HYDROGRAPHIC SURVEYING AND DATA PROCESSING

Transcription

1 HYDROGRAPHIC SURVEYING AND DATA PROCESSING by A. Thunberg Head o f Survey Techniques H ydrographic Departm ent Swedish Board o f Shipping and N avigation L ectu re given at the 9th In tern a tion a l H ydrographic Conference Summary Description of a system of autom atic data processing for use in hydrographic surveying produced fo r the H ydrographic Departm ent o f the Swedish Board of Shipping and Navigation. In recent years we have encountered m ajor problem s in m any fields when asked to try out new techniques and methods w ith a view to m aintaining or increasing production despite lack of qualified personnel and to few er w orking hours. This problem has been particularly acute in governm ent enterprises. Thus the Swedish hydrographic su rveyin g authorities have been faced w ith the necessity o f fin ding alternative means for m aking measurements, and assembling them fo r processing and recording the results. Th e principles of our w ork and the results achieved have already been described in a series o f articles in The In tern a tion a l H ydrographic Review. On the other hand the data processing system w orked out by the H ydrographic Departm ent in collaboration w ith a firm o f consultants has been touched on only b rie fly at the 196 International H ydrographic Conference and in a short announcement in the B u lletin. T h e A D P system has been demonstrated to hydrographers from several countries. H ow ever it is only now that we feel w e have had sufficient practical experience o f the system to be able to present it m ore thoroughly. T h e requisite subroutines have been w orked out and the program m e system can now be regarded as complete. W hen this report was w ritten the data processing system had been operating fo r somewhat m ore than two years, and the experience derived from it is m anifestly positive. This does not, however, mean that the system has not been revised during this time, and fu rth er m odifications w ill be made in order to deal w ith the num erical data m ore rapidly, thus im proving the econom y o f the processing w ork.

2 The background About ten years ago an analysis was made of the needs o f hydrographic surveying in Swedish waters. It was found that w ith the system then in use it w ould take about seventy years to survey the sea areas and archipelagoes satisfactorily. From various quarters we received requests for a less tim e-consum ing survey program m e. There was a particularly w ell ju stified need, fo r purposes o f national economy, for a better know ledge of hydrographic conditions fo r use in transport econom y studies in connection w ith harbour projects. During the last fe w years the H ydrographic Departm ent has been engaged in experim ental w ork w ith a view to increasing its capacity w ithin the fram ew ork o f the personnel and economic resources at its disposal. Th ere is no need to go into details; we shall here only indicate the measures w h ich first led to the adoption o f autom atic data processing methods. F orm erly the echo-sounder equipped surveying craft for sea surveys w ere large, relatively heavy, and slow (i.e. approxim ately seven knots). T h ey w ere also expensive, both to purchase and to m aintain. Usually these boats needed a crew o f six men, including two qualified surveyors. In this system a depth record was kept during the measuring routine by the boat s crew. These figures w ere later entered on charts by specially trained personnel. In relation to the amount of measurement data assembled this m ethod called for personnel and m aterials in considerable quantity, particularly during surveys at sea using vessels w ith large crews. Nor did this method offer much scope for increasing capacity w ithin the fram ew ork o f available resources. Th e amount o f data gathered, in relation to the m anpow er and m aterials available, has increased significantly, even for coastal and archipelagic waters, w ith the introduction of parallel sounding and electronic positioning. In a parallel-sounding form ation o f nine sounding boats for example, the personnel can consist o f two qualified surveyors and fou r conscripted seamen in the lead ship, and two conscripted seamen in each o f the eight boats running parallel. In addition two or three conscripted operators at a land station are needed. On account o f their small crew the parallel boats themselves can be small and relatively cheap to purchase and m aintain. Since the electronic positioning system keeps continuous check on the ship s course it has become possible to double the sounding speed, i.e. to m aintain 15 knots. This method, however, does not permit the evaluating and recording o f the data w hilst underway. Consequently the echogram s have to be collected afterwards and translated into depth figures on the plotting sheet. Th e new parallel-sounding m ethod has meant an im m ense increase in the number o f echograms, the manual processing o f w hich w ould constitute a practically insurmountable problem particularly in the realm o f recruiting draughtsmen. Inspection and drawing work o f this kind is extrem ely m onotonous and tiring, and this m ade it evident

