CN103377189A

CN103377189A - Method for collecting ground reference information of grain on basis of intelligent mobile terminals

Info

Publication number: CN103377189A
Application number: CN2012101054565A
Authority: CN
Inventors: 伍之昂; 毛波; 曹杰; 方昌健; 王嵩
Original assignee: NANJING CLIMBING INFORMATION TECHNOLOGY Co Ltd; Nanjing University of Finance and Economics
Current assignee: NANJING CLIMBING INFORMATION TECHNOLOGY Co Ltd; Nanjing University of Finance and Economics
Priority date: 2012-04-12
Filing date: 2012-04-12
Publication date: 2013-10-30

Abstract

The invention discloses a method for collecting ground reference information of grain on the basis of intelligent mobile terminals. The method is characterized in that the ground reference information is collected during grain production on the basis of the existing universal intelligent mobile phone terminals, and the method is used for analyzing remote sensing data. The method has the advantages that a data mining technology is utilized, the correctness of acquisition results is evaluated by the aid of systematic analysis for user feedback results, a corresponding user information collecting database is created, and contents which can be skillfully collected by users can be described, so that corresponding tasks can be recommended, and the accuracy of the acquired information is improved.

Description

Grain ground reference information collection method based on intelligent mobile terminal

一、技术领域 1. Technical field

本发明基于目前广泛使用的移动手持设备操作系统iOS或Android开发环境，充分利用目前iphone、htc等3G智能手机中强大的嵌入式传感器等功能，包括高清摄像头，高精度GPS定位系统，红外温度测量，感光传感器、加速度传感器等一系列嵌入式设备，动态采集有关粮食地面参考信息，并通过无线通信技术实时发送到远程服务器。 The present invention is based on iOS or Android development environment, which is currently widely used mobile handheld device operating system, and makes full use of functions such as powerful embedded sensors in 3G smart phones such as iphone and htc, including high-definition camera, high-precision GPS positioning system, and infrared temperature measurement. , a series of embedded devices such as photosensitive sensors and acceleration sensors, dynamically collect relevant grain ground reference information, and send them to remote servers in real time through wireless communication technology. the

二、背景技术 2. Background technology

对于大规模粮食种植这种开放式自然环境，环境因素包括温度、湿度、光照、海拔、土壤参数(土壤类型、土壤水分、土壤PH值等)等对作物生长周期的影响不容忽视。而且近些年，随着移动通信技术、遥感技术、计算机技术、嵌入式传感器技术等的迅猛发展，如何将这些技术融合在一起，丰富技术的应用范围，是广大科技工作者们亟待研究的重大战略课题之一。基于此，在本发明中我们利用智能手机中的GPS定位功能，自动获取当前粮食种植的海拔高度和经纬度等空间信息，并匹配卫星遥感监测，进行实时矫正；同时采集其他物理参数，并对其误差进行分类、校正，从而提高信息收集的准确度，然后与其他相关信息一起融合提交至服务器终端。 For the open natural environment of large-scale grain planting, the influence of environmental factors including temperature, humidity, light, altitude, soil parameters (soil type, soil moisture, soil pH value, etc.) on the crop growth cycle cannot be ignored. And in recent years, with the rapid development of mobile communication technology, remote sensing technology, computer technology, embedded sensor technology, etc., how to integrate these technologies together to enrich the application range of technology is a major issue that the majority of scientific and technological workers need to study urgently. One of the strategic issues. Based on this, in the present invention, we use the GPS positioning function in the smart phone to automatically obtain the spatial information such as the altitude and latitude and longitude of the current grain planting, and match the satellite remote sensing monitoring to perform real-time correction; Errors are classified and corrected to improve the accuracy of information collection, and then fused with other relevant information and submitted to the server terminal. the

三、发明内容 3. Contents of the invention

该发明通过匹配用户的设备使用能力以及参照信息收集任务的复杂度，自动将适合的任务推荐给用户，同时考虑到用户的通勤成本、时间成本、投入成本等综合要素，从而降低信息使用的成本。 The invention automatically recommends suitable tasks to the user by matching the user's equipment usage ability and referring to the complexity of the information collection task. At the same time, it takes into account the comprehensive factors such as the user's commuting cost, time cost, and input cost, thereby reducing the cost of information use. . the

基于通用传感器的粮田地面信息收集方法，该方法基于智能移动终端中的有关内置传感器，测量粮食种植中的环境参数，并通过软件实现的方法进行误差校正，从而提高信息收集的准确度。 A general sensor-based grain field ground information collection method, which is based on the relevant built-in sensors in the smart mobile terminal, measures environmental parameters in grain planting, and performs error correction through software implementation methods, thereby improving the accuracy of information collection. the

