Review Article

Applications of Smartphone-Based Sensors in Agriculture: A Systematic Review of Research

Table 3

Comparison of using smartphone-based sensors and traditional tools/methods in agriculture.

Application name Function Traditional tools/methods Advantages of smartphone-based sensors

Farming

Disease detection and diagnosis
n/a [13] Users snap photos of diseased leaves. Pictures are processed and sent to labs for further diagnosis (a) Expert diagnosis with the naked eye. (b) Lab diagnosis techniques such as microscopy, culturing, biochemical tests, and ELISA(a) Farmers with no expertise in disease identification can perform the diagnosis. (b) Farmers receive fast diagnosis compared to sending samples of diseases to labs
Magri [25] Users report information about pest diseases and receive advice from Magri regarding the pests Consult with pest experts Farmers can receive localized pest information

Fertilizer calculator
Baikhao [8, 16], BaikhaoNK [17] Using pictures of rice leaves, the applications analyze chlorophyll contents and the amount of nitrogen fertilizer needed for the rice field (a) Visual inspection using Leaf Color Chart (LCC) (b) SPAD analyzer (a) They can be more accurate and more reliable than visual inspection.a (b) They are cheaper than pricey SPAD analyzers

Soil study
n/a [14] Munsell soil color is determined using image sensors in order to suggest the structure and components of soil (a) Visual inspection using Munsell soil color chart. (b) Traditional instruments, for example, colorimeters and spectrometers (a) It is not subject to accuracy problemsb as in visual inspection. (b) Using smartphones as proximal sensors is simpler, cheaper, faster, less laborious, and more flexible than complex and costly instruments
SIFSS [15] SIFSS gives users access to the full soil dataset from the Soil Survey of Scotland Requesting information manually from the Soil Survey of Scotland It is a simpler and faster way to access information
SOCiT [15] SOCiT predicts topsoil organic matter and carbon content from pictures taken against a color correction card and other environmental characteristics, for example, elevation, climate, and geology Wet and dry combustion methods (which are analytically accurate but expensive) Soil analysis using SOCiT is quicker and more cost-effective

Water study
iDee [15] iDee provides a venue for river condition reporting and viewing for communities around River Dee catchment in Scotland Manually exchanging information within the communities via traditional communication channels, for example, surveys and forms It is a simpler and faster way to access information

Crop water needs estimation
n/a [8, 16] The smartphone image sensor is used to monitor the brightness of the light radiating on plants Lux meter Smartphones nowadays are highly available and ubiquitous
PocketLAI [9] PocketLAI determines Leaf Area Index (LAI), which is a key factor to calculate crop water requirements (a) Direct methods measuring LAI directly on leaves gathered from plantsc or by image processing of leaf images. (b) Indirect methodsd estimating LAI from gap fractione (a) PocketLAI is not destructive to the plants because it uses an indirect method. (b) PocketLAI is cheaper and more portable than usually expensive dedicated devices
RaGPS [26] RaGPS calculates solar radiation and equivalent evaporation from solar position using GPS in smartphones Obtaining such data from nearby agricultural weather station It can estimate crop water needs where there is no weather station nearby

Crop produce readiness analysis
n/a [8, 16] Smartphone image sensors serve as a two-dimensional sensor measuring ripeness level of bananas (a) Manual estimation with the naked eye. (b) Quantitative estimation methods, including analyses for mechanical, chemical, and aromatic components, some of which are destructive to the fruit itself [34] It is a nondestructive method, which is ideal for preserving the actual fruits

Farm management

Water management
n/a [18] The application estimates the water level, surface velocity, and discharge rate by analyzing a short video of the water flow between two control points with a known distance To obtain such water flow data, it traditionally requires expensive measurement stations, resulting in sparse and low quality data used for water management decision making (a) Smartphones facilitate the measurement tasks to be done much easier at lower costs (very low initial costs and no maintenance cost). (b) Users can take water flow measurement anywhere. (c) Less time is required to take measurements and data collection is easy via SMS

HR management and farm activity journaling
iFarm [27, 28] iFarm allows farmers in the field to record field data and communicate with farmers in the office that can form a work plan and communicate back to field farmers immediately as a location-aware task list on their smartphones Writing field data on paper and later manually entering those data into a computer at the office (a) Traditional field data collection approaches often result in nonintegrated data, making it difficult to extract valuable information from data. (b) Farmers can record field data and manage farm resources right at the field. (c) GPS makes a location-specific data collection and a location-aware task list creation possible
n/a [29] Using data from accelerometer, GPS, and microphone sensors, the application automatically detects farm workers’ activities, for example, harvesting, bed making, standstill, and walking Various kinds of dedicated body wearable sensors have been developed to detect human activitiesf Compared with conventional body wearable sensors, smartphones are more integrated, hence reducing users’ rejection due to the poor aesthetic value and intrusiveness of the traditional devices