3 from the outset that considerable staff problem s w ould be involved. A large-scale turnover o f staff w ould obviously be a great disadvantage in terms o f quality. There is also a considerable risk o f errors arising due to fatigue and to the tediousness caused by m onotonous and repetative w ork. In other spheres autom atic data processing has proved to be the only solution w hen dealing w ith very large quantities o f figures. T h e possibility o f using A D P fo r hydrographic surveying was closely exam ined at the H ydrographic Departm ent, and it was very soon evident that in principle there w ere no insuperable problem s in the application o f data processing to echograms. The principal difficulties seemed to be connected w ith finding a w ork routine com bining sim plicity o f procedure w ith dependability and cla rity in the presentation o f the results. Preliminary discussions In the Swedish H ydrographic O ffice data processing has been used fo r numerous routines in the fields of geodesy and cartography fo r some years now. F or certain purposes the program m es have been w orked out by our own staff, w hilst for others the computations have been made w ith existing standard program m es. In other words, w e have already had some experience o f A D P. Th e first discussions on the evaluation o f echogram s by means o f data processing w ere begun in the spring of W ith Mr. A. Thunberg as leader o f the project, the w ork was com m enced in cooperation w ith Mr. B. Lindblom, o f the Board s D epartm ent o f Ligh ts and Electrical Engineering, together w ith specialists on autom atic data processing from the firm o f Consulting Engineers, N ordisk A D B. T h e main object was to fin d a reasonable w ay o f approaching the problem from the point of view o f cost. A vailab le econom ic resources w ere lim ited, and this m eant that a number o f alternatives that in them selves w ere interesting and attractive had to be discarded. A ttention was firs tly devoted to the question of how the echosounder data w ould be converted to digital form. Am ongst other things was discussed the possibility o f recording sounding figures from the echo-sounder direct onto punched tape or m agnetic tape on board the surveying ship or boat. Th is procedure has since been developed, and is now being tested by the Finnish B oard o f N avigation. H ow ever at the prelim in ary discussions o f the Swedish H ydrographic Departm ent is was decided that this alternative w ould be altogether too expensive, particularly in regard to parallel sounding, since each boat w ould require its own electronic apparatus either for punching the tape or card or else fo r transm itting the data to the lead-ship. It w as also felt that there w ould be great difficu lty in filterin g out interference and irrelevant echoes from such things as fo r exam ple fish shoals. Y et another argument against the direct collecting o f digital data w as that it w ould have to be recorded at short intervals as otherw ise im portant inform ation would be missed, particu larly when sounding areas having an exceptionally h illy seabed and frequ en t narrow deeps. Th ere w ou ld thus be a

4 é large number o f tapes, and it w ou ld take a long tim e to feed the digital m aterial into the computer. A ccordingly a big reduction in the amount o f data is very desirable, although the bottom p rofile on the fathogram must not be generalized too much. It w ill be another m atter when echogram profiles can be recorded directly onto m agnetic tape. T h e Finnish Board o f N avigation has solved the problem o f reducing the quantity o f punched sounding data in a very ingenious m anner : in an electronic storage unit w hich is coupled to the echo-sounder eight successive values are compared, and only the m inim um depth in each group is punched. Y et despite this reduction the number of punched depth values is much greater than in the evaluation procedure w here the bottom p rofile is described by a series o f soundings selected by a trained operator. In m any instances by this means the profile can he described quite satisfactorily through punching only tw o or three per thousand echo-soundings. T o this should be added the im portant rem ark that the human eye is a good filte r fo r detecting anomalies. As a result o f the consultations, in the spring o f 1961 it was decided that the digitalization o f the echogram s analogue data should be carried out afterw ards on land. W h a t are the analogue depth data which have to be digitalized, i.e. translated into inform ation readable by a computer? Figure 1 shows an exam ple o f a pattern where a tw o-range system is used for position fixing. In the same figu re an echogram w ith the requisite annotations for position identification is also shown. Th e echogram records the seabed p rofile along the track of the ship or boat. In order that the depths registered along this track m ay be accurately entered on the chart, in fo r m ation is needed concerning both the geodetic position o f the track and the positions o f the individual soundings along this track. The track and the intersecting position lines are either straight lines, arcs o f circle, hyperbolas or ellipses. Th e positions o f the individual soundings along the track are found through interpolation between the intersecting position lines. These are m arked autom atically or m anually on the echogram when the ship, or boat, crosses the intersecting position lines. In this w ay the echogram is divided up into short sections usually about 00 to 500 m etres in length the position o f each section being clearly determined. Th e echograms in figures 1 and which refer to a tw o range positioning are from the first starboard parallel-sounding boat. A positioning instrum ent (H ydrodist) is mounted at each o f the tw o triangle points 9 and 1. The distance from the lead-ship to these points can be measured accurately. In this particular case the track is shown as an arc w ith point 9 as centre. Measurements made in this w ay give a continuous check on the ship s position on its course, and this is a great advantage particularly when sounding in an area o f uneven bottom and frequent narrow deeps, fo r here the courses of the surveying boats must be kept close together and no gaps are permissible. In the exam ple in question four boats are shown as fo llow in g a parallel course abreast the lead-ship and at a constant distance from it, tw o boats being to starboard (numbers 1 and ) and two to port (num bers and 4). Before starting a sounding course a printed panel on the echogram is fille d in, showing the exact geodetic extent o f the sounding

