Data classification in presence of noise will cause a lot of worse results than expected for pure patterns. In the proposed work we tend to investigate the drawback within the case of deep convolutional neural networks so as to propose solutions which will mitigate influence of noise. The main contributions presented in this proposed work include using convolution neural network as an image classifier for detecting noise in the images and using different opencv2 inbuilt methods to mitigate noise in the images. Though a number of techniques are introduced for this purpose, using neural networks we can achieve a greater accuracy.

Quality Testing of Rice Grains is testing of grain to evaluate the planting value and the authenticity of the certified lot. There are certain limitations to human eye to observe the Grain. So, the electronic world helps us to separate the faulty Grains from quality Grains. The specific target to be achieved is the development of a rice quality detection system that can assess the quality of rice using digital image processing. The evaluation of the rice grains on the basic grain size and shape using image processing edge detection algorithm is used to find the region of boundaries in each grain. Grain testing is done mainly for moisture, germination and physical purity of Grains. The performance of Image Processing reduces the time of operation.