Unmanned Aerial Vehicle for Agriculture Surveillance

*Alphanos Mahachi,Trymore Aloni and Lucious Mashevedze*

#### **Abstract**

The design of a fixed-wing Vertical Take-Off and Landing (VTOL) Unmanned Aerial Vehicle (UAV) for agriculture pest monitoring is the subject of this thesis. It is primarily concerned with the sugarcane pest problem in the KZN region. After seeing the impact of sugarcane on this region and South Africa as a whole, this design was created. The wing, fuselage, empennage, and tilt-rotor mechanism of the UAV are all designed to meet the mission requirements. The aerodynamics, performance, and stability of the UAV are next examined. The highest sustained turning performance was determined using the SEP chart. The UAV has a cargo capacity of 2 kg, a range of 96.7 m, a stall speed of 13.7 m/s, and a flight time of 1.48 hours. Because the UAV is a fixed-wing VTOL system, it can reach more geographically demanding regions and maneuver in windy conditions. The design was followed by the development of an IR thermography camera with 12 Megapixels and a 45 HFOV for the detection of pests. Following that, the tilt rotor mechanism was meticulously designed.

**Keywords:** agriculture pest surveillance, unmanned aerial vehicle, UAV design, UAV analysis, VTOL, multi-rotor

#### **1. Introduction**

#### **1.1 Problem definition**

It's tough to tell whether pests and diseases are affecting plants like sugarcane. In this study, the region of KwaZulu Natal (KZN) in South Africa was investigated. KZN is well recognized for its favorable geographical circumstances for sugarcane production, which include high annual rainfall, wetlands, low height, mild temperatures, and the presence of alluvial clay soil. Sugarcane production benefits not only KZN community but also South Africa as a whole; for example, 19.9 million tonnes of sugarcane produced in South Africa come from KZN, resulting in the bulk of sugar mills being based in this province [1]. **Figure 1** shows that 12 of South Africa's 14 sugar mills are located in KZN. Through the production of jobs and the sale of products, society benefits both socially and economically. For example, some people can work in mills for the processing step, while others work on farms. The sugar cane sector in KZN is expected to employ 79,000 people directly and 350,000 indirectly, accounting for about 2% of the South African population and a major portion of the entire

#### **Figure 1.**

*Sugarcane plantations regions in the Mpumalanga and KwaZulu-Natal provinces [2].*

agricultural labour. Sugarcane cultivation contributes over R8 billion to the South African economy, according to the South African Sugar Association (SASA) [2].

Given its agricultural and industrial investments, foreign exchange revenues, high employment, and connections with major suppliers, support industries, and customers, the South African sugar sector contributes significantly to the national economy. It's a diversified industry that combines sugar-cane farming with industrial factory production of raw and refined sugar, syrups and specialized sugars, as well as a variety of by-products.

Pests, on the other hand, have emerged as the most pressing worry when it comes to increasing sugarcane yields in KZN. This is due to the fact that not only does this location have the pest's host plant, but the conditions are also conducive to the pests' rapid reproduction, resulting in a wide variety of pest species. Stem borers, such as Eldana saccharina, the most common sugarcane pest in KZN, are one of these pest species types. According to the South African Sugarcane Research Institute (SASRI), there are four main pest categories that affect sugarcane production: leaf-eaters, leaf suckers, stem borers, and soil pests, but the Eldana is the industry's most serious pest, causing a total loss of over R1 billion per season if left unchecked [3]. Furthermore, with these losses in yield quality and quantity, action must be taken to reduce pest attacks in KZN (and the rest of the world), which can only be done by introducing precision farming, which is a farm management approach that uses information technology to ensure that crops and soil receive exactly what they need for optimum health and productivity [4]. When it comes to pest control, information technology can be used to monitor pests, spray pests, or do both to ensure that crops are healthy and generate a good yield. Monitoring pests before taking action has several advantages: it allows the farmer to implement timely interventions that ensure optimal yields at the end of the season, it reduces the amount of pesticide used because only the affected area can be sprayed before the pests spread throughout the farm, resulting in less wastage of resources and a reduction in the losses that can be incurred by the farmer due to the purchase of excess pesticides. The majority of pest monitoring in the investigated region is done by workers as they go about their everyday

operations, and the difficulty with this method is that by the time an infestation is discovered, a lot of harm has already been done. Another issue is that it is labourintensive and time-consuming.