5 I 'i, irrj ^ "?Û9 695Q 9 $ 1 i A / / % A / y > i.. Zi f 1I. i 1 1 L T h e resu lt from the hydrographic su rve yin g is a fa th ogram with notations, which m ake it possib le to tra n s fe r any de p th -fig u re to n c o r r e c t position on a plotting sheet. / /5 /+ r N u m b ers identifying the positions of dec-ca antennas. B y w riting the num bers r e fe r r in g to the positions in va riou s ways the fixing method can be describ ed to the com p u ter F ig. 1. Some patterns of sounding lines suitable for Automatic Data Processing. line. Data on the position lines intersecting the lead-ship s course are entered next to the traverse m arkings on the echogram. B elow the stylized bottom echo p rofile in figu re is shown how the seabed p ro file is

6 Automatically punched F i g. approxim ated by a polygon line during the evaluation (digitalization). The black points w ith cross marks are those points which should be evaluated in order to obtain a satisfactory profile. A polygon line o f this type is best

7 Lim it for the area to be surveyed. T h e sounding w ill affect 8 plotting sheet Sounding lines. A A A A -D Kach sounding line is ^ described by four data groups. M apnetic tape m e m o r y fo r each plotting- sheet. D u rin g the data p roc es sin g the data concern ing one sounding- line is divided in parts, each co n cern in g a p a rticu la r plotting sheet and stored on a particu lar m agnetic tape. F ig.. The editing of soundings being processed. described by indicating the coordinates for the break points : (these w ill be called respectively depth coordinates and length coordinates). Th e coordinates are referred to the analogue-digital converter s axis system.

8 The line-printer out-put. E ach plotting sheet is divided in four squares. E ach square is divided in 100 x 60 sm all rectan gles. Only one sym bol can be printed in each rectangle o. t?. a figu re, a letter or any other sym bol availab le in the lin e- printer. T w o kinds of in fo rm a tion a re im portant : a) Depth contour lines. T h e m inim um value of a contour is printed if m ore than one depth contour passes the r e c tangle. b) D e c im e tre - and units of m etre a re printed reg a rd in g the evaluated points along the bottom profile. figu res bs 8 8 a- f', A V 7 :6¾ 9! B 1 1! 71 8?! '1 B dm M in depth within 'he contour line is 5.7 m etres. m / t{ Kit;. 4. T h e lin e -p rin ter out-put.

9 W h a t then are the different stages o f w ork leading up to the com pletion o f the hydrographic chart? In the processing there are five individual phases : (a ) Transferring the inform ation on the echogram to the punched tape; (b ) Exam ining the plausibility o f the data punched on the tape by running the tapes through the com puter; ( c ) Correcting any errors; ( d ) The final input o f the punched tapes into the computer, and the storage of all surveying data on m agnetic tape; (e ) Editing and printing o f lists o f depth data by a rapid line-printer connected on line to the com puter. These lists o f data show the soundings in their correct position. (See figures and 4). Transfer onto punched tapes A fte r certain experim ents it was found advisable to divide the in fo r m ation assembled on the echogram into fou r data groups (see figu re 5). Data group 1, consists o f eight numbers from the echogram stamp, and is punched m anually; it defines the survey project, describes the geodetic extent o f the sounding line and gives the m ethod by w hich the positioning is carried out (i.e. two-range, hyperbolic, etc.). Figu re 6 shows the im pression o f the echogram stamp. Since the 1966 surveying season the recording paper fo r the echo-sounder has this stamp printed on it. P a n e l 1 identifies the survey project by means o f a jo b number. Panels, and 4- are used fo r d efin in g the position fixin g method. The positions ashore for the hydrodist, the Decca antennae, etc., are determ ined geodetically and are given specific numbers. B y placing these point numbers in different com binations in panels, and 4, the m ethod em ployed for positioning is described. B y means o f data processing, the m achine program m e through logical selection can itself select the appropriate subroutine fo r coordinate com puting. Pa n el 5 describes the track o f the lead-ship in a parallel-sounding group (i.e. by the radius o f a distance-circle or a hyperbole num ber). Panels 6 and 7 are used fo r describing the position o f the parallelsounding boats in relation to the lead-ship. Panel 6 shows the boat s distance from the lead-ship, and panel 7 the position of the boat in the sounding sweep (odd numbers to starboard and even num bers to port). Panel 8 is filled in afterwards and contains a correction for tide and transducer depth. In the early stages o f the data processing it was assumed that the zero adjustment of the echo-sounder was alw ays correct. In principle the recorder must be adjusted so that the zero echo on the echogram corresponds to the transducer depth. H ow ever it w ould be unreasonable to expect such a result alw ays. Experience has also shown

