module is configured to judge quality of the different batches and varieties of the fruits to obtain quality grades of the different batches and varieties of the fruits and send the quality grades to the user terminal by the server, and the user terminal is configured to obtain quality situations of the different batches and varieties of the fruits based on the quality grades;
wherein the standard characteristic data comprises: standard fruit weight values, standard fruit length values, standard fruit width values, and standard fruit height values of the different varieties of the fruits;
wherein the standard growth data comprises: standard environmental temperatures, standard environmental humidity, standard soil humidity, standard soil potential of hydrogen (pH), and standard light duration of the different varieties of the fruits;
wherein each identification code comprises a name, an origin place and a batch number of the corresponding fruit;
wherein the real-time growth data comprises daily environmental temperatures, daily environmental humidity, daily soil humidity, daily soil pH, and daily light duration of the different batches and varieties of the fruits in a last month before picking the fruits;
wherein the real-time characteristic data comprises daily weight values, daily length values, daily width values and daily height values of the different batches and varieties of the fruits when picking the fruits;
wherein a formulation process of the standard formulation module comprises following steps:
selecting corresponding batches and corresponding varieties of the fruits based on the standard characteristic data, and classifying the fruits with same batches and same varieties into fruit samples;
detecting each fruit sample to obtain sample weight values, sample length values, sample width values and sample height values of the fruit samples;
iterating through and comparing the sample length values of the fruit samples to determine a sample weight upper limit value and a sample weight lower limit value of the fruit samples, wherein the sample weight upper limit value and the sample weight lower limit value together form a weight interval for the fruit samples;
obtaining a length interval, a width interval and a height interval of the fruit samples according to the above step, wherein the weight interval, the length interval, the width interval and the height interval are the standard detection parameters of the fruits with same batch and variety as the fruit samples; and
obtaining the standard detection parameters of the different batches and varieties of the fruits according to the above steps;
wherein a working process of the non-destructive detection module comprises following steps:
obtaining the daily weight values, the daily length values, the daily width values and the daily height values of the different batches and varieties of the fruits when picking the fruits, obtaining the standard detection parameters of the different batches and varieties of the fruits, and obtaining the weight intervals, the length intervals, the width intervals and the height intervals of the different batches and varieties of the fruits;
performing next steps when the daily weight value of the fruit is in the weight interval corresponding to the fruit, the daily length value of the fruit is in the length interval corresponding to the fruit, the daily width value of the fruit is in the width interval corresponding to the fruit and the daily height value of the fruit is in the height interval corresponding to the fruit; or generating the fruit abnormal signal when any one of below conditions occurs: the daily weight value of the fruit is not in the weight interval corresponding to the fruit, the daily length value of the fruit is not in the length interval corresponding to the fruit, the daily width value of the fruit is not in the width interval corresponding to the fruit and the daily height value of the fruit is not in the height interval corresponding to the fruit;
wherein the working process of the non-destructive detection module further comprises the next steps:
obtaining the daily environmental temperatures, the daily environmental humidity, the daily soil humidity, the daily soil pH, and the daily light duration of the different batches and varieties of the fruits in the last month before picking the fruits;
adding the daily environmental temperatures of each batch of the fruits to obtain a real-time environmental temperature of each batch and each variety of the fruits;
obtaining real-time environmental temperature humidity, real-time soil humidity, real-time soil pH and real-time light duration of each batch and each variety of the fruits accordingly;
obtaining the standard growth data comprising the standard environmental temperature, the standard environmental humidity, the standard soil humidity, the standard soil pH, and the standard light duration of each batch and each variety of the fruits;
calculating a deviation value denoted as WCui between the real-time environmental temperature and the standard environmental temperature of each batch and each variety of the fruits, wherein WCui represents the deviation value between the real-time environmental temperature and the standard environmental temperature of a u-th fruit in an i-th batch, u represents a serial number of the fruit, i represents a serial number of the batch corresponding to the fruit, u=1, 2, . . . , z, z is a positive integer, i=1, 2, . . . , x, and x is a positive integer;
calculating accordingly a humidity deviation value SCui, a soil humidity deviation value TSCui, a soil pH deviation value SJui and a light duration deviation value GTui of each batch and each variety of the fruits, wherein SCui represents the humidity deviation value of the u-th fruit in the i-th batch, TSCui represents the soil humidity deviation value of the u-th fruit in the i-th batch, SJui represents the soil pH deviation value of the u-th fruit in the i-th batch, and GTui represents the light duration deviation value of the u-th fruit in the i-th batch; and
calculating the growth deviation value SPui of each batch and each variety of the fruits through a formula: SPui=SCui×a1+TSCui×a2+SJui×a3+GTui× a4, where SPui represents the growth deviation value of the u-th fruit in the i-th batch, a1, a2, a3 and a4 represent weight coefficients which are fixed values, and a1+a2+a3+a4=1;
wherein the non-destructive detection module is configured to send the fruit abnormal signal or the growth deviation value of each batch and each variety of the fruits to the server, the server is configured to send the fruit abnormal signal to the user terminal in response to the fruit abnormal signal, the user terminal is configured to check the quality situations of the fruits after receiving the fruit abnormal signal, the server is configured to send the growth deviation value of each batch and each variety of the fruits to the quality judgment module in response to the growth deviation value of each batch and each variety of the fruits;
wherein the quality judgment module is configured to judge the quality of the different batches and varieties of the fruits, a judgment process comprises following steps:
obtaining the growth deviation value SPui of each batch and each variety of the fruits;
in response to the growth deviation value SPui being less than X1, determining the quality grade of the corresponding batch and variety of the fruits is excellent;
in response to the growth deviation value SPui being less than X2 but not less than X1, determining the quality grade of the corresponding batch and variety of the fruits is average; and
in response to the growth deviation value SPui being not less than X2, determining the quality grade of the corresponding batch and variety of the fruits is inferior;
wherein X1 and X2 represent growth deviation thresholds and X1 is less than X2; and
wherein the quality judgment module is configured to send the quality grade of each batch and each variety of the fruits to the server, the server is configured to send the quality grade of each batch and each variety of the fruits to the user terminal, and the user terminal is configured to obtain the quality condition of each batch and each variety of the fruits based on the quality grade.
|