Freelancework-casestudy
- 1. Image enhancement Algorithms Image Pattern matching Video Fire Detection
- 2. Fire detection by wavelet Face Recognition
- 3. Facecount 1.8 Hair recognition 1.9 Hand Gesture Recognition
- 4. Vision based Inspection A vision based inspection product is developed for testing the defects in a manufacturing plant by using CCD /CMOS cameras .The system involves acquisition cameras with IEEE 1394 Fire wire interface and software development on a PC based platform. Software for enhancement, blob detection, pattern matching, segmentation and classification. We have developed expertise in inspection and surface defects detection using machine vision cameras and image processing techniques. Defective parts are automatically detected using machine vision image processing technology. The system consists of CCD camera and optics, Frame grabber, lighting, part sensor, PC and inspection Image processing software with hardware interfaces. Inspection software consist of developed image processing software to detect defects of parts manufactured in a production process (Real time ) automatically .Some of the function of the software is to detect defects such as rust, scratch, parts presence /absence ,measurements & gauging studies, etc. A sample inspection Results are given below
- 5. Figure 3 Radiator cap and Ring travellers Inspection PCB defectInspection ¨CCase study Objective: ? PCB Defect detection & classification ¨CFatal &Potential ? Finding Fiducial Marks based on Circularity,area,perimeter,etc-auto ? Finding PCB soldering Position-auto ? Golden Board (Learning Samples-feature statistics) - Alignment - Parts /component extraction from PCB board based on color information ,measurement Good PCB Defective PCB Registered-aligned
- 6. Image difference showing defects Sample case Study: Original PCB Defective PCB Registered Defects Difference (Positive + Negate) Subtract-positive values highlighted Negative value highlighted Image 255-0 0-255 Breaks , circuit shots and conductor without holes detected
- 7. Noise free threshold image-median filtering A- Finding Fiducial marks location Auto -original & defective original PCB with FM Edge + fill holes Aircraft Shape Detection& Classification-Casestudy ? Using Fourier descriptors ? Using invariant Moments Feature descriptors -45 Parameters (Red color in results-Elongation note() Image,area,mean,Sd,Min,max,XM,YM,perim,bx,by,width,heidght,major.minor,angle.circle,feret,intden. skew,kurt,%area,raw,slice1,feretx,ferety,feretangle,AR,round,solid,cutoff,feret,mass,xc,yc,xxv,yyv,xyv, yyv,xsk,ysk,xkurt,ykurt,orient,elongation(last) 1 SF16_P0_Y0_R0_2-1.png 119140 216.042 91.743 0 255 225.558 129.865 0 0 0 460 259 519.054 292.250 0 0 NaN 25739190 -1.930 1.726 0 25739190 1 NaN NaN NaN NaN 1.776 0.563 1 0 1 25739190 225.558 129.865 19014.543 6507.194 121.767 0.057 -0.009 -1.276 -1.437 0.558 4.028E15
- 8. Fingerprint Minutia Extraction Problem- Extracting Minutia (termination & bifurcations) from Fingerprints automatically Process: - Enhancing the Image - Converting to Binary image - Thinning the image - Identifying ridges & Bifurcations - Termination identification - Bifurcation identification - Removing spurious edges using distance function - Making ROI mask on Image - Overlaying Image on Mask - Extracting co-originates and angles for each terminated, bifurcation points(angles using LUT) - Saving it in a file for each fingerprints - Sample results & code Original Fingerprint image Binary image
- 9. Thinned image Bifurcations, termination identified Position of points marked Termination, Bifurcation Mask area Image overlay on Mask distribution of major points
- 10. Drowsy Driver Detection Lane Blur detection Lanes Traffic sign board recognition Ortho-image Processing¨CAuto-vectorization
- 11. Car number Plate Recognition Car Number Plate Recognition Face Detector
- 12. Face Detector Red light Traffic Detector Traffic Light Detector-Red signal