10 D e s c r ib e s the geod etic e x tent of the sounding line. ManuaU^ punched. D a ta gro u p D e s c r ib e s the geo d etic lines g iv in g r e f e r ence points alon g the sounding line (x and y c o ord in ates in the geod etic s y s te m can be com puted) M a n u ally punched. D a ta gro u p D e s c r ib e s the re fe r e n c e line on the fathogram (length and depth co ord in a tes r e fe r to the axes of the coordinate tab le) A utom atic punching. u a lê t fc 'i'o U { J 4 D e s c rib e s the bottom p ro file as a p o ly gon line (1 and d coordinates r e f e r to the axes o f the coordinate table) A utom atic punching. T e s t r e g a rd in g the p lau sibility of the n u m erica l content of the punched tape. L is t reg a rd in g prob ab le e r r o r s. Y es Any e rro rs? No P a p e r tape m anually c o rr e c te d o r the en tire p ro file (four datagrou ps) is r e evaluated T e s t r e g a rd in g the plau sibility of the paper tap e, with re-eva lu a ted profiles T o the final computation. F ig. 5

11 AREA NO MAIN SHIP LANE GUIDING SLAVE DIST. (metres) CROSSING SLAVE REL. POS. (No.) DECCA MASTER TIDE ( + or ) dm TRANSD. DEPTH drn SURV.-RANGE USED FREQV. USED I s I 6 * 7 I 8 WIND/SEA DATE YEAR TIME SHIP/LAUNCH SIGN. CONTR. EVALUAT. F i g. 6 that considerable deviations regarding the adjustment of the zero echo occur from time to time. Since the distance between the zero echo and the bottom profile represents the actual measured depth the zero echo can be used w ith advantage as a reference line. In this w ay it has been possible to dispense w ith the printed graduations on the record paper, because if fo r any reason the zero echo alters its position then the bottom p rofile follow s. Data Group is punched manually and relates to intersecting position lines. I f the intervals are o f equal length all along the sounding course then it is sufficient if the operator punches the data fo r the first, second and last intersections only. Thus fo r the planning o f the sounding w ork we endeavour to achieve regular divisions o f the echogram all along the bottom profile. H ow ever there are occasions when it is necessary to deviate from this principle; for example i f the sounding sweep has to alter speed. In the case o f irregular divisions all the intersecting position lines have to be punched on the tape. Data group is punched autom atically and describes a reference line on the echogram (norm ally the zero echo). A t those points where the zero echo line is intersected by the transverse m arkings the evaluation table s m obile measuring mark is accurately set, and by pressing a button a pair o f coordinates (length and depth), referred to the axis system o f the evaluation table, are registered on the punched tape. Data group 4 is also punched autom atically and describes the bottom profile w hich is approximated by a polygon line. Each point is defined by a depth and a length coordinate, always w ith reference to the axial system o f the evaluation table. Th e depth at each individual point along the sounding line can now be easily calculated as being the difference between the reference line and the bottom profile depth coordinates. Corrections for tide and transducer depth are added to this difference in the further computations. (See figu re 5).

12 Figure 7 shows a print-out from a paper tape with data regarding one sounding line. P R IN T O U T F R O M A P A P E R T A P E. Data r e g a rd in g one sounding line Data group A A A A T h is " e x t r a " figu re is punched by the o p erator to give a second definition of the positioning method used t Gooul G i t D _ ? S * f C P o s itiv e values in data groups and 4 a re length coordinates and negative valu es depth coordinates. T h e coordin ates r e f e r to the axes of the evaluation table and have been given differen t signs to m ake it possible to distinguish betw een depth and length coordinates. F ig. 7 Test of the punched tape content for plausibility W h en all the echograms relating to a particular survey project have been transferred to punched tape there w ill be material for further processing. Th e w ay in which the transfer has to be carried out does not perm it double punching fo r content checking. It is possible, however, to make the punched tape content conform to certain requirements. The number o f intersecting position lines in accordance w ith data group must agree w ith the number o f registrations in data group ; the length coordinates in data groups and 4 must always increase; in data group 1,

