Cayenne pepper (Capsicum frutescens L.) is a horticultural plant originating from America. Cayenne pepper plants have high economic value and are suitable for cultivation in tropical areas such as Indonesia. According to agricultural statistical data, national cayenne pepper production reached 1,335,595 tons in 2018 with a growth rate from 2014-2018 of 15% per year (BPS and Director General of Horticulture). The productivity of cayenne pepper in Indonesia is still low, when compared to other countries, especially in the Asian region. The factors that because low productivity are the low use of technology to control whitefly pests and ineffective monitoring efforts that are still an obstacle faced by farmers, then poor crop management. Whitefly pest attacks can directly damage plants and act as vectors of plant diseases. Symptoms of whitefly attack can be seen directly on the curling and yellowing of the leaves, while the signs of attack are found in the whitefly population under the chili leaves. The whitefly also sucks the chili leaf tissue and transmits the yellow virus. Whitefly pests can inhibit the growth of chili plants to a critical point, which can have an impact on crop failure. 

Agricultural product predictions are currently used in terms of planning and decision making for food security. Currently, precision agriculture has been developed, namely agriculture that provides precision treatment at all nodes of the agribusiness chain for efficiency in the world of agriculture [1], According to Sari [2], that monitoring technology using drones is able to provide direct impacts, solutions and convenience in spatial analysis repeatedly, continuously and in a relatively wide area coverage. Various studies have been carried out on crop yield prediction which have been widely used in agriculture, especially using satellite imagery, both from fashionable imagery and Landsat imagery [3]. Like Holik and Bachtiar [4], about predicting rice yields using drones. However, the majority can be seen directly or by laboratory tests. So it is necessary to innovate to be able to know the estimated results of damage caused by whitefly attacks automatically to make it more effective and efficient.                

This research was carried out on ±800m² of cayenne pepper plantations owned by farmers located in Brejel Lor Hamlet, Pucuk Village, Dawarblandong District, Mojokerto Regency in July 2022. The stages of this research are work procedures for research starting from literature study to completion. by getting the research results.

Preliminary surveys were conducted to determine the condition of chili fields attacked by whitefly pests. Several land surveys were carried out to obtain land with symptoms of whitefly infestation which could be observed on a digital camera. Secondary data on land history were obtained from observations and interviews with farmers to determine the history of land use. Climatological data was obtained from the Meteorology, Climatology and Geophysics Agency (BMKG). The data is secondary data that is used as supporting data. Sampling of whitefly pests was carried out to identify the types of pests that attack and determine the bioecology of pests on plants.

Figure 1: Research Stages

Figure 2: Image Processing Steps [4]

Sampling for calculating the distribution pattern was carried out by purposive sampling on a land area of ​​±800 m² which was attacked by whitefly pest symptoms. sampling plots using a quadrant measuring 2 x 2 m². Sampling was done manually by looking at the symptoms of attack and signs of pests. Samples that have been taken from the field are then taken with a triple camera photo and then given a clear description of the supporting data from the literature.

Aerial image capture using drones. Based on the literature, good image capture is carried out when the weather is sunny with light intensities ranging from 20,000-60,000 lux [5]. In this study, image capture was carried out in the morning starting at 07.00 - 09.00 WIB. Once ready, the drone is flown at an altitude of 10 m, 20 m and 30 m for a full swath. The height of this drone is consistent with research by Zhang et al. on estimating tree diameter based on drone imagery; Su, [6] and Guo et al. where the optimal height for drones is 10-30 m. In this study, image processing produces the main object, namely the image of symptomatic leaves caused by whitefly pests. Image processing in this study was carried out using the multi-thresholding method. In this study, 3 segmentations were carried out to distinguish symptomatic leaves due to whitefly attacks, healthy leaves and weeds. Segmentation results are calculated based on the percentage of pixel segmentation. The algorithm design flowchart can be seen in the figure 2.

The cayenne pepper weighing is carried out after the process of taking drone images and visual observations has been completed. Weighing is done to find out the actual yield of cayenne pepper in each grid. Cayenne pepper is taken directly and weighing is done with a digital scale. In this study there were two variables, namely the results of image analysis of leaves with whitefly symptoms and the yield as the actual weight. Yields are predicted using image analysis that has been processed using MATLAB software. After the data is obtained, then a linear regression analysis is performed to determine the level of accuracy.

Survey and Identification

Field survey activities for cayenne pepper plants were carried out directly on the farmers' land in Pucuk Village, Dawarblandong District, Mojokerto Regency. The survey activity aims to find cayenne pepper plants in the generative phase that are attacked by whitefly.

Based on the overall morphology of the specimen with the key of determination, the insect belongs to the whitefly pest with the general name spiraling whitefly. According to Kajita et al. [7], The species A. dispersisus is a whitefly that is polyphagous and cosmopolitan. A. dispersisus generally live in groups. In high populations, whitefly colonies can cover almost the entire underside of leaves.

Image Capture and Processing

The images produced in this study are then grouped into blocks to facilitate image processing. Blocks are labeled from block A1 to block F5, so there are a total of 30 data resulting from image acquisition. The block labeling process can be seen in Figure 4.

