VISUAL INSPECTION IN THE LEATHER INDUSTRY

ATTENTION:

BEFORE YOU READ THE ABSTRACT OR CHAPTER ONE OF THE PROJECT TOPIC BELOW, PLEASE READ THE INFORMATION BELOW.THANK YOU!

INFORMATION:

YOU CAN GET THE COMPLETE PROJECT OF THE TOPIC BELOW. THE FULL PROJECT COSTS N5,000 ONLY. THE FULL INFORMATION ON HOW TO PAY AND GET THE COMPLETE PROJECT IS AT THE BOTTOM OF THIS PAGE. OR YOU CAN CALL: 08068231953, 08168759420

WHATSAPP US ON 08137701720

VISUAL INSPECTION IN THE LEATHER INDUSTRY

Belgium the spreading of the moulds, 6 min for the cutting and 6 ~h~ leather industry is one that is certainly in need of min to sort the cut parts: About 75 % of the total area of automation. As an example we will deal in this paper with the leather hide is wed. the production of leather seats. Many steps in that proThese steps are certainly good candidates for automaduction process are subject to automation. These include tion. the inspection of the leather hides, stock management, seThe incoming inspection is a first step to automatize, leetion of hides needed for a particular set of seats, using some kind of vision system. Thereafter, the hides are nesting of the (taking into account the defects in put in stock, ordered according to their color, texture and the leather) and the cutting of the hides. These steps quality. Also these manipulations and the stock will be explained in somewhat more detail, however the are subject an extensive largest part of the paper will be devoted to the automatic The next step to automatize is the detection of the visual inspection of the leather. The goal is to detect and the hide when it is placed On the to be describe t,he hides together with their defects for the subworked UP. These the boundary Consequent nesting of the moulds. The algorithms presented ditions for a subsequent nesting the moulds. here are based on the measurement of means, variances An automatic visual inspection follows to detect and and edges, and the application of confidence intervds on classify the defective zones or the zones of lower quality. them. A few results will be shown. This topic is dealt with in the subsequent sections of this paper. After the determination of these boundary conditions INTRODUCTION an automatic nesting of the moulds is started. Therefore, global optimization routines are needed, such as simulated At the present time the leather industry is not automannealing. This will result in savings of material over what atized, let alone a few exceptions. The production process a human operator achieves. The last step is then to cut is tt~ostly based on hand labour, which results in an output the leather on a X-Y controllable table with a waterjet. that is heavily depending on subjective criteria. AutomaLaser cutters are not so adequate here because of focussing tion based on visual inspection can lead to an increase in problems. efficiency, a higher productivity and a more stable quality Except for the difficulties of the automatic fault detecdue to objective criteria (computer). The use of computers tion and nesting of the moulds, also the needed resolution in the production steps gives the possibility, through the poses some problems. A leather hide has a typical area of use of statistics, to follow up certain parameters related to about 4 m2 and the required level of detail is about one the leather. Such a system makes it possible to deliver the mm. This means that some kind of X-Y controlable table designer in an optimal way (gain of materials) the desired with a matrix or array camera is needed to acquire the hides with the specified characteristics as quality norms images. Due to the size of the hides different subimages and color for every shape or for a part of the shapes (e.g. will be created during the scanning so that they have to be e lower quality may be tolerated in hidden zones). fit together to form one large image. Also a special vision The leather hides are delivered in bales of about the system with parallel architecture is needed for the st,orage same color, coarseness, thickness, texture and quality. At and fast processing of the images. this moment the hides are sorted manually (incoming inThe mentioned possibilities of automation are studied spection) and worked up as soon as they are needed. At by different companies, in a research project founded by that time thenumber ofhides satisfying the required shape, the government. The role of our lab is to develop a fast quality, color and texture are fetched from the stock and contourfollower and to study methods for the automatic the set of moulds, representing the shapes to be cut out, defect detection. In what follows, a defect detection will are taken. The leather hides are spread out on large tables, be treated. whereafter the moulds are distributed on the hides. This is done so that the best efficiency is obtained as well as GENERAL CHARACTERISTICS OF THE the required quality, taking into account the zones where HIDES some small defects and a minor quality are acceptable. Then the leather is cut by hand. All this requires about 6 In contrast with automatic fault detection for the most min for the quality control, 6 min for the optimization of sorts of textiles, no periodic structure is found in hides IAPR Workshop on CV Spedal Hardware I and Industrial Applications OCT.12-14. 1988, Tokyo Arep faults are visible due to a local deviation in grey value or variance. Different forms and sizes occur Figure 1: correlation in function of distance (in pixels) upon which a defect detection can be based. In fact we cannot speak of a real texture, so that defects are mainly found by a statistical analysis of the image data. Among different hides, a big difference between the mean grey value and the range of the grey values (coarseness) can be present. This depends on the dye with which the hide is colored and on the presence (or absence) of well visible folds of the skin. Another elementary characteristic of an image is its histogram. However almost none of the faults can be localized using the histogram. Even the presence (or absence) of defects cannot be discovered out of it because as well for faultless parts as for parts with defects the histogram is almost symmetric around a certain grey value. A histogram is of course not sufficient to characterize the grey value distribution around a given point in the image. This can be represented by the correlation between neighbouring points. As Fig 1 shows, neighbouring points are not uncorrelated, which is probably due to camera defects and correlations present in the hides.

HOW TO RECEIVE PROJECT MATERIAL(S)

After paying the appropriate amount (#5,000) into our bank Account below, send the following information to

08068231953 or 08168759420

(1) Your project topics

(2) Email Address

(3) Payment Name

(4) Teller Number

We will send your material(s) after we receive bank alert

BANK ACCOUNTS

Account Name: AMUTAH DANIEL CHUKWUDI

Account Number: 0046579864

Bank: GTBank.