I. Introduction
Celery (Apium graveolens) is a vital vegetable grown all over the world for its food, health, and nutrition importance. In India, people are paying more attention to growing celery [1]. It's popular because it can be used in many traditional foods. Also, holistic medicine these days recognizes its good effects on health and so does everyone else nowadays. Farmers need to know and tell apart different types of celery [2]. This helps them take better care of their crops, separate products at the market correctly, and assist customers in making good choices when buying food. The need for the right and quick sorting of celery types has made farmers look at high-tech tools, mostly using machine learning systems. This research endeavors to classify three prominent celery varieties in India: Pusa Jyoti, Utah, and Pusa Lehar used CNN along with K-NN. Celery, which belongs to the Apiaceae family and has many types called cultivars, looks very different from each other. The three types being looked at - Pusa Jyoti, Utah, and Pusa Lehar- have their special physical features. They include differences in leaf shapes, stem colors, and overall plant structures [3]. These differences, even if not easily seen by people, are important for sorting different types. The farming land in India has a wide range of weather and soil, which helps grow different types of vegetables. In this situation, knowing and understanding different types of celery is important for both farmers and people who eat it. Farmers get help from exact planting methods made for certain types of plants, which improves both the amount and quality they produce. At the same time, people get information to make smart choices based on what they like and need for good health. The reason for this study is that it needs to improve the way celery types are identified and sorted in India. Traditional ways of sorting often use people's watching and knowledge. This can take a lot of time, have errors based on feelings, and be wrong many times. Using the latest machine learning tech can make this sorting process better and faster [4]. The main goal of this study is to make and test models using CNN and K-NN methods. These will help correctly classify the given celery types [5]. Using the power of CNNs, or systems that are good at classifying pictures. This helps find hidden details in celery photos and lets it strongly label them. Also, using the K-NN algorithm for recognizing patterns can give a comparison and maybe add another way to classify different things [6]. This study uses a step-by-step way that includes getting data, cleaning it up, making models, teaching them, and checking to see if they are working well. The group of pictures includes different types of celery - Pusa Jyoti, Utah, and Pusa Lehar. These were gathered from open fields with the help of farming specialists. Preparation methods like making images normal, changing their size, and taking out features get the data ready for model practice. The CNN model uses layers for getting features and fully connected ones for classification. Training is about making the model better step by step using the data collected. In the same way, the K-NN algorithm uses features to sort celery photos according to how much they are similar. This research is important because it could change how the types of celery in India. The good use of CNN and K-NN models could give a solid, speedy, and big way for farmers, researchers, and suppliers in the farming world to do their jobs well. It also helps in the ongoing talk about using machine learning for identifying plants.