13 only a certain number o f the position points w ill be quoted, etc. Thus w ith the aid o f a computer the plausibility o f the punched tape figu res can be checked, and this requires very little com puter time. Th is checking results in a line-printer list of presumed errors and the nature o f the errors is also ca refu lly described. Th e list could fo r exam ple give the fo llo w in g in form a tion : Tape No. 8, 1th sounding course, data group, number value 6, fo llow ed by comm ents on the character o f the error. A ll the data concerning the sounding courses involved are also assembled on a record. W h en all the tapes relating to one jo b number have been processed the sounding courses are sorted, and a com plete list is obtained via the line-printer. Such a list makes it possible to check that all the echogram s have been included in the processing. In m ost cases the indicated errors can now be rectified m anually direct on the paper tape by means o f a sim ple tool (a punch). If, how ever, this is not possible then an incorrect sounding course can be erased from the tape by m anually punching special symbols on the paper tape. T h e echogram in question has then to be retransferred to the analogue-digital converter and the tape tested again (see figu re 5). Storage on magnetic tapes A fter testing and correcting w ork the m aterial is prepared fo r com puter processing, the fin al result o f w hich is a list o f data w here the soundings are printed (by a line-printer) in their correct positions. This list o f data is used as a m anuscript w hen the fa ir sheet is drawn. T h e fo llow in g are the phases o f the w ork involved (see figures and 4). 1. D eterm ining the geodetic plane coordinates for the intersecting points o f the sounding lines w ith the position lines. Th e com puter itself selects the subroutines, based upon in form ation in the data group I, fo r determ ining the coordinates. Since the position lines fo llow ed by the sounding ship and boats are usually second degree curves, the com puter program m e has a sub-routine fo r inserting extra reference points along the p ro file so that the positions o f individual depth figures can be determ ined through linear interpolation in the fu rth er com putations.. Possible correction o f data groups and 4 fo r errors arising from the fact that the echogram has been inadequately adapted to the evaluation table.. Correction fo r tide and transducer depth. 4. Collocation o f the com puted x and y geodetic coordinates w ith the length and depth coordinates fo r the reference line and the bottom profile. T h ere is a know n num erical relation between the x and y coordinates o f the transverse m arks and the length and depth coordinates. 5. The division o f each sounding course into parts, each constituted by all the data referrin g to one plotting sheet. Th e size o f the plotting

14 sheets can be very varied, but these w ill always be square. The decisive factor here is that as the m agnetic tape m em ory is lim ited only a certain amount o f depth data for each plotting sheet m ay be stored. W h en processing m aterial from an area w here the general depth is not very great and w here the seabed is uneven, and w here in consequence the sounding lines must be run close together, the plotting sheet should cover an area o f 500 X 500 metres fo r exam ple. F or sparser soundings however, at greater depths, the plotting sheet could cover perhaps 10 x 10 kilom etres. 6. Storing the part-profiles on magnetic tapes, grouped according to plotting sheets. About 000 w ords o f 40 bits are available for each plotting sheet. In the com puter used only thirty plotting sheets can be processed at one and the same time. A certain number o f plausibility tests are also c a rrie d out at this stage in the processing. Drawing up and printing the plotting sheet manuscript The topographic characteristics o f the seabed w hich it is required to record are the general depth conditions, w hich can be indicated through depth curves, and also certain extrem e values such as m inim um and m axim um depths. F or these last greater accuracy in depth inform ation is required. Th e evaluated breakpoints along the bottom profile in the fathogram consist m ainly o f such extrem e values. Thus decim etre figures are generally present in the digital depth data available on the tape. Th e positions fo r different depth contours along the profile can be computed through interpolation. In an earlier version o f the data processing system a tape-guided plotter was used fo r printing the plotting sheet, but a large number o f punched tapes was needed to guide the plotter and the punching o f these tapes via the computer involved great expense. The plotting w ork also took both time and money. It thus very soon became evident that another method was needed. A rapid output device is the line-printer which prints 15 lines per second, and in w hich each line can hold up to 10 different symbols. If this could be used fo r printing the final results the cost o f renting the com puter w ould be reduced and the same time the tim e- consum ing plotting w ork could be dispensed with. T h e line-printer is typographically designed so that if 60 lines w ith 100 symbols per line are printed the result is a square of 55 x 55 mm. I f every square plotting sheet w ere divided into only 60 X 100 small rectangles, the quadratic content of each area would be rather large. I f for exam ple the plotting sheet square covers an area 10 X 10 kilometres then each rectangle w ill cover 100 X 160 metres. As the line-printer cannot print m ore than one symbol in each rectangle the inform ation which could be obtained from such a plot w ould be much too generalized. In order to attain the m axim um resolution in the chart area it is im portant to divide up the plotting sheet into as m any sub-rectangles as possible. This could be done by dividing the square plotting sheet into 4, 8 or 16 sub-squares then divided into 60 X 100 small rectangles. A s all depth data belonging