#### **1.2 Current solutions to the problem**

The application of pesticides without understanding of the pest they are dealing with is a prevalent approach used by farmers in KZN to tackle pest problems. They frequently use a calendar to schedule the spaying practices, which entails researching the pests' life cycles and spraying pesticides when they are most destructive to the plants. This strategy has a number of drawbacks, including the farmer's potential to over- or under-apply pesticides due to a lack of understanding about the pest population in the affected area. Overuse of pesticides can result in losses if too much money is spent on pesticide purchases. Another issue is that the farmer may apply the pesticide too early or too late, because the calendar and analyzing the pest's life cycle cannot always be depended on because there are other aspects to consider, such as meteorological circumstances. Furthermore, if pesticides are applied too late, the affected area may lose most or all of its crops, as in the case of an armyworm infestation.

Furthermore, the problem is being treated by employing human labour to monitor the farm, which is roaming around the farm looking for infested plant areas. This procedure is time-consuming to the point that employees may arrive late at the afflicted area, causing many crops to be harmed. When Eldana borers attack 5-monthold sugarcane plants, it is difficult for workers to notice the frass since the farms are dense at this point, which means utilizing human labour for pest scouting will take longer and may be unsafe because sugarcane can harbor harmful animals such as snakes.

However, by employing more methodical ways, not only is the problem of sudden insect invasion minimized, but it is also possible to detect pests early. As a result, a gadget to monitor the pest and keep the farm informed is required before the infestation spreads throughout the entire farm. To investigate the entire farm, this equipment must be able to go from one spot to another, either by ground or by air. The device must also be able to take photographs in order to visualize and identify pests even inside stems, to store detection data and the GPS location of the infested area, and to be serviceable, which refers to the ease and speed with which corrective and preventive maintenance can be performed on the system.

#### **2. Literature survey**

#### **2.1 Pest detection methods**

These are the techniques for distinguishing the pest from the crops and collecting data. This can be accomplished by taking photos, listening for pest sounds, and inspecting the crops for pests. The sub-sections that follow will go through some of the present pest detection methods.

#### *2.1.1 Acoustic sensor*

The noise level of the pest is monitored by an acoustic device sensor, which alerts the farmer of the precise location where the infestation is occurring whenever the

noise level above the threshold [5]. The acoustic sensors node is connected to the base station, and each sensor will send the noise levels to the base station whenever the noise level exceeds a predetermined threshold. The red palm weevil has been detected using this technology in palm tree plantations [6]. The sensors, coupled with communication modules including a transceiver, are affixed to a tree during the detecting phase and latched to the network of neighboring access points. The optical fiber acoustic sensor is the most commonly utilized acoustic sensor in the detection of red palm weevil.

#### *2.1.2 Imaging sensors*

To identify pests, these image sensors employ optics and electronics. The hyperspectral camera, multispectral camera, RGB cameras, and thermal camera are among the imaging techniques used for surveillance, particularly in precision agriculture and military applications, as well as aerial mapping. The hyperspectral remote sensor is concerned with extracting information from objects or scenes on the Earth's surface using light collected by airborne or spaceborne sensors. It has a larger bandpass, which can approach 2000 at times. It's utilized in precision agriculture to discern apart plant species with identical spectral fingerprints, determine plant biochemical composition, and calculate chemical characteristics. Food safety, pharmaceutical process monitoring and quality control, biomedical, industrial, biometric, and forensic applications are all examples of lab-scale applications [7]. **Table 1** lists some of the existing hyperspectral cameras as well as their specifications. The thermal camera is another imaging sensor that may be used to create a heat zone image using infrared radiation with a wavelength of 1400 nm. It is used in agriculture to monitor water stress and irrigation uniformity in crops, as well as to calculate vegetation indices. Food preparation, safety and fire inspections, plastic molding, asphalt, maritime and screen printing, measuring ink and dryer temperature, and diesel and fleet maintenance are some of the sectors that employ them. The Noyafa NF-521 Thermal Imaging Camera, for example, has a temperature range of 10 to 400°C, a basic accuracy of 2%, and an adjustable emissivity of 0.1 to 0.99, with a measurement resolution of 0.1°C. The FLIR Scout TK Compact thermal monocular camera, which is mostly used in wildlife when locating wild creatures and has a temperature range of 40 to 60 degrees Celsius and can catch things up to 90 meters away, is the second option. The multispectral cameras are the next in line, with five bandpass interference filters: red, green, near-infrared, blue, and red edge [8]. It is currently used in agriculture to monitor crop diseases and pests, evaluate the status of the vegetation, and measure nutrient deficit.


**Table 1.** *Parameters of hyperspectral sensors.*