信息实时上传匹配联动体系。通过卫星运行轨道的分析，使得地面信息收集时间与遥感数据获取时间基本同步，从而使得所获得的地面参照信息根据价值和容量与卫星遥感数据进行自动匹配，构建联动体系。同时，可将获取的地面信息与卫星数据及时传输到控制端，以实现对粮食种植的实时监控。 Information is uploaded in real time to match the linkage system. Through the analysis of the satellite orbit, the ground information collection time is basically synchronized with the remote sensing data acquisition time, so that the obtained ground reference information is automatically matched with the satellite remote sensing data according to the value and capacity, and a linkage system is constructed. At the same time, the obtained ground information and satellite data can be transmitted to the control terminal in time to realize real-time monitoring of grain planting. the

本发明实施方式提供一种基于协作过滤(CF)的用户任务推荐算法，其具体内容如下： Embodiments of the present invention provide a user task recommendation algorithm based on collaborative filtering (CF), the specific content of which is as follows:

协作过滤又称为合作过滤或社会过滤，它基于如下假设：如果一些用户对一些项目的评分较相似，则他们对其它项目的评分也比较相似。所以，其实现的基本思想是系统采用某种技术找到目标用户的若干邻居(与目标用户有相似兴趣的用户)，然后根据最近邻居对目标项目的评分，产生推荐。协作过滤分析的是“用户一用户”的关系，根据用户兴趣的相似性来推荐资源。 Collaborative filtering, also known as collaborative filtering or social filtering, is based on the assumption that if some users rate some items similarly, they rate other items similarly. Therefore, the basic idea of its realization is that the system uses a certain technology to find several neighbors of the target user (users with similar interests to the target user), and then generates recommendations based on the ratings of the nearest neighbors on the target item. Collaborative filtering analyzes the "user-user" relationship, and recommends resources according to the similarity of user interests. the

基于用户的协作过滤算法中度量用户相似性的方法主要有以下两种：余弦相似性和相关相似性，计算公式分别如式(1)和式(2)所示： In the user-based collaborative filtering algorithm, there are mainly two methods for measuring user similarity: cosine similarity and correlation similarity. The calculation formulas are shown in formula (1) and formula (2):

$Sim Sim ((u u,, n no)) = = \frac{{Σ Σ}_{m m = = 11}^{N N} {R R}_{u u,, m m} * * {R R}_{n no,, m m},,}{\sqrt{{Σ Σ}_{m m = = 11}^{N N} {R R}_{u u,, m m} * * {Σ Σ}_{m m = = 11}^{N N} {R R}_{n no,, m m}}} - - - - - - ((11))$

$Sim Sim ((u u,, n no)) = = \frac{\underset{c c &Element; &Element; I I}{Σ Σ} (({R R}_{u u,, c c} - - {\overset{&OverBar; &OverBar;}{R R}}_{u u})) * * (({R R}_{n no,, c c} - - {\overset{&OverBar; &OverBar;}{R R}}_{n no}))}{\sqrt{\underset{c c &Element; &Element; I I}{Σ Σ} {(({R R}_{u u,, c c} - - {\overset{&OverBar; &OverBar;}{R R}}_{u u}))}^{22} * * \underset{c c &Element; &Element; I I}{Σ Σ} {(({R R}_{n no,, c c} - - {\overset{&OverBar; &OverBar;}{R R}}_{n no}))}^{22}}} - - - - - - ((22))$

式中：N——资源的数目； In the formula: N - the number of resources;

R_u，m和R_n，m——用户u和n表示对资源项目m的评分； R _{u, m} and R _{n, m} - users u and n represent the ratings of resource item m;

I——用户u和n共同评分过的项目集； I——the item set that users u and n have rated together;

R_u，c和R_n，c——用户u和n对I中某一项资源的评分； R _{u, c} and R _{n, c} ——the rating of a resource in I by users u and n;

Figure 871251DEST_PATH_GSB00000919948800021

和

Figure 266461DEST_PATH_GSB00000919948800022

——用户u和n对资源的平均评分。

and

- the average rating of the resource by users u and n.

信息推荐时，根据最近邻居对项目u的评分产生推荐，采用以下公式(3)计算目标用户u对未评分项目i的预测评分： When recommending information, recommendations are generated based on the ratings of the nearest neighbors for item u, and the following formula (3) is used to calculate the predicted rating of target user u for unrated item i:

$Pred Pred ((u u,, i i)) = = {\overset{&OverBar; &OverBar;}{R R}}_{u u} + + \frac{{Σ Σ}_{a a = = 11}^{b b} (({R R}_{a a,, i i} - - {\overset{&OverBar; &OverBar;}{R R}}_{a a})) * * Sim Sim ((u u,, a a))}{{Σ Σ}_{a a = = 11}^{b b} Sim Sim ((u u,, a a))} - - - - - - ((33))$

式中：

Figure 840979DEST_PATH_GSB00000919948800024

——用户u和其最近邻a对资源项目的平均评分，R_a，i——最近邻a对项目i的评分，b——最近邻的个数，Sim(u，a)——用户u和a的相似度。 In the formula:

——the average rating of user u and its nearest neighbor a on the resource item, R _a,i ——the rating of the nearest neighbor a on item i, b——the number of nearest neighbors, Sim(u, a)——user u similarity to a.