Vehicle monitoring
SafeDriving [30] Based on data from accelerometer, gyroscope, and GPS, SafeDriving detects a tractor rollover and reports an emergency event Roll Over Protective Structures (ROPS) are required by the US Occupational Safety and Health Administration (OSHA) to be installed on all agricultural tractors to protect operators from rollovers [30] Although ROPS can reduce the severity of the rollover accidents, they do not provide an emergency reporting to rescue teams as in the smartphone application

Agricultural land management
MapIT [19] MapIT is a crowdsourcing application for collecting geographic information of small objects and small agricultural areas. A user takes a photo of an area and draws an outline of the area on a smartphone screen To map large objects, it usually requires a user to physically traverse the geographic object with a GPS device. Some tools use satellite images as baseline data With the traditional methods, mapping small objects is a time-consuming task since GPS is usually unreliable at this level of detail, which requires considerable effort to revise the outline data. MapIT is a camera-based application that is accurate, easy-to-use, and suitable for relatively small objects

Information system

Information localization
n/a [31] Farmers can access agricultural information that is relevant to their location via mobile and web technology Farmers can access information via mobile and web technology but have to manually sort through information that are not relevant to them (e.g., language not in the local dialect, disease outbreak in distant area) Location information from farmers’ mobile phone allows only relevant information to be shown to farmers and reduces information overload

Pest and disease information
SMILEX [32] Field worker’s fire blight reports from the field are shown on the map for better fire blight management Alerts and warnings have been distributed over the internet since early 1990s, but the use of modern technology (e.g., GPS) has been limited Smartphone enables real-time reports from the field with location information using built-in GPS. This allows better management of events and better information dissemination
VillageTree [20] Users can send images along with locations of pest incidents to be processed and alert other relevant users Pest report, inspection, and warning have traditionally been done manually, which is slow and tedious Pest images and locations from smartphone allow the system to detect pest and warn other users who might be affected

Extension service

Pest and disease information by experts
Agricultural Advisory System (AAS) [21] Farmers can call this call center to seek personalized solution for their farming problems. Images taken from their camera and location for their phones can provide information for the call center about the problem This paper targets farmers in rural areas where there are not enough experts to provide solutions. Therefore, traditional method is the insufficient expert field visits Smartphone allows farmers to better communicate their problems to expert remotely and receive prompt and personalized solution
m-Sahayak [22] Farmers can take photos or videos of the plants and use audio to record their query using their local spoken language Telephone services for agricultural solution exist for the target users of this paper, but communication failure occurs due to farmers’ inability to describe their disease Smartphone sensors (microphones and cameras) assist farmers in communicating their problems
mKRISHI [24] Farmers can use audio-visual facilities on a mobile phone to articulate their queries to experts with minimal use of textg Traditional tools and methods are similar to those of m-Sahayak [22] Smartphone sensors (microphones and cameras) assist farmers in communicating their problems

Tools for extension worker
GeoFoto [23] Field workers use this application in land plot identification Field workers manage multiple specialized devices and manually transfer data between devices Smartphone application provides a single device to do the same task as multiple devices, which makes management and data transfer easier and faster

aLCC is cheap, but the chart is likely to deteriorate over time, causing the reading of the rice leaf color level to be incorrect.
bThis is due to “illumination conditions, sample characteristics, and observer’s sensitivities” [14].
cAn example of dedicated instruments is Li-3100C.
dThese include systems using direct sunlight such as TRAC and DEMON, different models of ceptometers such as the SUNSCAN or the AccuPAR, and sensors such as the LAI-2000 or LAI-2200 Plant Canopy Analyzers.
eA viewable proportion of sky from beneath the canopy (or that of soil from top of the canopy).
fHowever, those dedicated sensors still have been targeting urban applications, for example, using body worn sensors to measure physical activity for motivating more workout, or for fall detection and prevention. The work in [29] aims at utilizing smartphone-based sensors to detect human activities in the context of rural farming.
gWhile this paper discusses the use of external sensors to create contextual data for experts to provide more informed solution, this review focuses on the part of paper that utilizes internal sensors.