15 to the plotting sheet are stored on one magnetic tape m em ory the tape could be read by the computer 4, 8 or 16 different times, each reading resulting in depth data regarding one sub-square edited on 60 x 100 possible places fo r depth inform ation. On the other hand in order to save com puter tim e it is desirable to carry out the w ork in such a w ay that inform ation stored on magnetic tape need not be read m ore than once. I f the com puter used has a large core store capacity then we can afford to divide each plotting sheet into an adequate number o f sub-squares, each one containing 100 X 60 possible places fo r inform ation. Th e computer used by the Swedish H ydrographic Departm ent permits the simultaneous editing o f eight such sub-squares. Four o f these sub-squares are used for editing data regarding the seabed topography, w hile the other four contain data for extrem e values, e.g. m axim um and m inim um depths. In principle the extrem e value lists should contain all the breakpoints w hich the punching operator has selected for describing the bottom p rofile on the fathogram. Th is means that every plotting sheet is divided up into 4 x 100 X 60 small rectangles (figu re 4). Th e scale factor can be determ ined quite sim ply. If fo r exam ple the plotting sheet covers 500 X 500 m etres and each quarter plotting sheet measures 55 x 55 mm then the scale w ill be 1/4 900, i.e. very close to 1/ Each sm all rectangle then represents an actual area on the sea bottom o f 0.8 x 1.5 metres. A n y o f the lineprinter s 6 different symbols can now be inserted in each rectangle. Through the optim al use of the core storage o f the computer a total of depth data figures per plotting sheet can be accommodated. O nly in cases w here the seabed is very uneven and the general depth is small need each sm all rectangle be filled in. Certain checks o f the data processing can also be made at this stage. It can happen, for example, that inform ation for a small rectangle is derived from tw o different sounding lines, and in this case the soundings fo r insertion in the rectangle should o f course be much the sam e; if, however, there is too great a divergence the computer produces a paper tape showing the nature of the error. Figures 8a, 86 and 8c show examples o f line-printed plotting sheet manuscripts. Figure 8d shows the result after the contour lines and some depth figures have been m anually drawn using line-printer tables as basic m aterial. I f w ith in a line-printed plotting sheet certain areas are o f particular interest, and the resolution required is higher than is the case fo r a standard line-printed plotting sheet, then it w ill be easy to enlarge to any scale any sub-square chosen w ithin the lim its o f the p lottin g sheet. As m entioned above, each enlargem ent w ill be divided into 4 X 100 X 60 ( = 4 000) rectangles. The enlarged plotting sheet w ill be o f the same size as the standard plotting sheet but o f course at a different scale. If, for instance, one quarter of a plotting sheet is enlarged the scale fo r the lineprinter plotting sheet w ill be tw ice that o f the standard plotting sheet. If during the processing it becomes necessary to exclude one or m ore sounding lines from the computations this can be easily done. If required, sounding lines m ay also be transferred to other positions, fo r exam ple if data have been w ron gly noted on the fathogram stamp.

16 TA B LE ( I ) (from line-printer) WITH SYMBOLS REGARDING DEPTH CONTOURS. F i g. 8 a

17 TABLE (from lin e-prin ter) WITH FIGURES REGARDING EVALUATED DEPTHS. < NV-X NW-Y SE-X 7U075O SE-Y 4 a 9 9 U / / ^ B ¾ * U t>

18 TABLE ( g ) (from lin e-printer) WITH FIGURES REGARDING EVALUATED DEPTHS. NW-X NW-Y SE-X SE-Y U a i A a Z i i) V 'i U V b V w / b b o « e

19 D E PT H CONTOURS AND DEPTH FIGURES. F ig u r e s only w here n e ce ssa ry for the com p le te n e ss of the ch art. T h e draw ing is based on the lin e printer output.

