We are going to check the quality of the seed using image processing. We will use fpga and interface the camera with it to get the output.

Demo Video

Project Proposal

1. High-level project introduction and performance expectation

Introduction:

The seed quality identification is very important in agriculture. Before boring the seed  in-farm it must be viewed properly and then sowed. In the current scenario the farmers  are taking more effort in their farm and also spending more time and money. But in spite of their hard-work they do not get proper profit. So, technology can come to the rescue here. There are certain limitations to the human eye to observe the seed. So, the electronic world helps us to separate the faulty seeds from quality seeds.

Purpose:

Seed testing is done to assess seed lot attributes and determine overall quality and  value for seedling production and storage. Seed testing standards are based on  scientific evidence and provide set procedures for facilities to conduct tests in a  uniform manner and ensure comparable results for seed owners. The seed testing  standards described below are closely aligned with the International Rules for Seed  Testing and follow these procedures for sampling, moisture content determination,  purity analysis, and seed weight determination. Seed analysis and classification are made to obtain information about seed type,  variety, quality and the production. Pure, disease-free and insect-free seeds can be  defined as quality seeds.  

Image Analysis shows many important advantages over manual techniques. It  provides rapid analysis as compared to any of the conventional methods. Seeds are  not subjected to any kind of treatment or damage. Once the system that works has  been designed then the whole process can be automated.  

Image processing has been applied to various processes of the agricultural industry in order  to achieve fast and accurate operation. Applying image processing techniques to  classify the seeds based on their varieties is also an objective method in  real time applications. Seed analysis and classification can provide additional  knowledge in their production, seeds quality control and in impurities identification.  Also, it is very important to confirm the variety of the seed before  planting. Because each variety of seed needs its own condition for good yield.  

The paper is proposed to present a review of how to check the quality of  the seeds using image processing.

FPGA:

We are using FPGA in image processing, as they are often used as a platform for real time image processing applications. Their structure is able to exploit spatial and temporal parallelism. The approach used is a windowing operator technique to traverse the pixels of an image and apply the filters to them. This makes it easy for us to perform further operations on the image. 

2. Block Diagram

Block diagram:

3. Expected sustainability results, projected resource savings

Parameters:

There are many parameters to determine the quality of the seed. Parameters we are using to determine the quality of the seed are size, shape, color, seed damage, etc.. after analyzing the parameters the quality of the seed is determined. The seed image must qualify the parameters from above to get the required output.

4. Design Introduction

Design introduction:

We are going to interface the camera to the FPGA. After interfacing we are going to take the image from the camera. 

Using Verilog code, operations will be performed on the image. We are going to dump the code into the FPGA, the result is being displayed on VCA 

By observing the outputs the seed quality can be assessed.

5. Functional description and implementation

Functionality:

After the input image is taken we need to read and write the image. The input image will be undergoing image pre-processing, in this step the image will be resized and cropped.

The next step is image edge detection we are using an edge detection process and binarization of the image. In edge detection, we will get the outline of the seed, and by binarization the image will become black and white. The seed part becomes black and the remaining part will be white.

After the bit file is generated, the image generated here will be in hexadecimal. In order to view the image we have to convert the image to bit file 

This bit file will be given to the FPGA board, FPGA will perform the operation and the output will be displayed on the connected screen.

6. Performance metrics, performance to expectation

The design will be able to determine the quality of the seed after analysising the image. After analysing the image we determine the quality of the seed.

Performance expectation: 70%

FPGA benefit:

FPGA clock rates are on the order of 100 MHz to 200 MHz. These rates are significantly lower than those of a CPU, which can easily run at 3 GHz or more. This will help our process to be done fast. 

7. Sustainability results, resource savings achieved

Design scheme: 

The image can be taken prior or it can be taken at the time of excecution. The image taken will be converted to hexadecimal code. After running the code we will dump the code on the board. We will get an output dipslayed on the screen.

Hardware design block diagram and software flow:

8. Conclusion

Conclusion: 

Once the output is obtained we are able to analyse the image. However, there are many more parameters to determine the quality of the seed. Here, we considered some of the important factors for it. 

We can use this quality check for any kind of seed while we used canola. The output we got was 75%  accurate.