用户u对任务i的适应程度为：用户对所有任务的平均适应度

Figure 108012DEST_PATH_GSB00000919948800025

加上与其最相似的用户群

Figure 306912DEST_PATH_GSB00000919948800026

对任务i的适应程度与其平均适应程度之差。通过对该公式，便可以确定将那些任务推荐给相应的用户。 The fitness of user u to task i is: the average fitness of user to all tasks

plus the user group most similar to it

The difference between the fitness for task i and its average fitness. Through the formula, it can be determined to recommend those tasks to corresponding users.

本发明提出了一种基于用户协作过滤信息推荐模型，该模型在对资源项目分类的基础上，将用户对项目评分转换为用户对资源项目类别的评分，对数据预处理之后，在对用户聚类，最后在目标用户所在的簇中寻找其最近邻居并产生推荐。同时，可以通过加权过滤数据预处理以提高算法的扩展性和推荐质量。 The present invention proposes an information recommendation model based on user collaborative filtering. On the basis of classifying resource items, the model converts user ratings on items into ratings on resource item categories. class, and finally find its nearest neighbors in the cluster where the target user is located and generate recommendations. At the same time, weighted filtering data preprocessing can be used to improve the scalability and recommendation quality of the algorithm. the

四、附图说明 4. Description of drawings

图1远程客户端功能架构 Figure 1 Remote Client Functional Architecture

图2粮食种植地面信息采集流程图 Figure 2 Flowchart of grain planting ground information collection

五、具体实施方式 5. Specific implementation

本发明实施方式提供GPS定位信息采集和农田一般信息采集相结合的方法，并通过卫星遥感技术，匹配所采集信息的准确度，并加以校正和数据融合，最后通过无线传输发送到服务器终端，具体步骤如下： The embodiment of the present invention provides a method of combining GPS positioning information collection and farmland general information collection, and through satellite remote sensing technology, the accuracy of the collected information is matched, corrected and data fused, and finally sent to the server terminal through wireless transmission, specifically Proceed as follows:

步骤1启动系统，进行自检、初始化程序； Step 1 Start the system, carry out self-inspection and initialization procedures;

步骤2启动GPS和智能农田信息采集传感器，GPS采集有关的海拔、经纬度信息，农田采集系统询问是否开启高清摄像头，如是，则采集图像信息，否则，采集温、湿度，土壤、光照等条件信息并保存； Step 2 Start GPS and smart farmland information collection sensors, GPS collects relevant altitude, longitude and latitude information, and the farmland collection system asks whether to turn on the high-definition camera, if yes, collects image information, otherwise, collects condition information such as temperature, humidity, soil, and light save;

步骤3通过GPS采集的地面信息和卫星遥感所提供的数据进行比较，如有误差，则进行修正GPS信息参数，然后保存至存储卡中； Step 3 compares the ground information collected by GPS with the data provided by satellite remote sensing, if there is an error, correct the GPS information parameters, and then save it to the memory card;

步骤4将保存的数据汇总之后通过无线通信网络发送到远程服务器终端，并动态实时显示； Step 4 sends the saved data to the remote server terminal through the wireless communication network after summarizing, and displays it dynamically in real time;

步骤5此时，所采集到的农田动态种植信息均可体现用户的资源需求，按照某种原则，如通过加权平均过滤数据等，全局画出动态实时曲线，并返回。 Step 5 At this point, the collected dynamic planting information of the farmland can reflect the user's resource requirements. According to a certain principle, such as filtering data through weighted average, the dynamic real-time curve is drawn globally and returned. the

Claims

1. A method for collecting ground reference information for grain based on an intelligent mobile terminal. The characteristics of the implementation process of the method are as follows: use the relevant sensors of a general intelligent mobile terminal to design a method for collecting ground reference information for grain; evaluate and mine the results uploaded by users , analyze the user's characteristics in the collection process and establish a corresponding database; based on the user's ground reference information collection characteristics, perform task recommendation to achieve efficient and accurate data collection.

2. According to the grain ground reference information collection method based on intelligent mobile terminal described in claim 1, its data collection is characterized in that: make full use of the sensor of general mobile intelligent terminal, design corresponding ground reference information collection method, to reduce data The cost of collection. Specifically, it is first necessary to deeply analyze the sensors contained in existing mobile devices such as smartphones, such as cameras, thermometers, GPS locators, gyroscopes, etc. Make full use of the above smart sensors through software to obtain ground reference data that meets the accuracy requirements without increasing hardware costs.

3. According to the evaluation and mining method of user upload results described in claim 1, it is characterized in that by using a collaborative filtering algorithm, through clustering of users, similar users are explored and their uploaded information is cross-validated, thereby Ensure the accuracy of data collection.

4. According to the user information collection database described in claim 1, it is characterized in that the database is automatically created and updated through a collaborative filtering algorithm, that is, by comparing the acquired user information and the ground reference data uploaded, automatically Establish a historical database uploaded by users, and based on this, carry out information collection content recommendation, obtain feedback information, and update the database.