20 Figure 8e carries the legends for symbols used in the line-printer output. E xplanation o f T A B L E ÇF) (fro m lin e -p r in te r ) with sy m b o ls regard in g depth con tou rs. D epth 1-5 m e tre s " m in terval. 5 m h 5 m ii E xp lanation o f T A B L E ( ) (fro m lin e -p r in te r ) with fig u re s rega rd in g evaluated depth from the bottom p ro file Depth le ss than 5 m etres. T h e printed fig u re stands fo r the value o f the unit of m etres. (Depth = is printed) D epth m ore than 5 m etres. T h e printed figu re gives the value in tens of m etres. (Depth = is printed) T h e m ain depths a re given in the table u sin g sy m b ols fo r depth con tou rs. E xp lanation fo T A B L E (s ) (fro m lin e -p r in t e r ) with fig u re s rega rd in g evaluated depths fro m the bottom p rofile D epth le ss than 5 m etres. The printed fig u re gives the the dm. value (Depth = is printed) D epth m ore than 5 m etres. The printed figu re stands for the value o f the unit o f m e tre s. (Depth = is printed) T h e m ain depths a re given in the table u sin g sym b ols fo r depth con tou rs. Fig. Se SYM BOLS used when printing m an uscripts for draw ing of depth con tou rs. 1 = 1 = = 4 = 4 5 = 5 6 = = 7 Q 9 = 9 = = 10 A = 11 B = 1 C = 1 D = 14 E = 15 F = 16 G = 17 H = 18 I = 19 + = 0 K = 1 L = M = N = 4 O = 5 * = 0 P = 5 X = 40 R = 45 = = 50 S = 55 + = 60 T = 65 * = 70 U - 75 X = 80 V = 85 = = 90 w = 95 % = 100 X =1 5 Y = 150 Z = 175 % = 00 O = 5 = = 50 u =75 % = 00

21 The analogue-digital converter Certain difficulties were experienced in getting hold of an analoguedigital converter which would fulfil the requirements of the Hydrographic Department which were as follows : 1. The apparatus must be easily transportable, and it must be possible to use it for field work if required.. The operator must be able to seek new breakpoints on the echogram profile immediately after having given the punch command, and while the punching of the preceding value is going on. This requires a storage function capacity.. It must be perfectly clear from the analogue data (the echogram) by what means the surveying points were selected during digitalization. 4. The number of recordings made per echogram must be easily countable for use in work studies at a later date. 5. The resolution in the analogue-digital converter should not be greater than what is reasonable in relation to all the other factors affecting the accuracy of the echogram profile. 6. It must be possible to place the echogram on the coordinate table without too great a necessity for accurate adjustment. 7. The echo-sounders are used for different depth intervals (0-50 m, m, m). Thus in the converter it must be possible to add the requisite depth : 50 or 100 metres. It was evident that at that time there was no suitable analogue-digital converter on the market at a reasonable price. Consequently we had to design our own apparatus, using the electronic equipment available on the market at that time. There are in principle three different types of converters : (a) pulsecounting instruments, (b) mechanisms equipped with binary-coded discs or photo-electric converters, and (c) precision potentiometers, the voltage of which is read off on a digital voltmeter, whereupon the signal is recoded in binary form. Alternative (c) proved fairly simple to follow, and the cost of building an evaluation apparatus according to this principle seemed reasonable. It might appear circuitous to transform analogue data (presented graphically) to another analogue form (voltage) for digitalization, but in practice this has functioned quite well. The principle is used in many commercial analogue-digital converters. During the last two years Dobbie Mclnnes Ltd. of Scotland, amongst others, have designed an analogue-digital converter a coordinate reader whose performance is adequate for most purposes connected with the digitalization of echograms. The analogue-digital converter built by the Swedish Board of Shipping consists of a coordinate table fitted with a conveyer for a carriage along the echogram (see figure 9). This carriage carries a smaller carriage which moves vertically. A ten-revolution precision potentiometer is attached to the larger carriage. This potentiometer is driven by a wire and a pulleywheel. A three-revolution potentiometer is attached to the smaller carriage

22 F lu. 9 which is rack and pinion driven. In a second digitalization apparatus the wire and rack and pinion system has been replaced by a more efficient transmission a cogged belt of reinforced plastic. The demand for potentiometer precision has been set somewhat higher than the degree of precision ( ± 0. mm) with which the operator can set the measuring mark or visually read the sounding. A plexiglass sheet with a measuring mark has been mounted on the smaller one. In the new apparatus an illuminated measuring mark is projected onto the echogram. The two carriages on the coordinate table are mechanically directed in the same way as the rulers on a large drawing board, so that there is practically no risk of obliqueness. On the smaller carriage there is also a marking device which makes a black dot on the echogram 7 mm above the place on the profile being evaluated. This makes possible a later check on how the evaluation is carried out. The number of evaluated points per echogram can be recorded in a counting device (eall-meter) on the coordinate table. Also coupled to the coordinate table with its potentiometers is an electronic apparatus which handles the measuring, storing and punching of the values on the fivechannel tape. The equipment consists o f : 1. Digital voltmeter. Solartron, with condenser storage.. Encoder. Solartron. Programmer (designed by Solartron, Stockholm).