Figure 3: (A) Field of Cayenne Pepper Attacked by Whitefly Pests (B) Whitefly Population On the Underside

Figure 4: Image Acquisition with Block Quadrants

Figure 5: Mobile Phone Image (A) Original Image (B) Multi-Thresholding Results (C) Filter Results for Whitefly Infested Leaves

Figure 6: The Results of Drone Image Processing at A Height of 10m (A) Original Image (B) Multi-Thresholding results (C) Filter Results of Leaves Attacked by Whitefly

Figure 7: The Pattern of Direct Relationship Calculations and Image Processing Results

Figure 8: Correlation of the Calculation of the Whitefly Attack Directly and The Results of Image Processing

Table 1: Percentage of Leaf Pixels with Yellow Symptoms

The acquired image is then segmented using multi-thresholding. The first process of image processing is to display the image in the Matlab application. Furthermore, the image is converted into greyscale by extracting the yellow color component from the original image. This aims to simplify the image object. The next stage is a multi-thresholding process with threshold values ​​in the range 100 and 230. Pixel intensity values ​​between 0-100 enter region:

• Pixel intensity values ​​between 100-230 enter region

• And pixel intensity values ​​between 230-255 enter regions 

• Image processing is applied using drone imagery, drone image processing is expected to be easier and more efficient than using a cellphone camera

Figure 9: Graph of Whitefly Attack Levels On Cayenne Pepper Production

The calculation of the percentage value of pixels with yellow symptoms is with the Matlab software, namely the number of symptomatic pixels divided by the total number of pixels multiplied by 100 so that the percentage of leaves with panda symptoms can be found in Table 1.

The whitefly attack pattern was carried out to see the relationship between attack percentages. Whitefly pest attack calculation is done directly by counting the number of plants that are sick with pandas in each quadrant. The relationship and pattern between direct calculations and the results of image processing can be seen in Figure 7.

Based on Figure 13, the trend of direct whitefly attack calculations and the results of image processing show a similar pattern of intensity although there is still a gap between the two. Then to see the level of relationship or R value between the calculation of the attack intensity directly and the results of image processing, you can see the panda in Figure 8.

Based on Figure 8, shows that the results of the correlation level of attack R directly with the results of image processing is 0.6. According to the table describing the correlation test values ​​according to Pratama et al. [8], this value shows a strong correlation between direct calculations and image processing of pandan cayenne pepper plants. The effect of whitefly pest attacks on the production of cayenne pepper can be seen in Figure 9.

Based on Figure 9, it shows that at a high level of whitefly attack there is a decrease in the productivity of cayenne pepper, while in the quadrants that have a low attack rate the productivity is higher. Thus it can be said that the attack of whitefly pests on the cayenne pepper plantations affects the production of cayenne pepper. The potential for damage caused by whitefly attacks is quite high, this is because whitefly attacks can cause damage to the physiology of cayenne pepper plants. Physiological damage caused by whitefly pests because they suck the liquid on the leaves so that the leaves turn yellow and curl and even fall off, besides that they also emit black sooty dew which can inhibit the process of photosynthesis and become a vector for the Gemini virus. Tuhumury and Amanupunyo, [9], reported that cayenne pepper farmers in Waimital Village, Kairatu District, experienced damage to up to 42% of their plants due to whitefly pests.

Drone imagery analysis can be used to see the level of attack and estimate the level of damage caused by the whitefly pest attack with a value of 0.6 in the strong category in Cayenne Pepper plantations in Brejel Lor Hamlet, Pucuk Village

Acknowledgment

The authors thank to my lecturer Dr. Ir. Herry Nirwanto MP and dr. Ir. Wiwin Windriyanti MP. In addition, the authors also thank to colleagues especially apple’s farmers at Pucuk village, Dawarblandong sub-district, Mojokerto district, East Java who facilitated for field observation and greatly assisted the research

1. Seminar, K.B. "Paradigma Pendayagunaan Teknologi Informasi untuk Pertanian." Prosiding Seminar Nasional Seminar Informatika Pertanian Indonesia, 2011, pp. 34–42.

2. Sari, V.D. and B.M. Sukojo. "Analisa Estimasi Produksi Padi Berdasarkan Fase Tumbuh dan Model Peramalan Autoregressive Integrated Moving Average (ARIMA) Menggunakan Citra Satelit Landsat 8 (Studi Kasus: Kabupaten Bojonegoro)." Geoid, vol. 10, no. 2, 2015, pp. 194–203.

3. Johnson, M.D. et al. "Crop Yield Forecasting on the Canadian Prairies by Remotely Sensed Vegetation Indices and Machine Learning Methods." Agricultural and Forest Meteorology, vol. 218, 2016, pp. 74–84.

4. Holik, A. and R.R. Bachtiar. "Prediksi Hasil Panen Padi Menggunakan Pesawat Tanpa Awak." Jurnal Ilmiah Rekayasa Pertanian dan Biosistem, vol. 7, no. 2, 2019, pp. 230–238.

5. Putra, I.W.A.S. et al. Kualitas Foto Udara pada Berbagai Ketinggian Aerial Photo Quality on Various Altitude. n.d.

6. Su, R.D. "Application of UAV in Modern Agriculture." Jiangsu Agricultural Sciences, vol. 47, 2019, pp. 75–79.

7. Kajita, H. et al. "Discovery of the Spiraling Whitefly Aleurodicus dispersus Russell (Homoptera: Aleyrodidae) from Indonesia, with Notes on Its Host Plants and Natural Enemies." Applied Entomology and Zoology, vol. 26, no. 3, 1991, pp. 397–400.

8. Pratama, I. et al. "Distribusi Spasial Kerapatan Mangrove Menggunakan Citra Sentinel-2A di TAHURA Ngurah Rai Bali." Journal of Marine and Aquatic Sciences, vol. 5, no. 2, 2019, pp. 192–202.

9. Tuhumury, G.N.C. and H.R.D. Amanupunyo. "Kerusakan Tanaman Cabai Akibat Penyakit Virus di Desa Waimital Kecamatan Kairatu." Agrologia, vol. 2, no. 1, 2018.