23 4. Tape punch (Creed 5). 5. Keyboard for manual punching. 6. Phase selector for addition of specified currents, depending upon the range in which the echo-sounder is used. Comments 1. The Solartron LM 1010 digital voltmeter works rapidly and makes 60 measurements per second. Input impedance is very high (1 000 megohm), so that it can measure voltage on a condenser without any noticeable leakage.. The Encoder register contains 5 bits, and this permits the punching of two 5-figure numbers with symbols, as well as space for storage of a sexagesimal symbol. In this case, this symbol is used in such a way that the symbol for the carriage-return (i.e. a new line) is automatically stamped on the tape after each fifth pair of coordinates. This makes it possible to write out the tapes easily in an automatic typewriter in order to check the contents.. The programmer punches the above-mentioned carriage-return symbols and, among other things it also guides the measuring and punching cycles. Figure 10 shows a schedule describing the programmer function that makes it possible for the operator to set the next measuring point while the punching of the coordinates for the former one is still going on. While the operator sets the measuring mark for the first point on the echogram all the contacts are open. When the scanning is to take place contacts A and B are closed. The analogue voltages are stored in the Cx and C* condensers, after which the A and B contacts are opened. D is closed, and the C^. voltage is fed into the digital voltmeter and is punched. D is then opened and C closed, whereupon the C voltage is fed in and punched. When the A and B contacts have been opened the voltage in the r and z potentiometers can be changed without affecting the previous measured value.

24 4. At a punching rate of 5 symbols per second the Creed punch can easily complete the punching of 1 symbols (1 every fifth time) before the operator has had time to move the measuring mark to a new point. 5. For manual punching an electric typewriter is coupled to one of the Department s two apparatus, and this makes it possible to check the punching visually at once. A Lorenz tape-reader has been attached to this equipment so that when required a copy can be taken from the tape and duplicates of the new tape can be made. A keyboard unit has been attached to the other analogue-digital converter. Experience has shown that the error percentage in manual punching is very low, despite the absence of a written-out text. 6. An Atlas Monograf sounder is used for the sounding work. This functions in three different ranges of depth : 0-50 m m and 100 ~ 150 m; by altering the ratio these can be changed to m, m and m. This was borne in mind when the Department s digital converter was designed. By a system of precision resistance, voltage corresponding to 50 and 100 metres can be added to the voltage derived from the coordinate table on the depth potentiometer. Checking the functioning of the analogue-digital converter Great functional reliability is of course required of the analogue-digital converter, and consequently it is essential that the entire transfer procedure from the coordinate table to the punched tapes is checked at regular intervals. This checking takes place as separate operations for the length and the depth coordinates. The linearity in the potentiometers is checked, and also the punched symbols are checked for errors when processing data from such control measurements. The electronic apparatus and the tape punches are subjected to a complete overhaul after a certain number of hours of use. The number of hours the apparatus has been in use, together with details regarding its functioning, are recorded in an instrument log. Concluding remarks As already mentioned, the working group was not set too distant a goal on the question of how to apply data processing to hydrographic surveying. It seemed better to set up a system which might start to be used within a shortish time, and which could thereafter be extended to include further steps towards a fully automated system. Data from a hydrographic survey (depth, time, position, etc.) may be sampled along the sounding lines and automatically stored on magnetic tape during the survey. Thus the basic material on magnetic tape will consist in the Swedish application of ADP of nothing except what is stored on paper tape. There will probably be no difficulty in inserting an automated

25 system for collecting sounding data into the programme system. W ith the profiles as basic data, a detailed list of sounding information in tabular form can be edited in the computer. At present these lists are printed-out as tables via a line-printer, but the thousands of soundings which go to build up this list could be the foundation for drawing up the depth contours automatically. This would probably be very intricate to programme, especially when a rugged bottom topography allows different interpretations. In the very near future we shall have to change from the rather old computer we now use to a modern one, for example to an IBM 60. Thè entire ADP system will then of course be scrutinized and probably revised. There is however nothing to indicate that we shall depart from the main principles, although perhaps the analogue-digital converter will have to be reconstructed to write on magnetic tape instead of paper tape. Perhaps the T.V. screen on a modern computer could be used to present the results. Further studies will show whether or not this is realisable.